Citation
Multimethod assessment of depression and behavioral distress in a pediatric population

Material Information

Title:
Multimethod assessment of depression and behavioral distress in a pediatric population
Creator:
Rodriguez, Christina M
Publication Date:
Language:
English
Physical Description:
ix, 156 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Anxiety ( jstor )
Child psychology ( jstor )
Children ( jstor )
Depression ( jstor )
Diseases ( jstor )
Hospitalization ( jstor )
Hospitals ( jstor )
Mothers ( jstor )
Parents ( jstor )
Pediatrics ( jstor )
Adaptation, Psychological ( mesh )
Child ( mesh )
Child Behavior Disorders -- diagnosis ( mesh )
Child, Hospitalized ( mesh )
Department of Clinical and Health Psychology thesis Ph.D ( mesh )
Depression ( mesh )
Depression -- diagnosis ( mesh )
Dissertations, Academic -- College of Health Related Professions -- Department of Clinical and Health Psychology -- UF ( mesh )
Infant ( mesh )
Mothers ( mesh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 86-93).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Christina M. Rodriguez.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Christina M. Rodriguez. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
49674199 ( OCLC )
028084652 ( ALEPH )

Downloads

This item has the following downloads:


Full Text












MULTIMETHOD ASSESSMENT OF DEPRESSION AND BEHAVIORAL
DISTRESS IN A PEDIATRIC POPULATION















By

CHRISTINA M. RODRIGUEZ


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA

1993














ACKNOWLEDGMENTS

First and foremost, I must thank two individuals

who helped transform an idea into reality: Steve Boggs, for his guidance, incredible patience, unwavering support, and long hours; and Sheila Eyberg, as a female role model of unparalleled professionalism as well as for her steadfast enthusiasm and faith. Also, I thank Jim Rodrigue, Mike Geisser, and Faye Gary-Harris, whose excitement about this project energized me and whose suggestions, comments, and questions facilitated the direction and critical evaluation of the project. I thank all of these committee members for helping me pull this together.

Secondly, I thank my mother not only for her invaluable assistance in creating the Child-BUMP pictures but also for instilling in me a deep compassion for human suffering that led me to the field of clinical psychology. I also thank Becky, Brigette, Elena, Jennifer, Melodye, and Randi, my close friends and fellow graduate school survivors, for the emotional support that sustained me and for redefining the depths of friendship. I also appreciate the help of other friends, faculty, and staff at the Dept. of Clinical














and Health Psychology, whose daily contributions to my well-being and training made accomplishing this project meaningful. I thank the group of research assistants who helped with data collection and turned what initially seemed to be an insurmountable daily task into an organized demonstration of teamwork. Finally, I thank the parents and nurses who took the energy from their pressing concerns in order to help the sons and daughters of tomorrow.


iii
















TABLE OF CONTENTS
Page

ACKNOWLEDGEMENTS ................................... ii

LIST OF TABLES ..................................... vi

ABSTRACT ........................................... vii

CHAPTERS

1 INTRODUCTION .................................. 1

General Childhood Depression..................... 1
Prevalence ............................... 1
Assessment Issues ........................ 2
Risk Factors ............................. 4
Pediatric Population .......................... 8
Definition Issues ........................ 8
Prevalence ............................... 10
Assessment Issues ........................ 11
Risk Factors ............................. 14
Pilot Study ................................... 20
Introduction and Rationale................ 20
Methods .................................. 21
Results and Discussion.................... 25
Purpose of Study and Hypotheses................ 27

2 METHODS ....................................... 29

Subjects ...................................... 29
Measures ...................................... 32
Procedure ..................................... 38

3 RESULTS ....................................... 40

Analyses for Current Study .................... 41
Descriptive Results & Statistics.......... 41 Analyses of Background Data............... 43
Correlational Analyses ................... 51
















Analyses of Current Study and Pilot Study
Combined ..................................... 53
Comparison of Samples .................... 53
Analysis of Combined Sample............... 54
Factor Analysis of BUMPR-Hospital......... 56
Analyses of Fathers for Combined Sample....... 61

4 DISCUSSION .................................... 69

Background Variables Affecting Distress....... 69 Evaluation of the BUMP-R ...................... 77
Evaluation of the Child-BUMP................... 80
Implications .................................. 81

REFERENCES ......................................... 86

APPENDICES

A BEHAVIORAL UPSET IN MEDICAL PATIENTS--REVISED (BUMP-R) ...................................... 94

B BACKGROUND INFORMATION SHEET................... 96

C BEHAVIORAL UPSET IN MEDICAL PATIENTS-CHILD SELF-REPORT VERSION (CHILD-BUMP)............... 97

D DATA COLLECTION PROCEDURES TRAINING GUIDE..... 152

BIOGRAPHICAL SKETCH ................................ 156















LIST OF TABLES


TABLE Page

1 Means and Standard Deviations of the Outcome Measures........................................ 42

2 Demographic Differences for the BUMPR-Hospital.. 44 3 Demographic Differences for the PIC-Depression.. 45 4 Demographic Differences for the Child-BUMP...... 46 5 Demographic Differences for the CDI............. 47

6 Demographic Differences for the Nurse-BUMP...... 48

7 Spearman Correlations Between Illness-Related Variables and Outcome Measures .................. 50

8 Correlations Among Outcome Measures and Demographic Variables. ........................... 52

9 Item-Total Correlations for the BUMPR-Hospital.. 57 10 Eigenvalues for the BUMPR-Hospital Factor
Analysis....................................... 59

11 Factor Structure for the BUMPR-Hospital......... 62 12 Item Loadings for Factor One--Negativity/
Agitation ...................................... 63

13 Item Loadings for Factor Two--Amiability........ 64 14 Item Loadings for Factor Three--Dysphoria....... 65 15 Item Loadings for Factor Four--Noncompliance.... 66














Abstract of Dissertation Presented to the
Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy

MULTIMETHOD ASSESSMENT OF DEPRESSION AND BEHAVIORAL DISTRESS IN A PEDIATRIC POPULATION By

Christina M. Rodriguez

August 1993

Chair: Sheila Eyberg, Ph.D. Cochair: Stephen R. Boggs, Ph.D. Major Department: Clinical and Health Psychology

Growing acceptance of the existence of childhood depression is apparent in recent literature. Despite this scrutiny, research on childhood depression contains mixed results largely because of assessment issues. Studies obtain information from various sources (e.g., parents, children) and utilize different definitions of depression. Assessment of preschool children is further complicated because of skepticism about depression at this age and because of the scarcity of measures appropriate for this age group.

The study of depression in pediatric psychology translates into adjustment in medical settings. Although pediatric populations seem more likely to exhibit depressive symptoms, measures suitable for


vii








these children are scarce. The utility of a new parent report measure, the Behavioral Upset in Medical Patients-Revised (BUMP-R), was examined in a pilot study of 81 mothers, which found increased behavioral distress upon hospitalization. Research on variables influencing adjustment has found that gender, race, socioeconomic status, and family composition are not significant factors. Greater distress may appear with younger children, maternal anxiety, and limited parent contact during hospitalization. Effects of diagnosis, duration of illness, length of hospitalization, and previous hospitalizations are unclear.

The current study explored variables influencing

depressive symptoms in hospitalized children aged 4-12. Parent ratings of child distress were compared to nurse ratings and to children's responses to a pictorial measure designed in this study for hospitalized preschoolers. The day following hospital admission, an assessment battery containing standard measures of depression and measures of distress for hospitalized children was administered to 70 mothers and their children. Thirty-two nurse ratings were also obtained.

Results indicated that demographic and illnessrelated variables were not risk factors for hospital adjustment difficulties. Based on parent ratings, children exhibiting behavioral distress at home may be


viii









more likely to experience adjustment problems upon hospitalization. Concerning assessment, the BUMP-R demonstrated internal consistency and concurrent validity. A factor analysis of mothers from the current study combined with the pilot study revealed four factors identified as Negativity/Agitation, Amiability, Dysphoria, and Noncompliance. The child pictorial measure also demonstrated internal consistency and correspondence with mothers' ratings. Therefore, both measures are promising for use with hospitalized children.














CHAPTER 1
INTRODUCTION

General Childhood Depression Prevalence

After years of controversy, the recent

proliferation of research suggests a growing acceptance that depression exists in children. Several comprehensive reviews of the literature estimate that the diagnosis of major depression in childhood appears in approximately 2% of 7- to 12-year-olds in the general population (see Finch & Saylor, 1984; Kaslow & Racusin, 1990; Kazdin, 1987, 1988 for reviews). However, this estimate increases to 60% in outpatient clinical populations (Kashani, Husain, Shekim, Hodges, Cytryn, & McKnew, 1981). The wide variance in current estimates of the prevalence of childhood depression are primarily due to differences in assessment, definition, and age level (e.g., Angold, 1988; Finch & Saylor, 1984). Prevalence estimates may vary according to age level because depressive symptoms differ between prepubertal children and adolescents (Ryan, Puig-Antich, Ambrosini, Rabinovich, Robinson, Nelson, Iyengar, & Twomey, 1987). In addition, because depression is defined and assessed in a variety of ways (Angold, 1988), different studies use different










criteria for establishing prevalence in a given population.

Assessment Issues

The research on general childhood depression

contains numerous conflicting results. Much of the variability in estimating prevalence and in estimating the influence of variables may be attributed to difficulties in assessment. The definitions of depression determine the selection of assessment techniques (Angold, 1988). Although several studies utilize diagnostic criteria for depression (see Finch & Saylor, 1984 for review), the majority of studies define depression based on scores obtained on a variety of measures. However, many inventories gain only temporary popularity and are utilized in only a single study (Kerr, Holer, & Versi, 1987). Thus, depression is interpreted differently, which complicates questions regarding the influence of numerous variables.

With adults, assessment of depression is often based on self-report or interview data. However, different assessment methods are conducted at different ages. For instance, assessment of adolescents, like adult assessment, often utilizes self-report and interview methods. However, investigation of affective disturbance in preschool children is particularly challenging. A preschool child is not readily able to









3

distinguish emotions (Stone & Lamenek, 1990). Of those socioemotional measures available for assessment of preschool children, self-report measures are rare (Lichtenstein, Dreger, & Cattell, 1986). Consequently, preschool assessment relies heavily on parent report. For school age children, a larger variety of assessment techniques for depression are utilized, including self-report questionnaires, interview rating scales, parent ratings, and observational measures (see examples of observational measures in dental populations, Johnson & Baldwin, 1968 and Wright & Alpern, 1971; in medical populations, see Jay, Ozolins, Elliot, & Caldwell, 1983 and Katz, Kellerman, & Siegel, 1980). Rating scales are completed by a variety of sources, including parents, peers, and teachers.

Given the variability in sources, correspondence among raters becomes a significant concern. A comprehensive review of correspondence issues in measurement of both behavioral and emotional childhood problems found relatively low correspondence among raters (Achenbach, McConaughy, & Howell, 1987). Several other studies have documented low correlations between self-reported depression and parent ratings of depression (e.g., Kazdin, 1989; Kazdin, French, Unis, & Esveldt-Dawson, 1983; see Kaslow & Racusin, 1990; Kazdin, 1987, 1988 for reviews). A few studies have











demonstrated adequate correlations between raters (e.g., in normal children, Leon, Kendall, & Garber, 1980; in pediatric population, Eason, Finch, Brasted, & Saylor, 1985). Despite the general finding of low or poor correspondence, most researchers in the area of depression conclude that multiple sources are required in order to clarify the diagnostic picture (e.g., Kaslow & Racusin, 1990; Kazdin, 1988). In fact, Kazdin, Colbus, & Rodgers (1986) describe the application of discriminant analyses which combines a battery of measures to maximize the classification, i.e., diagnosis, of depression in children. Therefore, in spite of the poor correspondence between raters, current research encourages assessment of childhood depression from multiple sources. Risk Factors

In addition to variability due to assessment methods, a number of variables may affect the prevalence of childhood depression, including gender and age of the child and demographic characteristics, such as race, socioeconomic status, and family composition. A fair amount of research has investigated the effect of sex on depression. In childhood depression, no gender differences emerge for children ages 6-12 (Angold, 1988; Kaslow & Racusin, 1990). The prevalence, however, increases in females










throughout adolescence and approaching adulthood (Kazdin, 1987, 1988). Consequently, the sex ratio of depression for females to males nears 2:1 by adulthood (Boyd & Weissmann, 1981).

Nevertheless, for prepubertal children, several studies utilizing various assessment instruments have corroborated the 1:1 sex ratio. Among these studies, scores on a clinician-administered interview rating scale revealed no sex differences (Shanahan, Zolkowski-Wynne, Coury, Collins, & O'Shea, 1987). In addition, normative data on over 1400 schoolchildren grades 2-8 for one of the most widely-used self-report depression scales, the Children's Depression Inventory (CDI; Kovacs, 1983, 1985), found negligible gender differences, and thus separate norms were not recommended (Finch, Saylor, & Edwards, 1985). This finding was substantiated when no gender differences on the CDI were found in a smaller sample of 166 children grades 3-6 (Reynolds, Anderson, & Bartell, 1985). A study of parent-reported childhood depression also reflects this equal prevalence for children grades 3-6 (Leon et al., 1980; Reynolds et al., 1985). Furthermore, no statistically significant sex differences were found on a measure of peer-nominated depression (Lefkowitz & Tesiny, 1980).










Only a few studies have demonstrated sex differences in depressive symptomatology for prepubertal children. For instance, minor differences were found in which girls scored higher than boys on a different self-report measure (Children's Depression Scale; Reynolds et al., 1985). A study of 8- to 13year-olds also found a trend for girls to report more depressive symptoms than boys on the CDI (Seligman, Peterson, Kaslow, Tanenbaum, Alloy, & Abramson, 1984). These conflicting findings may reflect variability in assessment or in populations, or they may indicate that observers respond differentially based on the child's sex (Saylor, Finch, Spirito, & Bennett, 1984). Despite these mixed results, the bulk of the literature continues to support the 1:1 sex ratio for depression in prepubertal children.

As mentioned above, an interaction between sex and age appears, leading to an increased incidence of depression in adolescent girls (Kazdin, 1987, 1988). A significant interaction effect between gender and age was found in a study of over 1000 schoolchildren, with adolescent girls reporting the most unhappiness in a structured interview (Webb & VanDevere, 1985). Although depressive disorder is relatively uncommon in prepubertal children, the prevalence of depression rises with increasing age (Angold, 1988; Kaslow &











Racusin, 1990). Developmental differences in the manifestation of depression may also lead to different age groups exhibiting different features of depression (Kazdin, 1987). For example, although Ryan et al. (1987) found little variation in depressive symptomatology between children and adolescents, results did identify several features on which the age groups differed (e.g., lethality of suicide attempts, hopelessness, symptoms of anxiety).

In contrast to studies of school-aged and

adolescent children, depression in the lower end of the age range, children under the age of 6, is studied infrequently. Skepticism regarding depressive features in preschoolers prevails despite clinical reports of sadness and suicide attempts in children as young as three years (e.g., Pfeffer & Trad, 1988). Although research using preschool populations has been limited, one study utilizing a multimethod assessment of normal 2- to 7-year-olds found 9 of 109 (8%) children displaying depressive symptoms (Kashani, Holcomb, & Orvaschel, 1986). Therefore, age differences in prevalence may represent measurement variance given that different assessment devices are administered to different age groups (Kazdin, 1988).

Other demographic characteristics have been less well studied. With regard to ethnicity, no clear










racial differences have been found in depressed prepubertal children (Angold, 1988; Kaslow & Racusin, 1990). Although a few studies have suggested that depression may be associated with lower socioeconomic status (see Angold, 1988 for review), depressive symptomatology was not found to be related to father's education or family income (Kandel & Davies, 1982). Finally, no differences in socioeconomic status or family composition (i.e., child living in single-parent home, two-parent home) were found between depressed or nondepressed clinic children (Kaslow, Rehm, Pollack, & Siegel, 1988). The effects of these demographic variables on childhood depression, however, are seldom explored.


Pediatric Populations

Definition Issues

Given the relationship between stressful life

events and depressive symptoms (e.g., Mullins, Siegel, & Hodges, 1985), pediatric psychology has evidenced a growing interest in depressive symptoms in medical populations. Few studies have examined the prevalence of major depression in pediatric populations. In contrast, several studies have focused on adjustment in medical settings. The depressive symptoms associated with poor adjustment in pediatric populations










encompasses a broad range of symptoms, including general subjective distress, dysphoric mood, behavioral manifestations of emotional distress, and anxiety. This wide range of symptoms reveals the difficulties involved due to definitional variability. Many of the symptoms studied are features typically associated with depression, whereas other symptoms are externalizing responses associated with general emotional or psychological distress. Several studies have demonstrated a relationship between externalizing behavior problems and depression (e.g., Leon et al., 1980). Therefore, adjustment problems in pediatric populations are examined in a variety of ways.

Historically, psychologists have been interested in children's adjustment to medical procedures and hospitalization. A child's short-term reaction to medical intervention may vary widely. These depressive reactions can include restlessness, apathy, and sleep and appetite disturbances (Jessner, Blom, & Waldfogel, 1952), panic and crying (Prugh, Staub, Sands, Kirschbaum, & Lenihan, 1953), as well as anger and aggression (Jensen, 1955). Additionally, a large scale follow-up study compared 1000 children hospitalized before age 5 to a nonhospitalized control group (Douglas, 1975). Results indicated that the hospitalized children exhibited numerous difficulties










corresponding to early hospitalization, including conduct problems, delinquency, academic difficulty, and unstable work history. Another study of pediatric surgery patients found that, without intervention, children displayed more behavioral difficulties nearly one month after hospitalization (Melamed & Siegel, 1975). This early research demonstrated that children often adjust poorly to hospitalization or medical procedures.

Prevalence

The study of poor adjustment to hospitalization is susceptible to similar difficulties in estimating prevalence as those difficulties in general childhood depression. Different age groups, definitions, and assessment techniques are utilized. Although one study found nearly 90% of pediatric surgery patients displayed behavioral difficulties following hospitalization (Prugh et al., 1953), most estimates of depressive symptoms in pediatric populations are more conservative. In children with physical handicaps or chronic illness, poor adjustment is reported as ranging from 13-26% based on parent report, approximately two times the rate found in healthy children (Wallander, Varni, Babani, Banis, & Wilcox, 1988). In a review of studies using diagnostic criteria (derived from DSM-III criteria), estimates of depressive symptoms in both








11

inpatient and outpatient pediatric samples ranged from 7-40%, with the wide variability possibly due to different age ranges and different medical populations (Finch & Saylor, 1984). In a study of 7- to 12-yearold hospitalized children, 38% exhibited dysphoric mood (Kashani, Barbero, & Bolander, 1981). An examination of a pediatric psychology service in a children's hospital found 19% of consultations were referred for depression or suicide attempts and 12% of the consultations were referred for adjustment to chronic illness (Olson, Holden, Friedman, Faust, Kenning, & Mason, 1988). Thus, pediatric populations may be more likely to display depressive symptoms than the general child population.

Assessment Issues

As mentioned above, the assessment of pediatric

patients' adjustment involves symptoms from overlapping constructs and may include symptoms of major depression, dysphoric mood, behavioral concomitants of emotional upset, and symptoms related to anxiety. Assessment in pediatric populations shares difficulties similar to the assessment of depression in the general population. Problems with correspondence among raters appear in the pediatric psychology literature as well as in the general childhood depression literature. Rating scales are typically administered to parents,










teachers, and health personnel (see Katz et al., 1980 for example of nurse ratings in pediatric populations).

Compounding the usual assessment issues, the evaluation of emotional distress in pediatric populations is further complicated by the scarcity of measures appropriate for the assessment of these children. The application of standard measures of depressive symptoms, such as the CDI, may not be suitable for hospitalized children. For example, the CDI did not differentiate between a sample of chronic headache sufferers and a non-headache control group (Wisniewski, Naglieri, & Mulick, 1988). Similarly, pediatric inpatients attained CDI scores comparable to gender and age-matched schoolchildren (Saylor, Finch, & McIntosh, 1988). In fact, one study found that normal children tended to report more depression on the CDI than cancer patients, which the researchers suggest reflects denial of symptoms (Worchel, Nolan, Willson, Purser, Copeland, & Pfefferbaum, 1988). In contrast, another study found that children with recurrent abdominal pain scored significantly higher on the CDI than healthy children (Walker & Greene, 1989). These mixed results with the CDI question its utility in pediatric populations.

Likewise, a clinician-administered structured interview, the Children's Depression Rating Scale








13

(CDRS; Poznanski, Cook, & Carroll, 1979), appears to be as limited as the CDI. A study of migraine patients found no difference in CDRS scores compared to a control group of children (Cunningham, McGrath, Ferguson, Humphreys, D'Astous, Latter, Goodman, & Firestone, 1987). A closer examination of the CDRS administered to pediatric cancer patients found significant overlap between depressive symptoms and impairment due to illness (Heilgenstein & Jacobsen, 1988). Thus, the authors found that measures which include somatic symptoms may overestimate the presence of depression.

Instruments which target emotional upset in pediatric populations are few in number. The Observation Scale of Behavioral Distress (Jay et al., 1983) interprets behavioral distress as behaviors indicative of anxiety and pain. Similarly, an observational pain rating scale for children aged 2-6 includes some "depression-like" items (Gauvain-Piquard, Rodary, Rezvani, & Lemerle, 1987). However, both observation rating scales were based on cancer patients and both scales feature an emphasis on pain behavior rather than on depressive symptoms. In contrast, a parent rating scale, the Behavioral Upset in Medical Patients--Revised (BUMP-R; Saylor, Pallmeyer, Finch, Eason, Trieber, & Folger, 1987), focuses on emotional








14

distress and was devised for hospitalized children with a variety of diagnoses.

Risk Factors

Factors which influence the prevalence of

childhood depression in the general population may also affect pediatric populations. With respect to gender differences, results have been mixed. Consistent with findings in the general childhood depression literature, a study of pediatric cancer patients ages 2-20 found no gender differences on an observational measure of behavioral distress during a painful medical procedure (Jay et al., 1983). Moreover, no sex differences were found in hospitalized pediatric patients ages 5-15 on self-reported depression, interview-rated depression, or parent-reported behavioral upset (Saylor et al., 1987). One study of hospitalized children resulted in a greater number of boys diagnosed with depression, although this subsample was small (Kashani, Barbero, & Bolander, 1981). Some studies have found girls demonstrated more behavioral distress during painful medical procedures (e.g., Katz et al., 1980; Melamed & Siegel, 1975). Although the literature suggests that girls may exhibit more distress, evidence for overall gender differences in depressive symptoms for pediatric populations is limited.










The effect of age on the adjustment of pediatric patients contrasts with findings in the general childhood depression literature. In the general population, the prevalence of depressive symptomatology increases through childhood and adolescence. However, in a study conducted with cancer patients, age was the strongest predictor of distress during medical procedures, with younger children exhibiting greater distress (Jay et al., 1983). Another study of a pediatric cancer population aged 1-17 also found younger children expressed more distress during a painful medical procedure (Katz et al., 1980). Additional support for the inverse relationship between age and distress was found when younger hospitalized children were rated by parents as more distressed (Saylor et al., 1987). Furthermore, an interaction between sex and age may occur. In contrast to the trend in general child depression for adolescent females to demonstrate more depressive symptoms, younger females may exhibit more distress in pediatric settings (e.g., Melamed & Siegel, 1975). Overall, younger pediatric patients appear more likely to display behaviors suggestive of depressive symptoms and emotional distress.

Information on the influence of other demographic variables (such as race, socioeconomic status, or










family composition) on adjustment in pediatric populations is even more limited than that found in research on general childhood depression. Of the few studies available, appropriate behavior in a dental setting was associated with socioeconomic status, with upper SES children exhibiting less negative behavior (Wright & Alpern, 1971). Influence of race and family composition are not known and thus may correspond to data obtained on general childhood depression. Consequently, future research should explore the influence of these demographic variables on the emotional adjustment of pediatric patients and whether these influences correspond to those found in general childhood depression.

Several other variables related to pediatric populations may also affect adjustment to hospitalization. These variables include maternal anxiety, parental presence in the hospital, diagnosis, prior medical experience, onset of illness, and length of hospitalization. With regard to maternal anxiety, mothers' self-report of anxiety was positively related to their children's negative behavior during dental visits (Johnson & Baldwin, 1968; Wright & Alpern, 1971). Moreover, mother's state anxiety was positively correlated with a physiological measure of anxiety in pediatric patients (Vardaro, 1978). Maternal anxiety











and depression were significantly higher in children with recurrent abdominal pain than in healthy controls (Walker & Greene, 1989). Another study of pediatric cancer patients demonstrated a significant positive correlation between parental trait anxiety and the child's behavioral distress (Jay et al., 1983). Furthermore, poor parental coping patterns were associated with the depressive symptom of hopelessness in children with cancer (Blotcky, Raczynski, Gurwitch, & Smith, 1985). Thus, a mother's response to her child's hospitalization may affect the child's adjustment.

A concept related to maternal anxiety, whether the parent is present as well as the duration of contact during hospitalization, may also influence the emotional adjustment of the hospitalized child. For example, maternal presence was associated with more negative behavior from children receiving injections (Shaw & Routh, 1982). However, in another study, the amount of time spent with a hospitalized child did not predict depression or behavioral distress (Saylor et al., 1987). Assuming that a factor in poor hospital adjustment lies in separation, Peterson, Mori, and Carter (1985) argue for the importance of encouraging parent contact and enlisting parental assistance during hospitalization. Although parental contact may ease










the child during hospitalization, the conflicting findings suggest that the role of parent contact is either unclear or setting-specific.

The effect of diagnosis on adjustment to

hospitalization has not been studied closely in pediatric populations. The literature on depression in adult medical patients notes differences in prevalence of symptoms as a function of medical diagnosis (e.g., von Ammon Cavanaugh, 1986). For instance, the highest self-reported depression scores were found in patients with gastrointestinal disease, cancer, bone and connective tissue disease, renal disease, and neurological disease. Few studies have investigated this aspect in pediatric samples. One investigation found no differences in behavioral distress due to type of cancer diagnosis (Jay et al., 1983). Another study comparing Crohn's disease, ulcerative colitis, and cystic fibrosis found significantly greater prevalence of depression in Crohn's disease compared to cystic fibrosis (Burke, Meyer, Kocoshis, Orenstein, Chandra, Nord, Sauer, & Cohen, 1989). However, research has not fully examined the differential adjustment of children based on the wide variety of medical diagnoses.

Previous medical experience may also influence the pediatric patient's emotional adjustment to hospitalization. A study of hospitalized children










found that the total number of days spent in previous hospitalizations predict higher parent ratings of their child's emotional distress during the current hospitalization (Saylor et al., 1987). In contrast, Jay et al. (1983) suggest that children may habituate to painful medical procedures because behavioral distress was negatively correlated with number of previous medical procedures. With regard to dental visits, no relationship was found between behavior during the visit and history of unpleasant medical experiences (Johnson & Baldwin, 1968). Thus, these mixed results indicate that the relationship between prior medical experience and emotional adjustment is not well understood.

Onset of illness has also not been extensively

investigated. Time since diagnosis was significantly negatively correlated with behavioral distress during a medical procedure (Jay et al., 1983), further supporting the idea that children habituate to aversive medical procedures. Nevertheless, chronicity of illness as a variable in hospital adjustment has not been studied. A related concept, length of hospital stay, has also not received much research attention. In an adult medical population, length of hospitalization was not related to psychological adjustment (Levenson, Hamer, Silverman, Rossiter,










1986-1987). Both variables, onset of illness and length of hospitalization, warrant further investigation.


Pilot Study

Introduction and Rationale

Given the many unanswered issues regarding

adjustment, a pilot study was conducted to examine the applicability of a new measure in studying variables influencing distress in a hospitalized pediatric population aged 4-12. Few studies have investigated depressive symptomatology in pediatric populations of preschool and pre-literate children. Moreover, measures specifically designed to assess distress in medical settings are limited.

The pilot study focused on the application of a new parent rating scale of behaviors associated with depression and anxiety (Saylor et al., 1987). This rating scale involves the evaluation of the frequency of specific behaviors, not a parental interpretation of emotional distress. Moreover, this brief questionnaire does not require direct observation by a trained clinician. Thus, preliminary findings regarding the influence of background variables on adjustment to hospitalization were gathered.










Methods

Subjects. The sample consisted of 81 mothers of hospitalized children (49 boys, 32 girls) from consecutive pediatric admissions at Shands Hospital in Gainesville, Florida. Ages of the children ranged from

4 years, 1 month to 11 years, 8 months (M = 6 years, 8 months; SD = 1 year, 11 months). The sample of children was 73% White and 27% African-American; 64% of the children lived in two-parent homes and 32% lived in single-parent homes. The sample was predominantly lower to middle SES (95%), based on the five levels of social class as assessed by Hollingshead's (Myers & Bean, 1968) Two Factor Index of Social Position (Class

1 = 3%; Class 2 = 2%; Class 3 = 17%; Class 4 = 40%; Class 5 = 38%).

With respect to illness-related variables, 27% of the children had never been previously hospitalized, 24% had one prior hospitalization, and the remaining 49% had multiple hospitalizations (ranging from 2 to 28), with the number of hospitalizations M = 6.1, SD = 15.8. Duration of illness ranged from diagnosis at birth to diagnosis upon admission to the hospital, with 47% of the patients diagnosed within the past six months.

Medical diagnoses obtained from mothers were

subsequently categorized into illness groups derived








22

from Nelson's Textbook of Pediatrics (Behrman, Vaughan, & Nelson, 1987), with 79% of the sample falling into one of six groups. Twenty percent of those children classified were diagnosed with cardiovascular or respiratory illnesses (e.g., pneumonia, tetralogy of fallot, cystic fibrosis). Fourteen percent were hospitalized for immunity, allergy, or related diseases (e.g., asthma, reflex sympathetic dystrophy). Another 17% were experiencing illnesses interfering with the digestive system (e.g., Hirshsprung's disease, inflammatory bowel disease, appendicitis). Nineteen percent of the patients were diagnosed with urinary problems (e.g., reflux, nephrotic syndrome, kidney infection). An additional 17% were classified with diseases of the nervous system (e.g., spina bifida, Hallervorden spatz, hydrocephalus). Diagnoses had not been identified for another 13% of the children. The seventeen patients not categorized into the above groups had been diagnosed with several different illnesses (e.g., burns, autism, leg fracture) and did not conform with the six illness groups.

Furthermore, descriptions of the clinical course of illnesses were obtained form Nelson's Textbook of Pediatrics (Behrman et al., 1987) to identify diagnoses based on chronicity. Sixty-eight children in the sample were thus classified, with 28% of these










receiving diagnoses for acute illnesses and 72% receiving diagnoses for chronic illnesses. The remaining thirteen diagnoses not categorized on chronicity represent children with ambiguous or unknown illnesses.

In addition to data collected on mothers, eleven

fathers also participated in the pilot study. Three of these fathers completed the measure along with their spouses, whereas the remaining eight fathers were the only parent available for participation in the study.

Children who exhibited mental or physical

limitations which would interfere with parental report of distress were not included in the sample. For instance, children with a developmental delay or hearing impairment would complicate parental response to items regarding conversational ability (e.g., "Refuses to speak") or on items requiring comprehension of requests (e.g., "Accepts advice or instructions easily"). Determination of whether the child met these exclusionary criteria was made through interaction with the child and soliciting the judgment of the parent and/or nurse.

Measure. The Behavioral Upset in Medical

Patients--Revised (BUMP-R; Saylor et al., 1987) is a 56-item parent rating of the child's behavior corresponding to emotional distress at the hospital and








24

at home (See Appendix A). This scale is a revision of the adult version (Zeldow & Braun, 1985) consisting of a 32-item checklist of behaviors that nonpsychiatric patients may exhibit in hospital settings. Patient behaviors were rated by nurses on a Likert scale ranging from 0 to 4. This scale range indicates the frequency of behavior, with 0 representing "never" and

4 representing "always."

Saylor et al. (1987) revised the scale for use with children by having five judges independently evaluate which items were inappropriate for children. Those items deemed inappropriate by a majority of judges were deleted, yielding a 28-item scale. A sample item is:

Looks depressed or sad..................Never
Sometimes

Often

Usually

Always

Scoring of the BUMP-R parallels the original adult version. However, instead of ratings by nurses, parents initially rate the child's behavior in the hospital followed by ratings of the same behaviors at home. Therefore, the BUMP-R provides a parent rating of the child's behavioral upset in the hospital and at home, prior to hospitalization.










Psychometric data are limited due to the recency of the measure's development. For the adult version, internal consistency was reported as .93 and testretest reliability over variable intervals was reported as .66. Four factors were identified from 213 inpatient adults, including behavioral regression, poor patient-staff relationship, depression and anxiety, and passivity and withdrawal (Zeldow & Braun, 1985).

Procedure. Participation was solicited the day following the child's admission to the hospital. Mothers were instructed to complete the BUMP-R in the hospital room, rating their child's behavior since the current hospitalization as well as rating their child's behavior at home.

Results and Discussion

Analyses indicated that age, SES, and the BUMP-R scores were normally distributed. However, the number of previous hospitalizations and the duration of illness were not normally distributed. Consequently, analyses utilizing these two illness variables are based on Spearman rho correlation coefficients.

Whereas no sex differences were found in the

BUMPR-Home rating of distress, a trend for girls (M = 29.8, SD = 11.4) to exhibit more distress in the hospital than boys (MI = 24.9, SD = 13.7) emerged (t =

1.67, R = .09). As expected, no race or family








26

composition differences were found. There were also no significant differences in the BUMPR-Hospital ratings due to diagnosis based on the six categories of illness (F(63) = 1.12, R > .05). Only marginal differences for chronicity of illness (t(66) = 1.73, R = .09) were found on the BUMPR-Hospital, with children diagnosed with acute illnesses (M = 32.1, SD = 14.4) scoring higher than those diagnosed with chronic illnesses (M = 26.3, SD = 11.6).

Age was significantly correlated with hospital distress (r = -.24, p < .05), with younger children obtaining higher distress ratings than older children. Behavioral distress in the hospital or at home was not significantly related to SES, duration of illness, or number of previous hospitalizations.

Ratings of hospital distress were significantly

associated with home distress (r = .48, R < .0001). In addition, maternal ratings of their child's behavioral upset in the hospital (M = 26.8, SD = 13.0) were significantly higher (t(79) = 2.67, p < .01) than ratings of distress behaviors at home (M = 23.1, SD = 11.3), suggesting that mothers observed increased behavioral distress following hospitalization.










Purpose of Study and Hypotheses

The purpose of the current investigation was twofold. First, variables affecting behavioral upset in a hospitalized population aged 4-12 were further explored. Second, issues regarding assessment modalities were studied. Specifically, differences among raters were investigated as well as a comparison of measures designed to assess distress in medical populations with measures more commonly used in the assessment of depression. As part of the assessment battery, a newly designed self-report scale appropriate for pre-literate children in hospital settings was included. The usefulness of this new measure for hospitalized children was of particular interest.

Assessment of depression was obtained following hospital admission via self-reported behavioral distress and depression, nurse report of the child's behavioral distress, and parental report of their child's depression and behavioral distress as well as the parent's own anxiety in the hospital setting. No significant differences in the outcome measures due to sex, race, socioeconomic status, or family composition (i.e., single v. two-parent homes) were anticipated. The following hypotheses were tested.

(1) Age and grade differences were expected, with younger children exhibiting greater emotional distress.








28

(2) Differences in depression and behavioral upset were hypothesized for different diagnostic groups.

(3) Onset of illness, number of prior

hospitalizations, and length of current hospital stay were expected to be positively related to behavioral upset and depression.

(4) Number of hours spent with parent was expected to be negatively correlated with adjustment to hospitalization.

(5) All outcome measures, including parental anxiety, were hypothesized to be intercorrelated.

(6) In addition, behavioral distress was expected to be rated higher in the hospital setting as compared to home, indicating distress at hospitalization.














CHAPTER 2
METHODS

Subjects

Seventy mothers and their children were recruited from consecutive pediatric admissions at Shands Teaching Hospital. The sample of children consisted of 29 males and 41 females between the ages of 4 years, 2 months and 12 years, 11 months (M = 8 years, 6 months, SD = 2 years, 8 months). Racial composition was 66% White, 27% African-American, and 7% Other. With regard to school grade level, there were 10% of the children not attending school, 4% in preschool programs, 12% in kindergarten, 16% in the first grade, 11% in the second grade, 10% in the third grade, 10% in the fourth grade, 13% in the fifth grade, 10% in the sixth grade, and 4% in the seventh grade.

The family composition of the sample consisted of 66% of the children living in two-parent homes and 34% of the children living in single-parent homes. Although primarily lower to middle SES (90%), the sample included all five levels of social class as measured by Hollingshead's (Myers & Bean, 1968) Two-Factor Index of Social Position (Class 1 = 3%; Class 2 = 7%; Class 3 = 16%; Class 4 = 37%; Class 5 = 37%).

29








30

Based on diagnoses reported by mothers, 77% of the children were divided into six diagnostic groups based on Nelson's Textbook of Pediatrics (Behrman et al., 1987). Of those diagnoses categorized, 19% of the children were hospitalized for cardiovascular or respiratory difficulties (e.g., cystic fibrosis, cardiac myopathy, coronary artery disease). Twenty-two percent of the children were diagnosed with immunity, allergy, or related diseases (e.g., HIV, asthma, juvenile rheumatoid arthritis) and 20% percent of the children were experiencing difficulties involving the digestive system (e.g., cleft lip and palate, alpha-antitrypsin deficiency, gastroenteritis). An additional 13% of the children were diagnosed with urinary difficulties (e.g., nephrotic syndrome, urinary tract infection, bladder infection). Thirteen percent of the patients had been hospitalized for diseases affecting the nervous system (e.g., spina bifida, seizure disorders). For another 13% of the children, diagnosis had not been determined (e.g., fever of unknown origin). The sixteen children not classified into diagnostic groups had received a variety of diagnoses, preventing assignment to a group of adequate number for analyses. These unclassified diagnoses included viral meningitis, diabetes, snakebite, and sickle cell anemia.








31
In addition, 62 of the diagnoses were categorized on chronicity based on Nelson's Textbook of Pediatrics (Behrman et al., 1987) description of pathogenesis. Thirty-five percent of diagnoses were described as acute illnesses and 65% of diagnoses were identified as chronic illnesses. The eight children not classified on chronicity had undiagnosed illnesses.

Forty percent of the sample of children had never been previously hospitalized, 12% of the children had been hospitalized once before, whereas 43% had been hospitalized on multiple occasions (ranging from 2 to 20 prior hospitalizations), with prior hospitalizations M = 3.2, SD = 5.1. The time spent hospitalized during the study ranged from 1 day to 29 days (M = 5.0, SD =

4.8) with 54% hospitalized for three days or less. The duration of illness associated with their diagnosis ranged from newly diagnosed to diagnosis at birth, with 41% diagnosed within one month, and an additional 20% within six months. The average number of hours the parent spent with the child in the hospital in a 24 hour period ranged from 3 hours to 24 hours a day, with 67% of parents reporting they spent 24 hours a day with their child and 17% spending between 18 and 23 hours with their child.

Thirty-two of the children involved in the study also received behavior ratings from their nurses. In










addition to the participation of mothers, thirteen fathers were also involved in the study. Five of the fathers were spouses of mothers who also participated in the study whereas eight fathers were the only parental respondents involved in the study.

Children who manifested mental or physical

deficits which would impede responses on the child self-report and parent measures were excluded (e.g., profound mental retardation, lack of language development, significant vision or hearing impairment). Comprehension problems or the inability to see the pictures, for example, would interfere with the child's responses on the Child-BUMP. The evaluation of whether a child met these exclusionary criteria was made through interaction with the child and from parent and/or nurse judgments. Informed parental consent and child verbal assent were independently obtained. Basic demographic information and information regarding illness was collected on a summary sheet (See Appendix B).


Measures

The Behavioral Upset in Medical Patients--Revised (BUMP-R; Saylor et al., 1987). As detailed in the description of the pilot study (p.23-25), the BUMP-R is a 56-item parent rating of the child's behavior








33

corresponding to emotional distress at the hospital and at home (See Appendix A). Patient behaviors are rated by parents on a Likert scale ranging from 0 to 4. In addition, the child's nurse completed the BUMP-R questions for behavior observed in the hospital.

The Personality Inventory for Children, Depression Scale (PIC-D; Lachar & Gdowski, 1979; Wirt, Lachar, Klinedinst, & Seat, 1977). The PIC-D is one of 16 profile scales of the PIC, a comprehensive 600-item, true-false parent rating measure of psychological adjustment for children ages 3-16.

The Depression scale contains 46 rationallyderived items. A sample item is:

My child is usually in good spirits .........True False

These items yield ten interpretable factors, with Brooding/Moodiness and Social Isolation accounting for 56% of the variance. The eight other factors are Crying Spells, Lack of Energy, Pessimism/Anhedonia, Concern with Death and Separation, Serious Attitude, Sensitivity to Criticism, Indecisiveness/Poor Self-Concept, and Uncommunicativeness. A Total Depression scale score is obtained.

With respect to validity, the PIC-D correlated significantly with all scales on the Conners Parent Questionnaire except the Antisocial Scale and










correlated with the Children's Depression Inventory (Leon et al., 1980). Moreover, the Depression scale correlated with social withdrawal, depression, uncommunicativeness, and social subscales of the Child Behavior Checklist (Kelly, 1982). Test-retest reliability was reported as ranging from .80 to .94 and internal consistency (coefficient alpha) of .86 (Wirt et al., 1984).

The State Trait Anxiety Inventory (STAI, Form Y) (Spielberger, 1983). The STAI consists of two self-report scales measuring state anxiety and trait anxiety. The form used in this study is a revision of an earlier version (Spielberger, Gorsuch, & Lushene, 1970). In contrast to the earlier version which had items related to depression, Form Y measures feelings of anxiety discriminated from symptoms of depression. The State Anxiety scale consists of 20 statements about how the subject feels "right now, at this moment." The Trait Anxiety scale consists of 20 statements about how the subject generally feels. Subjects are asked to rate the intensity of feelings on a Likert scale ranging from (1) not at all, (2) somewhat, (3) moderately so, (4) almost always. The following sample item is from the Trait Anxiety scale:










I feel nervous and restless........Not at all Somewhat

Moderately so

Almost always

The following sample item is from the State Anxiety Scale:

I feel strained ....................Not at all

Somewhat

Moderately so

Almost always

Test-retest correlations for the Trait Anxiety scale range from .73 to .86 for college students. Stability on the State Anxiety scale, ranging from .16 to .62, reflects that state anxiety is influenced by situational factors (Spielberger, 1983). With regard to internal consistency, the median alpha coefficient was reported as .93 for State Anxiety and .90 for Trait Anxiety (Spielberger, 1983). The Trait Anxiety scale has demonstrated concurrent validity with such measures as the Taylor Manifest Anxiety Scale (Taylor, 1953) and the scores on the state anxiety scale have been correlated with situational stress (see Spielberger, 1983 for review).

Parents were instructed to complete the State

Anxiety items for how they felt at that moment, during the evaluation. Then they were asked to complete the










Trait Anxiety items for their general feeling since their child was hospitalized.

The Children's Depression Inventory (CDI) (Kovacs, 1983, 1985; Finch et al., 1985). The CDI is a 27-item self-report depression scale appropriate for children ages 8 and older. Each item consists of three statements representing graded levels of severity of a depressive symptom and the child selects one of the three statements. A sample item is:

I feel like crying everyday.

I feel crying many days.

I feel like crying once in a while.

The choices are valued from 0 to 2, with high scores indicating depression. The internal consistency of the CDI (coefficient alphas) was reported as ranging from .70 to .94 (Kovacs, 1985; Saylor et al., 1984). Test-retest reliability ranged from correlations of .38 to .87 (Saylor et al., 1984).

The Behavioral Upset in Medical Patients-Child

Self-Report Version (Child-BUMP). The Child-BUMP is a 27-item pictorial scale designed for use in this study (See Appendix C). All questions from the BUMP-R for the hospital setting (with one exception) were rephrased in language understandable for children ages 4-12. The questionnaire was rephrased by three independent sources and the simplest items selected.










The scale was modeled according to the Harter &

Pike (1983) Pictorial Scale of Perceived Competence and Social Acceptance for Young Children. In this scale, children are asked to select from two items with corresponding pictures which is most like them. Then, for each item, two possible questions are asked regarding the frequency with which they exhibit the behavior. This two-step process facilitates attention and simplifies the items for young children. A sample item from the Child-BUMP is as follows:

This boy cries a lot. This boy cries a little.

(If child selects first picture,

ask next two questions)

DO YOU:

always cry OR sometimes cry

(If child selects second picture,

ask next two questions)

DO YOU:

hardly ever cry OR never cry

Each final response is scored on a scale of one to four, with high scores suggestive of greater behavioral distress. A total score was obtained by summing the item scores. Fourteen of the items begin with the most negative response (4) and thirteen of the items depict African-American children. Male and female versions of








38
the scale are identical except for the sex of the child in the picture and as phrased in the question.


Procedure

Parents of eligible patients were informed of the study and participation was solicited by trained undergraduate research assistants supervised by the experimenter. Research assistants underwent a training program involving several steps. Assistants were first given a written description of general data collection procedures (See Appendix D) that they rehearsed prior to attending a 2-hour training session. During the training session, the assistants watched a videotape simulation of the experimenter administering the measures to two children, a 5-year-old male and an 11-year-old female. Following the videotape, the assistants role-played and discussed solving potential difficulties that could be encountered during data collection. After the training session, each assistant observed the experimenter administering the measures to at least one subject. Finally, the experimenter observed each research assistant giving the measures until the assistant demonstrated proficiency in the data collection procedures.

Families were administered the questionnaires in the hospital rooms the day following the child's










hospital admission. This delay allowed approximately 24 hours of behavior for both the parents and children to assess reaction to hospitalization (time since hospitalization ranged from 17 to 35 hours). Following the collection of demographic information, parents were instructed to complete the BUMP-R and PIC-D based on their child's behavior and the STAI on their own feelings since their child's hospitalization. As the parents completed their forms, the research assistant read aloud the Child-BUMP to all children and read the CDI to children ages 8-12. To reward cooperation, the child was allowed to select a small prize (e.g., sticker, toy car, puzzle book). After the data was collected from the parent and child, the nurse completed the BUMPR-Hospital form provided that the nurse had interacted with the child for a minimum of eight hours.














CHAPTER 3
RESULTS

Initially, this section focuses on the results

obtained from the examination and analysis of the data collected from the 70 mothers, their children, and the nurses involved in the current study. Differences in demographic and illness-related variables were examined as well as the relationships among the test scores. Following this presentation, the information gathered from mothers in this study were compared to the data collected from the 81 mothers who participated in the pilot study to highlight significant differences between studies. Data from the mothers in the current study and the pilot study were then combined and analyzed to explore differences in demographic and illness-related variables on the BUMP-R for this larger, combined sample. In addition, a factor analysis of the BUMPR-Hospital ratings was performed for this combined sample.

Finally, analyses on a sample of fathers was also conducted. This sample consisted of 13 fathers from the current study and 11 fathers from the pilot study. A matched sample of mothers was obtained for comparison, in which eight of the fathers were compared 40










to their spouses who had also participated. For the remaining 16 fathers, mothers were selected for comparison based on the child's age and sex. Although this matched sample cannot address correspondence of ratings between mothers and fathers, the comparison addresses whether fathers rate distress in their children similar to mothers' ratings.


Analyses for Current Study Descriptive Results & Statistics

Scores used for all analyses were the raw scores of the BUMP-R, Child-BUMP, and CDI and the standard TScores of the PIC-D and STAI. Inspection of skewness, kurtosis, and histograms indicated that all of the outcome measures were normally distributed. However, the length of current hospitalization, the number of prior hospitalizations, the duration of illness, and the number of hours the parent spent with the child were not normally distributed. Therefore, analyses utilizing these variables were based on Spearman rho coefficients. Moreover, given the number of correlations computed, the significance level was reduced to alpha=.01 for correlational analyses in order to control for Type I errors.

Descriptive statistics on each test variable are displayed in Table 1. No normative data is available










Table 1: Means and Standard Deviations of the Outcome Measures

N Mean (SD)


BUMPR-Hospital 70 27.2 (13.3) BUMPR-Home 70 24.7 (11.2) PIC-D 70 60.9 (15.0) STAI-State 70 55.3 (12.2) STAI-Trait 65 52.2 (10.4) Child-BUMP 70 51.5 (8.4) CDI 33 6.9 (5.1) Nurse-BUMP 32 28.6 (18.8)



for the BUMP-R scores or for the new Child-BUMP. The sample mean on the PIC-D (M = 60.9, SD = 15.0) was one standard deviation above the normative sample mean, suggesting that mothers reported more depressive symptomatology in this population than in the general population. However, sample means on parent report of anxiety on both the STAI State scale (M = 55.3, SD = 12.2) and the STAI Trait scale (M = 52.2, SD = 10.4) were within normal limits. In addition, the current sample mean on the CDI (M = 6.9, SD = 5.1) is comparable to the mean reported for newly diagnosed diabetics (Kovacs, 1983) and is below the recommended cut-off for diagnosis of depression.









43

The new Child-BUMP measure appeared to function as designed, with the children readily grasping the twostep response process. Four-year-old children occasionally had difficulty with the measure if the parent had suggested that the child had comprehension problems. Repeating items was often helpful for children, particularly for those with more limited attention spans. Cronbach's coefficient alpha for the Child-BUMP was .76, suggesting fairly strong internal consistency. For the Nurse-BUMP, internal consistency was .93, indicating that the individual items are very strongly intercorrelated.

Analyses of Background Data

An examination of the demographic variables sex, age group, race, and family composition indicated no significant differences for the outcome measures (See Tables 2-6). No differences had been anticipated based on sex, race, or family composition. However, a marginal relationship was found between the CDI scores and family composition (See Table 5), with children of single parent homes reporting more depressive symptoms than children in two-parent homes (t = 1.99, p = .055). Contrary to the expectation that younger children (ages 4-8) would exhibit more distress than older children (ages 8-12), no differences between age groups were found on any of the outcome measures. Analyses also









44

Table 2: Demographic Differences for the BUMPR-Hospital N Mean (SD) t


Sex


Males Females


28.3 (14.2) 26.4 (12.7)


.60


Age Group


4-8

8-12


Race


White

African-American


29.0 (15.5) 25.7 (11.0)





28.1 (13.0) 26.7 (14.9)


25.4 (12.8) 30.6 (13.5)


Note: All t values are nonsignificant.


1.01


.36


Family Composition

2-Parent 1-Parent


1.55









45

Table 3: Demographic Differences for the PIC-Depression N Mean (SD) t


Sex


Males Females


62.1 (13.9) 60.0 (15.9)


.59


Age Group


63.2 (18.2) 58.9 (11.6)




60.8 (16.2) 60.6 (13.2)


Race


White

African-American


Family Composition


58.3 (12.8) 64.9 (18.7)


Note: All t values are nonsignificant.


4-8

8-12


1.20


.04


2-Parent 1-Parent


1.70











Table 4: Demographic Differences for the Child-BUMP N Mean (SD) t


Sex


Males Females


52.0 (8.8) 51.0 (8.7)


.48


Age Group


4-8

8-12


Race


White

African-American


50.7 (10.2)

52.0 (7.3)




52.6 (8.6) 48.5 (8.5)


51.5 (8.3)

51.4 (9.1)


Note: All t values are nonsignificant.


.65


1.75


Family Composition

2-Parent 1-Parent


.05










Table 5: Demographic Differences for the CDI


N Mean (SD) t


Sex

Males 12 8.3 (5.0) 1.12

Females 21 6.2 (5.2)


Race

White 21 7.0 (5.2) .28 African-American 10 6.4 (5.2)


Family Composition

2-Parent 21 5.4 (4.3) 1.99*

1-Parent 11 8.8 (5.1)


Note: All t values are nonsignificant.
* = .055
p = .055










Table 6: Demographic Differences for the Nurse-BUMP N Mean (SD) t


Sex


Males Females


31.3 (19.6) 25.5 (18.1)


.86


Age Group


4-8 8-12


Race


White

African-American


32.2 (17.8) 25.0 (19.6)




26.8 (17.8) 38.2 (24.7)


Family Composition


25.5 (17.6) 32.4 (20.5)


Note: All t values are nonsignificant.


1.08


1.23


2-Parent 1-Parent


.97








49

revealed no significant interaction between sex and age group or between sex and race.

An examination of the Spearman coefficients (See Table 7) indicated that the length of the current hospital visit and the number of prior hospitalizations were not associated with any of the measures of behavioral distress. Duration of illness was significantly negatively correlated with only parental state anxiety. Moreover, the relationship between the Child-BUMP scores and hours spent with the parent was positively correlated, a direction opposite of that predicted. Thus, the more time a parent spent with the child, the more distress the child reported on the pictorial measure. Although no other significant relationships with time spent with the parent emerged, a modest negative correlation with parental state anxiety was noted.

With regard to the six diagnostic groups, there were no significant differences on the BUMPR-Hospital scores (F(53) = .87, R > .05), on the CDI (F(26) =

1.92, p >.05), or on the Nurse-BUMP (F(24) = 1.68, p > .05). Only a marginal difference in diagnostic groups was found with the PIC-D scores (F(53) = 2.11, R = .08) and with the Child-BUMP scores (F(53) = 2.13, p = .08). Based on the classification of chronicity of illness, no significant differences appeared on any of the








Table 7: Spearman Correlations Between Illness-Related Variables and Outcome Measures


Length of Number of Prior Duration of Number of Hours hospitalization hospitalizations illness spent w/ parent

rBUMPR-Hospital .14 (70) .04 (7) .02 (6) .16 (70) BUMPR-Homespital .13 (70) .0421 (70) .0214 (64) .1605 (70)
BUMPR-Home .13 (70) .21 (70) .14 (64) .05 (70) PlC-Depression .04 (70) .11 (70) .00 (64) .05 (70) Child-BUMP -.14 (70) -.04 (70) -.17 (64) .29 (70)* CDI .00 (33) .09 (33) -.08 (31) .01 (33) Nurse-BUMP -.15 (32) -.15 (32) -.29 (30) .04 (32) STAI-State .20 (70) -.15 (70) -.30 (64)* -.26 (70)a STAI-Trait .03 (65) .09 (65) -.17 (59) -.13 (65)


R .01

a Because the significance level was reduced to .01, the marginal relationship was found at


only R < .05.









51

outcome measures. In contrast to the pilot results and the hypothesis, there was no significant difference between the BUMPR-Hospital and BUMPR-Home score (t(70) = 1.73, p > .05).

Correlational Analyses

The correlation matrix in Table 8 indicates that, as expected, many of the outcome variables were intercorrelated. Consistent with the pilot study, the BUMPR-Hospital ratings were significantly positively correlated with the BUMPR-Home ratings. In addition, the BUMPR-Hospital scores were also significantly positively associated with the PIC-D, Child-BUMP, and CDI. However, the parent ratings of hospital distress on the BUMP-R was only marginally positively correlated with the Nurse-BUMP. The parent report of distress on the BUMPR-Home form was significantly positively correlated with the PIC-D and CDI. Moreover, the parent ratings on the PIC-D were significantly positively correlated with the CDI and parental trait anxiety. Parental trait and state anxiety were also positively correlated. With regard to demographic variables, SES was significantly positively correlated with the BUMPR-Home rating and the PIC-D scores.

The BUMPR-Hospital ratings and the Nurse-BUMP

ratings were not significantly correlated with parent report of state or trait anxiety. In addition, the










Table 8: Correlations Among Outcome Measures and Demographic Variables


BUMPR-Hosp BUMPR-Home PIC-D Child-BUMP CDI Nurse-BUMP STAI-State STAI-Trait



BUMPR-Home .53 (70)

PIC-D .45 (70)" .49 (70)

Child-BUMP .36 (70)" .10 (70) .14 (70) CDI .43 (33)' .52 (33)" .54 (33)" .30 (33) Nurse-BUMP .39 (32)' .02 (32) .04 (32) .32 (32) -.17 (14) STAI-State .15 (70) .07 (70) .22 (70) -.05 (70) -.03 (33) -.16 (32) STAI-Trait .14 (65) .21 (65) .40 (65)- -.02 (65) .26 (31) -.19 (29) .64 (65)"

AGE -.15 (70) .04 (70) -.17 (70) .04 (70) .36 (33)' -.23 (32) -.03 (70) -.08 (70)

GRADE -.20 (70) -.04 (70) -.22 (70) -.01 (70) .30 (33) -.25 (32) -.02 (70) -.07 (65)

SES -.02 (70) .31 (70)" .36 (70)" .16 (70) .27 (33) .29 (32) .00 (70) .23 (65)



2 & .01

2 < .001

' Because the significance level was reduced to .01, this marginal relationship was found at only 2 < .05.
N)j










parent report on the BUMPR-Home form was not significantly correlated with the Child-BUMP, NurseBUMP, or parental state or trait anxiety. Similarly, the PIC-D was not significantly related with the ChildBUMP, Nurse-BUMP, or parental state anxiety. The Child-BUMP was also not significantly associated with the CDI, Nurse-BUMP, or parental state or trait anxiety. The CDI was not significantly correlated with the Nurse-BUMP or parental state or trait anxiety. Moreover, contrary to the hypothesis that younger children would demonstrate more distress, age and grade were not significantly correlated with the outcome variables. This finding does not replicate the modest correlation of age and BUMPR-Hospital scores revealed in the pilot study. Lastly, SES was not significantly related to scores on the BUMPR-Hospital, Child-BUMP, CDI, Nurse-BUMP, or parental state or trait anxiety.


Analyses of Current Study and
Pilot Study Combined

Comparison of Samples

Based on the results obtained from the pilot

study, a difference had been anticipated in the current study between maternal ratings of distress at hospitalization and maternal ratings of distress at home. Moreover, the effects of age and gender were not observed in the current study compared to the pilot.










Given that these findings were not replicated in the present study, a comparison of the two samples of subjects was performed.

Distribution of race, family composition, SES,

duration of illness, number of prior hospitalizations, and diagnostic grouping were comparable between samples. However, the pilot study was composed of predominantly males whereas the current study consisted of primarily females, a significant gender difference (X2 = 5.46, R = .01). Furthermore, the children in the pilot study were significantly younger than the children in the current study (F(150) = 23.86, R < .0001).

Analyses of Combined Sample

The data obtained from maternal ratings on the BUMP-R from both samples were combined and analyzed, resulting in a more evenly distributed sample with respect to sex and age. This combined sample consisted of 151 mothers of 78 boys and 73 girls. Ages of the children ranged from 4 years, 1 month to 12 years, 11 months (M = 7 years, 6 months, SD = 2 years, 5 months).

With this combined sample, no significant sex differences were found in the mothers' ratings of hospital distress (t(149) = .80, p > .05) or in the ratings of home distress (t(149) = .70, p > .05). Moreover, children aged 4-8 years obtained ratings










comparable to the children aged 8-12 years on the BUMPR-Hospital (t(149) = .91, p > .05) and on the BUMPR-Home (t(149) = 1.19, p > .05). Age evidenced only a small relationship (. = -.17, p < .05) with ratings of behavioral distress in the hospital.

There were no differences in the BUMPR-Hospital scores (F(117) = 1.55, p > .05) or in the BUMPR-Home scores (E(117) = 1.45, p > .05) for the six diagnostic groups. The combined sample demonstrated no differences for chronicity of illness on the ratings of hospital distress (t(128) = .78, p > .05) or the ratings of home distress (t(128) = .29, p > .05). No significant relationship was found between the number of prior hospitalizations and the BUMPR-Hospital scores (1 = .06, p > .05) or the BUMPR-Home scores (r = .06, p > .05). Similarly, duration of illness was unrelated to the BUMPR-Hospital ratings (r = .06, p > .05) or the BUMPR-Home ratings (r = .05, p > .05).

With this combined sample, maternal ratings of distress following hospitalization (M = 27.0, SD = 13.1) were significantly higher (t(151) = 3.15, p < .01) than ratings of distress at home (M = 23.8, SD = 11.2). Consistent with the results of both the pilot study and the current study, ratings of distress in the hospital were significantly correlated with ratings of distress in the home (r = .50, R < .0001).










Factor Analysis of BUMPR-Hospital

The BUMPR-Hospital scores obtained from the 151 mothers participating in the pilot study and the current study were then further analyzed in order to explore the psychometric characteristics of this measure. Because the BUMPR-Home items were designed to serve only as comparisons to the hospital items, the BUMPR-Home scores alone would not address hospitalization adjustment and thus were not further analyzed. Correlations between individual items on the BUMPR-Hospital rating and the total score are reported in Table 9. Cronbach's coefficient alpha assessing internal consistency of the total BUMPR-Hospital measure was .87, suggesting that individual items are strongly intercorrelated.

Factor analytic procedures were then performed on the BUMPR-Hospital scores. Principal factors were derived from the correlation matrix and the diagonal was replaced by communality estimates. For these communality estimates, Gorsuch (1983) maintains that, instead of unities for the diagnonals, a more accurate estimation of the communality for a given variable utilizes the squared multiple correlation with all other variables. To obtain the optimal factor solution, factor rotation was performed using the varimax orthogonal transformation, the promax oblique











Table 9:


Item 1 Item 2 Item 3 Item 4 Item 5 Item 6 Item 7 Item 8 Item 9 Item 10 Item 11 Item 12 Item 13 Item 14


Item-Total Correlations


.416 .484 .466 .568 .562

.453 .290 .398 .113

.242 .385

.435 .511 .638


57

for the BUMPR-HosDital


Item 15 Item 16 Item 17 Item 18 Item 19 Item 20 Item 21 Item 22 Item 23 Item 24 Item 25 Item 26 Item 27 Item 28


.332 .563 .328

.491 .333

.343 .109 .518

.486 .269 .376 .516 .301

.346


Item-Total Corlain


rp;%hl a Q !










transformation, as well as the quartimax orthogonal transformation. Comparison of the rotation methods revealed that the promax oblique rotation yielded factor structures of substantial factor complexity and thus complicated interpretability. Thus, orthogonal solutions were preferred over the promax oblique solution which assumes intercorrelated factors. The quartimax orthogonal rotation was selected in favor of the varimax orthogonal rotation because Gorsuch (1983) advises that the varimax method is inappropriate when the measure has high internal consistency. The quartimax rotation assumes that a single general factor accounts for a substantial amount of the variance, which is implied with high internal consistency, whereas the varimax rotation does not assume one large factor.

The number of factors extracted was based on three criteria. As Gorsuch (1983) described, Guttman recommended examination of the characteristic roots (eigenvalues) for factor extraction. The number of eigenvalues greater than one estimates the number of factors. Based on this criterion, four factors were extracted (See Table 10), which accounted for 85% of the variance.










Table 10: Eigcenvalues for the BUMPR-Hospital Factor Analysis

Eigenvalues


Factor 1 6.07

Factor 2 2.50

Factor 3 1.78

Factor 4 1.09



The next criteria involved a plot of the

eigenvalues, known as the scree test. Gorsuch (1983) describes that the predominant factors will account for a significant portion of the variance. Thus, the number of factors can be extracted where the plot of the eigenvalues levels off. The scree test shown in Figure 1 supports extraction of four factors.

Lastly, interpretability was used as the final

criterion. Comparisons to three factor and five factor solutions confirmed that the quartimax four-factor solution retained the greatest number of items and resulted in the most parsimonious factor structure. Moreover, the quartimax four factor rotation yielded the most readily interpretable solution.

Item loadings of .40 were considered significant, which is more conservative than the traditional .30 cutoff and would result in stronger factors. Only Item














Figure 1: Scree Test of Eigenvalues for Factor Analysis of


BUMP-R Hospital Ratings


2


2+

3



4
1+
5
6
7
89 01 23 4
0 + 5 67 89 01 23 45 67 8
--------------------*--- -+----+----+----+--------+----4-----4....4
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28

Number










14 (Is uncooperative) loaded on more than one factor, and two items did not have significant loadings on any of the four factors (Item 6: Refuses to speak; Item 7: Says he or she feels blue or depressed; See Table 11 for factor structure). Using the .40 item loading criterion, the first factor was extracted with 11 items, identified as Negativity/Agitation (See Table 12), with an internal consistency coefficient of .86. This first factor accounted for 41.2% of the variance. The second factor of nine items, labelled Amiability (See Table 13), encompassed sociability and agreeability. This factor attained an internal consistency coefficient of .79 and accounted for 26.3% of the variance. The third factor with four items was named Dysphoria (See Table 14), including vegetative symptoms of depression and emotional discomfort, with internal consistency at .68. The third factor accounted for 19.9% of the variance. The final factor was identified as Noncompliance (See Table 15), with four items and an internal consistency of .68. The last factor accounted for 12.5% of the variance.

Analysis of Fathers for Combined Sample

To examine similarities between parental raters, fathers were compared to a matched sample of mothers. Results indicated that fathers' ratings (M = 30.7, SD = 15.6) were similar to mothers' ratings (M = 28.3, SD =
































































I

I
I

I


Table 11: Factor Structure of the

FACTOR 1 FACTOR 2 Item 1 .63589 -.04183 Item 2 .45435 .12336 Item 3 .59320 .02689 Item 4 .65127 .13778 Item 5 .65284 .12495 Item 6 .30260 .37821 Item 7 .18094 .18013 Item 8 .22155 .17271 Item 9 -.06193 -.00589 Item 10 .32859 -.03857 Item 11 .19292 .08428 Item 12 .16359 .22701 Item 13 .53006 .07606 Item 14 .57713 .26186 Item 15 .51382 .06229 Item 16 .70806 .08321 Item 17 .12274 .07588 Item 18 .72155 .03593 Item 19 .55057 -.10472 Item 20 .21628 .10644 Item 21 -.14432 .39522 Item 22 .25396 .67377 Item 23 .28992 .64513 Item 24 -.03104 .53951 tem 25 .04713 .56132 tem 26 .24431 .63897 tem 27 .03600 .54897 tem 28 .23332 .48658


BUMPR-HospitRl


FACTOR 3

-.00338

.36104 .18392

.22307 .17103

.25504 .28743

.10444 .43279

-.10527 .67505 .70590 .30855 .08055

-.02597

-.01942 .32630

-.16548

-.12761

.45006 .02132

.14469

-.13552 .20288 .30167 .12487 .01670

-.27860


FACTOR 4 .02754 .00857

-.04462

-.05709

-.02268

-.06537 .01944 .66649 .14313 .45162

-.02277

-.03168 .02363 .39853

-.14439 .19862 .48533 .20505 .29440 .00852 .13843

-.12223 .15905

-.09263

-.11305 .01719 .08005 .27999


BUMPR-Hosnital










Table 12: Item Loadings for Factor One--Negativity/ Aqitation

Item Loading


Item 18 Demanding (.72) Item 16 Stubborn, negativistic (.71) Item 4 Becomes upset easily (.65) Item 5 Is irritable or grouchy (.65) Item 1 Is impatient (.64) Item 3 Gets angry (.59) Item 14 Is uncooperative (.58) Item 19 Manipulative (.55) Item 13 Clinging, needs lots of reassurance (.53) Item 15 Complains (.51) Item 2 Cries (.45)


Consistency Factor 1: Coefficient alpha = .86


Internal










Table 13: Item Loadings for Factor Two--Amiability

Item Loading


Item 22 Tries to be friendly (.67) Item 23 Accepts advice or instructions easily (.64) Item 26 Pleasant to be with (.64) Item 25 Laughs or smiles at funny comments

or events (.56) Item 27 Shows interest in recovery (takes

initiative) (.55) Item 24 Starts conversation (.54) Item 28 Does what he or she is told (.49) Item 21 Able to ask for help (.40)


Consistency Factor 2: Coefficient alpha = .79


Internal









65

Table 14: Item Loadings for Factor Three--Dysphoria

Item Loading


Looks depressed or sad (.71) Looks worried, tense (.68) Has sleep problems (.45) Sleeps unless directed into activity (.43)


Item 12 Item 11 Item 20 Item 9


Internal Consistency Factor 3: Coefficient alpha = .68









66
Table 15: Item Loadings for Factor Four--Noncompliance

Item Loading


Item 8 Has to be reminded what to do (.67) Item 17 Is incredibly passive (.49) Item 10 Has to be told to follow hospital

routine (.45) Item 14 Is uncooperative (.40)



Internal Consistency Factor 4: Coefficient alpha = .68










15.6) for the BUMPR-Hospital scores (t(46) = .54, R > .05). Ratings on the BUMPR-Home showed a similar pattern (t(46) = .57, R > .05), with fathers (M = 25.0, SD = 12.0) obtaining scores comparable to mothers (M = 23.0, SD = 11.7).

For the 24 fathers involved in the pilot study and the current study, no significant sex differences for the children emerged on the BUMPR-Hospital rating (t(22) = .80, R > .05) or for the BUMPR-Home rating (t(22) = .34, p > .05). In addition, no age group differences were found between 4-8 year old children and 8-12 year old children for the BUMPR-Hospital scores (t(22) = 1.03, R > .05) or for the BUMPR-Home scores (t(22) = .43, p > .05). Age was also not significantly correlated with the BUMPR-Hospital ratings ( = -.25, R > .05) or with the BUMPR-Home ratings (1 = -.11, R > .05). Previous number of hospitalizations was not associated with the BUMPR-Hospital scores (r = .10, R > .05) or with the BUMPR-Home scores (r = .36, p > .05). However, duration of illness was correlated with the BUMPR-Hospital ratings (r = -.49, p < .05) but not with the BUMPR-Home ratings (r = .03, R > .05).

Similar to the earlier analyses of mothers,

fathers also rated hospital distress (M = 30.7, SD = 15.6) as significantly higher (t(22) = 2.38, p < .05)









68

than home distress (M = 25.0, SD = 12.0). Moreover, the hospital distress score was also signicantly correlated with the home distress score (r = .66, p < .001).














CHAPTER 4
DISCUSSION

Background Variables Affecting Distress

The first purpose of the present study was to

examine variables influencing depression and adjustment in hospitalized children aged 4-12. The findings indicated no significant gender differences on any of the distress or depression measures. This result is consistent with the bulk of the literature (e.g., Jay et al., 1983; Saylor et al., 1987). However, it did not support the trend observed in the pilot study for girls to obtain higher BUMPR-Hospital scores than boys. A comparison of the pilot sample with the current sample revealed significantly more females in the latter sample. When the two samples were combined, no significant sex differences were found on the BUMPRHospital form, for mothers or fathers. In fact, the two studies cited earlier that suggested that girls exhibit more distress (Katz et al., 1980; Melamed & Siegel, 1975) consisted of samples with proportionately more boys. Consequently, when the sample is more evenly represented by males and females, any sex differences in depressive symptoms for pediatric populations may disappear.










Contrary to the hypothesis, younger children did not obtain higher distress scores in the current study. Many previous studies have found an inverse relationship between age and depressive symptoms in pediatric populations (e.g., Jay et al., 1983; Katz et al., 1980; Saylor et al., 1987). Moreover, age was found to be negatively correlated with BUMPR-Hospital scores in the pilot study. Thus, the current study did not confirm the negative correlation between age and distress found in the pilot study. Again, a comparison revealed that children in the pilot study were significantly younger than those in the current sample. With both samples combined, a low but significant negative correlation was found between age and mothers' BUMPR-Hospital ratings. The magnitude of this age correlation is comparable to that found in the study that introduced the BUMP-R (Saylor et al., 1987). Thus, younger children may show more distress upon hospitalization, although the relationship does not appear strong.

As anticipated, no significant differences were found for race on any of the outcome measures. The influence of ethnicity on distress in pediatric populations has not been studied. However, the absence of racial differences in distress found in the current study is consistent with research in depression for the








71

general child population (e.g., Angold, 1988; Kaslow & Racusin, 1990) as well as with the pilot study results.

With regard to family composition, only the CDI showed a small, nonsignificant difference, with children of single-parent households obtaining somewhat higher scores than children of two-parent households. This suggests that there may be a tendency for children to display more depressive symptoms because of the stressors associated with single-parent homes. However, given that this group difference was not significant and that family composition differences have not been reported in the general or pediatric population with regard to depressive symptoms, this finding may be spurious. Overall, the results confirmed the pilot study findings and supported the hypothesis that family composition does not significantly affect behavioral distress in pediatric populations.

No relationship between socioeconomic status and distress was expected. However, the PIC-D and the BUMPR-Home ratings were significantly positively correlated with SES index. Thus, mothers in lower socioeconomic levels reported more symptoms of depression and distress in their children, possibly reflecting the degree of stress associated with greater financial hardship. This finding is consistent with










the relationship found in a dental setting (Wright & Alpern, 1971). Given that the remaining measures used in this study that target distress in pediatric populations did not manifest this relationship, perhaps SES is not as influential in this group of children.

Maternal anxiety was predicted to be positively correlated with child's distress. The findings revealed that maternal reports of state and trait anxiety were generally not related to the child's behavioral upset. The only relationship that emerged involved the PIC-D and maternal trait anxiety in that mothers' reports of greater anxiety during the child's hospital stay was associated with their reports of more depressive symptoms in their children. This result may be attributed to source bias because anxious mothers may expect their children to display adjustment difficulties upon hospitalization. The absence of strong support for the relationship between children's distress and maternal anxiety contrasts with several previous studies (e.g., Blotcky et al., 1985; Jay et al., 1983; Walker & Greene, 1989). Perhaps maternal anxiety was not strongly related to child distress in the present study because those previous studies involved specific diagnostic groups (e.g., cancer, recurrent abdominal pain). When pediatric patients with diverse diagnoses are involved, parental anxiety










may not be significantly associated with the child's behavioral upset or depression.

With regard to the amount of parent contact during hospitalization, a negative correlation with child distress had been anticipated. However, only the Child-BUMP self-report scale was positively correlated with the number of hours the parent spent with the child. In other words, the more contact the child had with the parent upon hospitalization, the more behavioral upset the child reported. This finding supports studies relating maternal presence to increased negative behavior (e.g., Shaw & Routh, 1982). Children may have interpreted high contact with a parent during hospitalization as indicative of cause for concern, or the children may have felt more comfortable expressing their distress in their parents' presence. Alternatively, those parents with more distressed children may be more likely to spend more time at their child's bedside in an attempt to alleviate their distress.

In contrast, the remaining measures of child

distress were not associated with the amount of parent contact, which supports earlier findings on hospitalized children (Saylor et al., 1987). An interesting and unexpected low positive correlation between amount of parent contact and maternal state










anxiety emerged. Therefore, mothers reported greater anxiety the more time they spent with their child. This may reflect parent fatigue or it may suggest that concern for their child's illness may influence both amount of contact and anxiety ratings.

Overall, this pattern of results regarding parent contact does not support the Peterson et al. (1985) position promoting parental contact in order to avoid difficulties following separation upon hospitalization. One salient limitation in drawing conclusions about parent contact from the current study stems from subject participation being contingent upon parental presence. Obviously, those patients who spent little time with their hospitalized child were particularly likely to be unavailable for study participation. Thus, distress in children who could not participate in this study may indeed be greater with limited parent contact. In addition, parents themselves reported the number of hours spent with their child and they may have exaggerated the amount of time. Indeed, there was minimal statistical variability in reported parent contact, and, given the skewed distribution and the possible influence of outliers, results regarding amount of time the parent spent with the child should be interpreted cautiously.










A number of illness-related variables were also examined. Diagnostic group differences in depression and behavioral distress were not found in the current study. For this relatively understudied aspect of adjustment, categorization by type of illness and by chronicity of illness failed to support significant group differences, contrary to expectation. These findings are comparable to those in the pilot study and in a study of cancer patients (Jay et al., 1983). However, given the wide variety of diagnoses obtained in the current sample, categorization of illness was difficult and group sizes were possibly too small for adequate comparison. An ideal study would compare several clearly definable illness groups which might then yield differences in behavioral upset and depression.

The number of prior hospitalizations was expected to be positively associated with distress, although the results did not confirm this hypothesis. Previous medical experience was not correlated with behavioral distress in the current study or in the pilot study. Results from previous research have been mixed, with some investigations suggesting greater distress for frequently hospitalized patients (e.g., Saylor et al., 1987) and other investigations suggesting habituation and decreased upset with more hospital visits (e.g.,










Jay et al., 1983). The current findings suggest that previous medical history may be unrelated to the child's reaction to hospitalization. The current study included children with a wide variety of diagnoses and medical histories whereas the Jay et al. (1983) study involved a sample of children with cancer. Perhaps the quality of previous medical experience (i.e., negative or positive perceptions) is more influential in a child's adjustment to hospitalization than frequency of prior hospitalizations.

Contrary to the hypothesis, duration of illness and length of hospital stay were not positively correlated with the child's emotional distress. Although previous research has not studied these variables closely, children with chronic illness and longer hospitalizations were expected to obtain higher scores on measures of depression and behavioral upset. However, both the current study and pilot study did not find these relationships. One interesting negative association between maternal state anxiety and duration of illness emerged. The concept of habituation may account for this relationship, with mothers reporting less anxiety as they become accustomed to the chronic nature of their child's illness. With regard to length of hospitalization, most children were hospitalized for short periods and greater variability in length of










hospital stay may have revealed differences in adjustment.


Evaluation of the BUMP-R

The second purpose of this study was the

investigation of measures of distress particularly suitable for pediatric populations. The Behavioral Upset in Medical Patients--Revised designed by Saylor and her colleagues (1987) appears to be a promising parent report measure. The current study did not reveal significant differences between ratings of hospital and home distress, which did not support the results obtained from the pilot study. With the data from both the pilot and current study combined, a more evenly distributed sample emerged. Maternal hospital ratings in this combined sample were significantly higher than home ratings. Moreover, fathers also reported more behavioral upset in the hospital as compared to the home. Therefore, the overall findings suggest that hospitalization may precipitate emotional distress.

The current study confirmed the pilot study's

strong positive correlation between the BUMPR-Hospital scores and the BUMPR-Home scores. When the two samples were combined, this relationship remained strong for both mothers' and fathers' ratings. These results










corroborate the correlation found by Saylor et al. (1987). These authors suggested that premorbid psychological functioning, as measured by the BUMPRHome, predicts a child's adjustment to the stress of hospitalization. Consequently, children reportedly experiencing greater emotional distress at home may be more susceptible to adjustment difficulties upon hospitalization.

The BUMPR-Hospital and BUMPR-Home scores were also positively correlated with the other parent report measure, the PIC-D. This indicated that the BUMP-R taps into the construct of depression similarly to an instrument traditionally used for healthy children. Likewise, the BUMPR-Hospital and BUMPR-Home ratings were positively correlated with the children's selfreported depression on the CDI. Given that these correlations with measures commonly used in the general child population are only moderate, the BUMP-R appears to be assessing some unique aspect of child behavior. Indeed, that the BUMPR-Home ratings correlated more strongly with the PIC-D and CDI than does the BUMPRHospital would suggest that the latter is more sensitive to hospitalization behavior.

The BUMPR-Hospital rating was significantly

positively correlated with the Child-BUMP but only marginally positively correlated with the Nurse-BUMP.










On the other hand, the BUMPR-Home was not correlated with either the Child-BUMP or Nurse-BUMP. This further supports that the BUMPR-Hospital specifically measures adjustment in a hospital setting, not depressive symptoms in general.

Conclusions regarding correspondence between

mothers and nurses are limited given the small sample of 32 nurses available to complete the BUMPR-Hospital. However, from the marginal relationship found between mothers' and nurses' ratings and the strong internal consistency of the Nurse-BUMP, nurses may be able to reliably assess behavioral distress in a hospitalized child after as little as eight hours of contact.

Fathers were seldom available to participate in the study or they indicated that they preferred the mother complete the measures. Thus, this study did not address interrater agreement between fathers and mothers for the BUMPR-Hospital form. However, the sample of fathers (comprised of fathers participating conjointly with mothers and fathers participating alone) was compared to a sample of mothers matched for the child's age and sex. This comparison indicated that fathers provided BUMP-R ratings comparable to mothers. Therefore, fathers appear to be as reliable raters on the BUMP-R as mothers.








80

For the combined samples of the pilot and current studies, Cronbach's coefficient alpha for the BUMPRHospital rating was .87. Given this strong internal consistency, the most suitable and parsimonious solution in the factor analysis of the BUMPR-Hospital scores was based on the quartimax rotation. Four factors emerged, with the largest number of items representing Negativity/Agitation. The second factor was labelled as Amiability, the third factor as Dysphoria, and the fourth factor as Noncompliance. This factor structure appears to adequately capture those dimensions involved in adjustment to hospitalization.


Evaluation of Child-BUMP

A self-report pictorial measure, the Child-BUMP, was designed for the current study to tap emotional distress in hospitalized children as young as four years old. Based on a two-step selection process, children identify pictures reflecting their level of distress since hospitalization. Most children appeared to understand the instrument, particularly with repetition of items. Cronbach's coefficient alpha for the Child-BUMP was computed as .76, which indicates moderately strong internal consistency.










The Child-BUMP was moderately significantly
correlated with the parent report on the BUMPR-Hospital form. Thus, children and mothers demonstrated some agreement regarding the child's emotional distress, although they appear to hold different perspectives given that the correlation was moderate. As would be expected, the Child-BUMP, which assesses behavioral upset in the hospital, did not correlate significantly with the mothers' BUMPR-Home ratings, giving some initial indication of discriminative validity of the measure. In addition, the Child-BUMP was not associated with the parent's report on the PIC-D, possibly because the PIC-D was not intended for use in medical populations. Similarly, the CDI was not significantly correlated with the Child-BUMP. This may be attributed to the physical symptoms and schoolrelated items appearing on the CDI which would not be appropriate for hospitalized, physically-ill children. Finally, the Child-BUMP was not significantly correlated with the Nurse-BUMP, which suggests low agreement between these two sources.


Implications
The present study ventured into several relatively unexplored areas in pediatric psychology, including perspectives from various sources, incorporating self-










reports from preschool children, utilizing a heterogeneous sample of children, and examining a comprehensive array of potential risk factors. The findings from the current study highlight the need for continued research to determine which children are at risk for developing adjustment difficulties upon hospitalization. Most of the background variables examined did not clarify which children are most likely to exhibit distress in the hospital.

Of the variables investigated, maternal anxiety, parent contact, diagnosis, and previous medical experience specifically require further examination. Alternative means of assessing maternal anxiety (e.g., clinician-administered interview) may help resolve confusion regarding the influence of mothers' emotional reactions on child adjustment. With regard to parent contact, health personnel could log the amount of time the parent spends with the child to ensure accuracy of reporting. As suggested earlier, future studies could determine specific diagnostic groups to examine a priori and evaluate distress in these groups in order to obtain sufficient group sizes. In addition, investigation of previous medical experience should include the child's subjective evaluation of prior encounters with medical personnel.










Premorbid psychological functioning, as measured by the BUMPR-Home rating, was strongly related to distress upon hospitalization. Therefore, a child's emotional and behavioral difficulties outside the hospital is predictive of emotional upset upon hospitalization. Future research should assess children prior to hospitalization (e.g., for scheduled procedures) to determine which children are likely to experience adjustment difficulties. For children entering the hospital for unscheduled procedures, a brief assessment battery shortly after hospital admission could be investigated to facilitate identification of children who will likely experience difficulties in the hospital.

One intriguing direction for pediatric psychology research moves toward evaluating the influence of the context of health care delivery on adjustment. For instance, studies should more closely examine the impact of a family's access to health care. Parental experience of the availability of health care for their children may affect their attitudes towards medical intervention. Many families may also encounter poor continuity and follow-up by health care providers. Availability may also affect a family's health care orientation, i.e., their attitude towards health care and medical professionals. Belief systems also likely










influence the perceived acceptability of medical intervention. In addition, social support systems may contribute to health care orientation as well as provide assistance to parent and child during health care crises. These variables may affect the parent and child's perspective and adjustment when encountering a given hospitalization.

With regard to assessment of pediatric

populations, overall, the BUMP-R appears to be a useful instrument in the assessment of emotional distress in hospitalized children, with concurrent validity, high internal consistency, and appropriateness for various raters. Further investigation of the BUMP-R should evaluate the consistency of distress ratings over time as well as comparing scores of hospitalized children with outpatient pediatric samples in order to ascertain the specific effect of hospitalization. The current multimethod investigation utilized self-report, parentreport, and nurse-report. Correspondence between raters should be further explored, focusing on interrater reliabilities between mothers, fathers, children, and health personnel (e.g., nurses, doctors).

In addition, the Child-BUMP warrants further

investigation regarding its utility for pre-literate hospitalized children. Items should be further examined to ensure the suitability for these young








85

children. Both child and parent report of behavioral distress could also be compared to clinician ratings of depression and adjustment difficulties to further assess the validity of the BUMP-R and Child-BUMP. The BUMP-R and Child-BUMP appear to be promising instruments for assessing preschool children, an understudied age group because of the unique challenges of research with very young children.














REFERENCES


Achenbach, T. M., McConaughy, S. H., & Howell, C. T.
(1987). Child/adolescent behavioral and emotional
problems: Implications of cross-informant
correlations for situational specificity. Psychology
Bulletin, 101, 213-232.

Ammon Cavanaugh, S. von. (1986). Depression in the
hospitalized inpatient with various medical
illnesses. Psychotherapy and Psychosomatics, 45,
97-104.

Angold, A. (1988). Childhood and adolescent depression:
I. Epidemiological and aetiological aspects. British
Journal of Psychiatry, 152, 601-617.

Behrman, R. P., Vaughan, V. C. III, & Nelson, W. E.
(1987). Nelson's textbook of pediatrics (13th Ed.).
Philadelphia: W. C. Saunders Co.

Blotcky, A. D., Raczynski, J. M., Gurwitch, R., &
Smith, K. (1985). Family influences on hopelessness
among children early in the cancer experience.
Journal of Pediatric Psychology, 10, 479-493.

Boyd, J. H., & Weissman, M. M. (1981). Epidemiology of
affective disorders: A re-examination and future
directions. Archives of General Psychiatry, 38,
1039-1046.

Burke, P., Meyer, V., Kocoshis, S., Orenstein, D. M.,
Chandra, R., Nord, D. J., Sauer, J., & Cohen, E.
(1989). Depression and anxiety in pediatric
inflammatory bowel disease and cystic fibrosis.
Journal of the American Academy of Child &
Adolescent Psychiatry, 28, 948-951.

Cunningham, S. J., McGrath, P. J., Ferguson, H. B.,
Humphreys, P., D'Astous, J., Latter, J., Goodman, J.
T., & Firestone, P. (1987). Personality and
behavioural characteristics in pediatric migraine.
Headache, 27, 16-20.










Douglas, J. W. B. (1975). Early hospital admission and
later disturbances of behaviour and learning.
Developmental Medicine and Child Neurology, 17,
456-480.

Eason, L. J., Finch, A. J., Jr., Brasted, W., & Saylor,
C. F. (1985). The assessment of depression and
anxiety in hospitalized pediatric patients. Child
Psychiatry and Human Development, 16, 57-64.

Finch, A. J., Jr., & Saylor, C. F. (1984). An overview
of child depression. In W. J. Burns & J. V. Lavigne
(Eds.), Progress in pediatric psychology (pp.
201-239). Orlando, FL: Grune & Stratton.

Finch, A. J., Jr., Saylor, C. F., Edwards, G. L.
(1985). Children's Depression Inventory: Sex and
grade norms for normal children. Journal of
Consulting and Clinical Psychology, 53, 424-425.

Gauvain-Piquard, A., Rodary, C., Rezvani, A., &
Lemerle, J. (1987). Pain in children aged 2-6 years:
A new observational rating scale elaborated in a
pediatric oncology unit--preliminary report. Pain,
31, 177-188.

Gorsuch, R. L. (1983). Factor analysis (2nd Ed.).
Hillsdale, NJ: Lawrence Erlbaum Assoc.

Harter S., & Pike, R. (1983). Procedural manual to
accompany the Pictorial Scale of Perceived
Competence and Social Acceptance for Young Children.
University of Denver.

Heilgenstein, E., & Jacobsen, P. B. (1988).
Differentiating depression in medically ill children
and adolescents. Journal of the American Academy of
Child & Adolescent Psychiatry, 27, 716-719.

Jay, S. M., Ozolins, M., Elliott, C. H., & Caldwell, S.
(1983). Assessment of children's distress during painful medical procedures. Health Psycholoy, 2,
133-147.

Jensen, R. A. (1955). The hospitalized child: Round
Table, 1954. American Journal of Orthopsychiatry,
25, 293-318.

Jessner, L., Blom, G. E., & Waldfogel, S. (1952).
Emotional implications of tonsillectomy and
adenoidectomy of children. Psychoanalytic Study of
the Child, 7, 126-169.











Johnson, R., & Baldwin, D. C. (1968). Relationship of
maternal anxiety to the behavior of young children
undergoing dental extraction. Journal of Dental
Research, 47, 801-805.

Kandel, D. B., & Davies, M. (1982). Epidemiology of
depressive mood in adolescents: An empirical study.
Archives of General Psychiatry, 39, 1205-1212.

Kashani, J. H., Barbero, G. J., & Bolander, F. D.
(1981). Depression in hospitalized pediatric
patients. Journal of the American Academy of Child
Psychiatry, 20, 123-134.

Kashani, J. H., Holcomb, W. R., & Orvaschel, H. (1986).
Depression and depressive symptoms in preschool children from the general population. American
Journal of Psychiatry, 143, 1138-1143.

Kashani, J. H., Husain, A., Shekim, W. O., Hodges, K.
K., Cytryn, L., & McKnew, D. H. (1981). Current
perspectives on childhood depression: An overview.
American Journal of Psychiatry, 138, 143-153.

Kaslow, N. J., & Racusin, G. R. (1990). Childhood
depression: Current status and future directions. In
A. S. Bellack, M. Hersen, & A. E. Kazdin (Eds.),
International handbook of behavior modification and
therapy (2nd ed., pp. 649-667). New York: Plenum
Press.

Kaslow, N. J., Rehm, L. P., Pollack, S. L., & Siegel,
A. W. (1988). Attributional style and self-control
behavior in depressed and nondepressed children and their parents. Journal of Abnormal Child Psychology,
16, 163-175.

Katz, E. R., Kellerman, J., & Siegel, S. (1980).
Behavioral distress in children with cancer undergoing medical procedures: Developmental
considerations. Journal of Consulting and Clinical
Psychology, 48, 356-365.

Kazdin, A. E. (1987). Assessment of childhood
depression: Current issues and strategies.
Behavioral Assessment, 9, 291-319.

Kazdin, A. E. (1988). Childhood depression. In E. J.
Mash & L. G. Terdal (Eds.), Behavioral assessment of
childhood disorders (2nd ed., pp. 157-195). New
York: Guilford Press.










Kazdin, A. E. (1989). Identifying depression in
children: A comparison of alternative selection
criteria. Journal of Abnormal Child Psychology, 17,
437-454.

Kazdin, A. E., Colbus, D., Rodgers, A. (1986).
Assessment of depression and diagnosis of depressive
disorder among psychiatrically disturbed children.
Journal of Abnormal Child Psychology, 14, 499-515.

Kazdin, A. E., French, N. H., Unis, A. S., &
Esveldt-Dawson, K. (1983). Assessment of childhood
depression: Correspondence of child and parent
ratings. Journal of the American Academy of Child
Psychiatry, 22, 157-164.

Kerr, M. M., Holer, T. S., Versi, M. (1987).
Methodological issues in childhood depression: A
review of the literature. American Journal of
Orthopsychiatry, 57, 193-198.

Kovacs, M. (1983). The Children's Depression Inventory:
A self-rated depression scale for school-aged
youngsters. Unpublished manuscript, University of
Pittsburg School of Medicine.

Kovacs, M. (1985). The Children's Depression Inventory
(CDI). Psychopharmacology Bulletin, 21, 995-998.

Lachar, D., & Gdowski, C. L. (1979). Actuarial
assessment of child and adolescent personality: An
interpretive guide for the Personality Inventory for
Children profile. Los Angeles, CA: Western
Psychological Services.

Lefkowitz, M. M., & Tesiny, E. P. (1980). Assessment of
childhood depression. Journal of Consulting and
Clinical Psychology, 48, 43-50.

Leon, G. R., Kendall, P. C., Garber, J. (1980).
Depression in children: Parent, teacher, and child
perspectives. Abnormal Child Psychology, A, 221-235.

Levenson, J. L., Hamer, R., Silverman, J. J., Rossiter,
L. F. (1986-87). Psychopathology in medical
inpatients and its relationship to length of
hospital stay: A pilot study. International Journal
of Psychiatry in Medicine, 16, 231-236.










Lichtenstein, D., Dreger, R. M., & Cattell, R. B.
(1986). Factor structure and standardization of the
Preschool Personality Questionnaire. Journal of
Social Behavior and Personality, 1, 165-181.

Melamed, B. G., & Siegel, L. J. (1975). Reduction of
anxiety in children facing hospitalization and
surgery by use of filmed modeling. Journal of
Consulting and Clinical Psychology, 43, 511-521.

Mullins, L. L., Siegel, L. J., & Hodges, K. (1985).
Cognitive Problem-Solving and life-event correlates
of depressive symptoms in children. Journal of
Abnormal Child Psychology, 13, 305-314.

Myers, J., & Bean, L. (1968). A decade later: A followup of social class and mental illness. New York:
Wiley & Sons.

Olson, R. A., Holden, E. W., Friedman, A., Faust, J.,
Kenning, M., & Mason, P. J. (1988). Psychological
consultation in a children's hospital: An evaluation
of services. Journal of Pediatric Psychology, 13,
479-492.

Peterson, L., Mori, L., & Carter, P. (1985). The role
of the family in children's responses to stressful
medical procedures. Journal of Clinical Child
Psycholo 14, 98-104.

Pfeffer, C. R., & Trad, P. V. (1988). Sadness and
suicidal tendencies in preschool children.
Developmental and behavioral pediatrics, 2, 86-88.

Poznanski, E. D., Cook, S. C., & Carroll, B. J. (1979).
A depression rating scale for children. Pediatrics,
64, 442-450.

Prugh, D. G., Staub, E. M., Sands, H. H., Kirschbaum,
R. M., & Lenihan, E. A. (1953). A study of the
emotional reactions of children and families to
hospitalization and illness. American Journal of
Orthopsychiatry, 23, 70-106.

Reynolds, W. M., Anderson, G., & Bartell, N. (1985).
Measuring depression in children: A multimethod
assessment investigation. Journal of Abnormal Child
Psychology, 13, 513-526.











Ryan, N. D., Puig-Antich, J., Ambrosini, P.,
Rabinovich, H., Robinson, D., Nelson, B., Iyengar,
S., & Twomey, J. (1987). The clinical picture of
major depression in children and adolescents.
Archives of General Psychiatry, 44, 854-861.

Saylor, C. F., Finch, A. J., Jr., & McIntosh, J. A.
(1988). Self-reported depression in psychiatric,
pediatric, and normal populations. Child Psychiatry
& Human Development, 18, 250-254.

Saylor, C. F., Finch, A. J., Jr., Spirito, A., &
Bennett, B. (1984). The Children's Depression
Inventory: A systematic evaluation of psychometric
properties. Journal of Consulting and Clinical
Psychology, 52, 955-967.

Saylor, C. F., Pallmeyer, T. P., Finch, A. J., Jr.,
Eason, L., Trieber, F., & Folger, C. (1987).
Predictors of psychological distress in hospitalized
pediatric patients. Journal of the American Academy
of Child and Adolescent Psychiatry, 26, 232-236.

Seligman, M. E. P., Peterson, C., Kaslow, N. J.,
Tanenbaum, R. L., Alloy, L. B., & Abramson, L. Y.
(1984). Attributional style and depressive symptoms among children. Journal of Abnormal Psychology, 93,
235-238.

Shanahan, K. M., Zolkowski-Wynne, J., Coury D. L.,
Collins, E. W., O'Shea, J. S. (1987). The Children's
Depression Rating Scale for normal and depressed
outpatients. Clinical Pediatrics, 26, 245-247.

Shaw, E. G., & Routh, D. K. (1982). Effect of mother
presence on children's reaction to aversive
procedures. Journal of Pediatric Psychology, 2,
33-42.

Spielberger, C. D. (1983). Manual for the State-Trait
Anxiety Inventory (Form Y, Self-Evaluation
Questionnaire). Palo Alto, CA: Consulting
Psychologists Press.

Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E.
(1970). Manual for the State-Trait Anxiety Inventory
(Self-Evaluation Questionnaire). Palo Alto, CA:
Consulting Psychologists Press.




Full Text
ITEM 23
This girl starts talking to other people.
DO YOU:
This girl waits for other people to start talking to her.
DO YOU:
always
OR sometimes
sometimes
wait for
wait for
start talking
other
other
to other people
people to
people to
start
start
talking
talking
4
3
2
always
start talking
to other people
142


more likely to experience adjustment problems upon
hospitalization. Concerning assessment, the BUMP-R
demonstrated internal consistency and concurrent
validity. A factor analysis of mothers from the
current study combined with the pilot study revealed
four factors identified as Negativity/Agitation,
Amiability, Dysphoria, and Noncompliance. The child
pictorial measure also demonstrated internal
consistency and correspondence with mothers' ratings.
Therefore, both measures are promising for use with
hospitalized children.
ix


151


APPENDIX D
DATA COLLECTION PROCEDURES TRAINING GUIDE
(Make sure to present all of the following points:)
Hi. My name is I'm conducting a study with the
clinical and health psychology department. We're
looking at how children react to being hospitalized.
We'd like to ask you to participate in this study, if
you agree, I WILL first ask you a few general questions
about your child and their living situation. Then I
will give you some questionnaires about your child's
reactions and about your own feelings since coming to
the hospital. These questionnaires will take about
30-45 minutes to complete. Also, we'll ask your
child's nurse to fill out a brief questionnaire about
their opinion of your child's reaction. While you're
completing the questionnaires, we'll ask your child to
answer some questions about how they have been feeling
during their stay.
(At this point, you must make an assessment of the
child's ability to participate. Determine:
A. is the child sleeping?
1. if yes, does parent mind if we wake them up?
a. if yes, can we come back in an hour?
B. is the child sedated?
1. if yes, are they too sedated to participate?
a. if yes, will it wear off in an hour?
C. is the child developmentally delayed?
1. if yes, show some questions to parent and
ask would child be able to understand them?
D. is the child hearing or sight impaired?
1. if yes, can child see/hear well enough to
participate?
If the final answers to questions A, B, C, and D are
no, then we won't be able to evaluate child. Thank
parent for time and cooperation, explaining that we
must have data from both the child and the parent.
152


121


ITEM 13
This girl needs her parents to make her
feel better.
This girl doesnt need her parents to make her
feel better.
DO YOU:
DO YOU:
always OR sometimes
need your need your
parents to parents to
make you make you
feel better feel better
hardly ever OR never
need your need your
parents to parents to
make you make you
feel better feel better
4 3
2 1
122


105


This girl laughs and smiles at funny things.
DO YOU:
always OR
laugh and
smile at
funny things
sometimes
laugh and
smile at
funny things
1
2
ITEM 24
This girl doesnt laugh and smile at funny things.
DO YOU:
hardly ever
laugh and
smile at
funny things
OR never
laugh and
smile at
funny things
3
4
144


139


23
receiving diagnoses for acute illnesses and 72%
receiving diagnoses for chronic illnesses. The
remaining thirteen diagnoses not categorized on
chronicity represent children with ambiguous or unknown
illnesses.
In addition to data collected on mothers, eleven
fathers also participated in the pilot study. Three of
these fathers completed the measure along with their
spouses, whereas the remaining eight fathers were the
only parent available for participation in the study.
Children who exhibited mental or physical
limitations which would interfere with parental report
of distress were not included in the sample. For
instance, children with a developmental delay or
hearing impairment would complicate parental response
to items regarding conversational ability (e.g.,
"Refuses to speak") or on items requiring comprehension
of requests (e.g., "Accepts advice or instructions
easily"). Determination of whether the child met these
exclusionary criteria was made through interaction with
the child and soliciting the judgment of the parent
and/or nurse.
Measure. The Behavioral Upset in Medical
PatientsRevised (BUMP-R; Saylor et al., 1987) is a
56-item parent rating of the child's behavior
corresponding to emotional distress at the hospital and


59
Table 10:
Eigenvalues for the BUMPR-Hospital Factor
Analysis
Eigenvalues
Factor 1
6.07
Factor 2
2.50
Factor 3
1.78
Factor 4
1.09
The next criteria involved a plot of the
eigenvalues, known as the scree test. Gorsuch (1983)
describes that the predominant factors will account for
a significant portion of the variance. Thus, the
number of factors can be extracted where the plot of
the eigenvalues levels off. The scree test shown in
Figure 1 supports extraction of four factors.
Lastly, interpretability was used as the final
criterion. Comparisons to three factor and five factor
solutions confirmed that the quartimax four-factor
solution retained the greatest number of items and
resulted in the most parsimonious factor structure.
Moreover, the quartimax four factor rotation yielded
the most readily interpretable solution.
Item loadings of .40 were considered significant,
which is more conservative than the traditional .30
cutoff and would result in stronger factors. Only Item


85
children. Both child and parent report of behavioral
distress could also be compared to clinician ratings of
depression and adjustment difficulties to further
assess the validity of the BUMP-R and Child-BUMP. The
BUMP-R and Child-BUMP appear to be promising
instruments for assessing preschool children, an
understudied age group because of the unique challenges
of research with very young children.


107


ITEM 6
This girl doesnt say shes sad.
DO YOU:
never say OR hardly ever say
youre sad youre sad
1
2
This girl says shes sad.
DO YOU:
sometimes say OR always say
youre sad youre sad
3
4
108


61
14 (Is uncooperative) loaded on more than one factor,
and two items did not have significant loadings on any
of the four factors (Item 6: Refuses to speak; Item 7:
Says he or she feels blue or depressed; See Table 11
for factor structure). Using the .40 item loading
criterion, the first factor was extracted with 11
items, identified as Negativity/Agitation (See Table
12), with an internal consistency coefficient of .86.
This first factor accounted for 41.2% of the variance.
The second factor of nine items, labelled Amiability
(See Table 13), encompassed sociability and
agreeability. This factor attained an internal
consistency coefficient of .79 and accounted for 26.3%
of the variance. The third factor with four items was
named Dysphoria (See Table 14), including vegetative
symptoms of depression and emotional discomfort, with
internal consistency at .68. The third factor
accounted for 19.9% of the variance. The final factor
was identified as Noncompliance (See Table 15), with
four items and an internal consistency of .68. The
last factor accounted for 12.5% of the variance.
Analysis of Fathers for Combined Sample
To examine similarities between parental raters,
fathers were compared to a matched sample of mothers.
Results indicated that fathers' ratings (M = 30.7, SD =
15.6) were similar to mothers' ratings (M = 28.3, SD =


49
revealed no significant interaction between sex and age
group or between sex and race.
An examination of the Spearman coefficients (See
Table 7) indicated that the length of the current
hospital visit and the number of prior hospitalizations
were not associated with any of the measures of
behavioral distress. Duration of illness was
significantly negatively correlated with only parental
state anxiety. Moreover, the relationship between the
Child-BUMP scores and hours spent with the parent was
positively correlated, a direction opposite of that
predicted. Thus, the more time a parent spent with the
child, the more distress the child reported on the
pictorial measure. Although no other significant
relationships with time spent with the parent emerged,
a modest negative correlation with parental state
anxiety was noted.
With regard to the six diagnostic groups, there
were no significant differences on the BUMPR-Hospital
scores (F(53) = .87, p > .05), on the CDI (F(26) =
1.92, p >.05), or on the Nurse-BUMP (F(24) = 1.68, p >
.05). Only a marginal difference in diagnostic groups
was found with the PIC-D scores (F(53) = 2.11, p = .08)
and with the Child-BUMP scores (F(53) = 2.13, p = .08).
Based on the classification of chronicity of illness,
no significant differences appeared on any of the




79
On the other hand, the BUMPR-Home was not correlated
with either the Child-BUMP or Nurse-BUMP. This further
supports that the BUMPR-Hospital specifically measures
adjustment in a hospital setting, not depressive
symptoms in general.
Conclusions regarding correspondence between
mothers and nurses are limited given the small sample
of 32 nurses available to complete the BUMPR-Hospital.
However, from the marginal relationship found between
mothers' and nurses' ratings and the strong internal
consistency of the Nurse-BUMP, nurses may be able to
reliably assess behavioral distress in a hospitalized
child after as little as eight hours of contact.
Fathers were seldom available to participate in
the study or they indicated that they preferred the
mother complete the measures. Thus, this study did not
address interrater agreement between fathers and
mothers for the BUMPR-Hospital form. However, the
sample of fathers (comprised of fathers participating
conjointly with mothers and fathers participating
alone) was compared to a sample of mothers matched for
the child's age and sex. This comparison indicated
that fathers provided BUMP-R ratings comparable to
mothers. Therefore, fathers appear to be as reliable
raters on the BUMP-R as mothers.


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
LIST OF TABLES vi
ABSTRACT V
CHAPTERS
1 INTRODUCTION 1
General Childhood Depression 1
Prevalence 1
Assessment Issues 2
Risk Factors 4
Pediatric Population 8
Definition Issues 8
Prevalence 10
Assessment Issues 11
Risk Factors 14
Pilot Study 2 0
Introduction and Rationale 2 0
Methods 21
Results and Discussion 25
Purpose of Study and Hypotheses 27
2 METHODS 29
Subjects 29
Measures 3 2
Procedure 38
3 RESULTS 4 0
Analyses for Current Study 41
Descriptive Results & Statistics 41
Analyses of Background Data 4 3
Correlational Analyses 51
iv


8
racial differences have been found in depressed
prepubertal children (Angold, 1988; Kaslow & Racusin,
1990). Although a few studies have suggested that
depression may be associated with lower socioeconomic
status (see Angold, 1988 for review), depressive
symptomatology was not found to be related to father's
education or family income (Kandel & Davies, 1982).
Finally, no differences in socioeconomic status or
family composition (i.e., child living in single-parent
home, two-parent home) were found between depressed or
nondepressed clinic children (Kaslow, Rehm, Pollack, &
Siegel, 1988). The effects of these demographic
variables on childhood depression, however, are seldom
explored.
Pediatric Populations
Definition Issues
Given the relationship between stressful life
events and depressive symptoms (e.g., Mullins, Siegel,
& Hodges, 1985), pediatric psychology has evidenced a
growing interest in depressive symptoms in medical
populations. Few studies have examined the prevalence
of major depression in pediatric populations. In
contrast, several studies have focused on adjustment in
medical settings. The depressive symptoms associated
with poor adjustment in pediatric populations


Figure 1: Scree Test of Eigenvalues for Factor Analysis of
BUMP-R Hospital Ratings
60
/ +
6 +
5 +
4 +
3 +
2 +
6
7
8 9
0 1
2 3 4
5 67 89 01 23 45
+ + + + + + + + + +--
6 8 10 12 14 16 18 20 22 24 26
28
Number
VO +


This girl can ask others for help.
DO YOU:
always ask OR sometimes
others for ask others
help for help
1 2
ITEM 20
This girl cant ask others for help.
DO YOU:
hardly ever OR never ask
ask others others for
for help help
3
4


82
reports from preschool children, utilizing a
heterogeneous sample of children, and examining a
comprehensive array of potential risk factors. The
findings from the current study highlight the need for
continued research to determine which children are at
risk for developing adjustment difficulties upon
hospitalization. Most of the background variables
examined did not clarify which children are most likely
to exhibit distress in the hospital.
Of the variables investigated, maternal anxiety,
parent contact, diagnosis, and previous medical
experience specifically require further examination.
Alternative means of assessing maternal anxiety (e.g.,
clinician-administered interview) may help resolve
confusion regarding the influence of mothers' emotional
reactions on child adjustment. With regard to parent
contact, health personnel could log the amount of time
the parent spends with the child to ensure accuracy of
reporting. As suggested earlier, future studies could
determine specific diagnostic groups to examine a
priori and evaluate distress in these groups in order
to obtain sufficient group sizes. In addition,
investigation of previous medical experience should
include the child's subjective evaluation of prior
encounters with medical personnel.


154
(and quickly read the first question not already
completed. If you're at ease, the parent will think
this is normal procedure and won't be uncomfortable.
Once the parent questionnaires are underway, introduce
the study to the child, saying the following more or
lessbe sure to adapt wording to child's age level:)
As I told your mom (and dad), I'm going around talking
to the kids in the hospital asking them how they feel
about being here. I'll be asking you some serious
questions which will be kept secret. We want to try to
find out what makes kids have a hard time in the
hospital so that we can help other kids when they have
to come here. There's no right or wrong answers to the
questions. I just want to find out how you feel since
you came to the hospital. When we're done, I have some
prizes you can choose from to thank you for helping us.
Okay? First, I'll show you two pictures and ask you to
pick which picture is most like how you have been
feeling in the hospital...
(Continue by reading the questions, periodically
reminding them that they are to chose the picture which
is most like them in the hospital.
If the child is between 8 and 12, then administer the
CDI. The revised directions should be :)
Now I'm going to read to you some groups of sentences.
For each group, pick one sentence that describes you
best since you got to the hospital. Remember, there is
no right or wrong answer. Just pick the sentence that
best describes the way you have been since you got
here. Here is an example of how it works.
(Then read the sample question and continue with the
remaining items. Again, remind them periodically that
they are to chose the sentence that describes them
since they got to the hospital.
When finished with the CDI, select two appropriate toys
for the child and ask them to choose one. Thank them
for their help.
Check to see if parent is done with questionnaires. If
not, decide if you want to wait in the room or come
back later to pick them up. If done, thank them
profusely. Remind them that someone will be back to
collect information before the child is discharged if
they are in the hospital for at least another day.
Thank them again on your way out.


This girl listens when others try to help her.
DO YOU:
ITEM 22
This doesnt listen when others try to help her.
DO YOU:
always OR
sometimes
hardly ever
OR never listen
listen when
listen when
listen when
when others
others try
others try
others try
try to help
to help
to help
to help
1
2
3
4
140


78
corroborate the correlation found by Saylor et al.
(1987). These authors suggested that premorbid
psychological functioning, as measured by the BUMPR-
Home, predicts a child's adjustment to the stress of
hospitalization. Consequently, children reportedly
experiencing greater emotional distress at home may be
more susceptible to adjustment difficulties upon
hospitalization.
The BUMPR-Hospital and BUMPR-Home scores were also
positively correlated with the other parent report
measure, the PIC-D. This indicated that the BUMP-R
taps into the construct of depression similarly to an
instrument traditionally used for healthy children.
Likewise, the BUMPR-Hospital and BUMPR-Home ratings
were positively correlated with the children's self-
reported depression on the CDI. Given that these
correlations with measures commonly used in the general
child population are only moderate, the BUMP-R appears
to be assessing some unique aspect of child behavior.
Indeed, that the BUMPR-Home ratings correlated more
strongly with the PIC-D and CDI than does the BUMPR-
Hospital would suggest that the latter is more
sensitive to hospitalization behavior.
The BUMPR-Hospital rating was significantly
positively correlated with the Child-BUMP but only
marginally positively correlated with the Nurse-BUMP.


14
distress and was devised for hospitalized children with
a variety of diagnoses.
Risk Factors
Factors which influence the prevalence of
childhood depression in the general population may also
affect pediatric populations. With respect to gender
differences, results have been mixed. Consistent with
findings in the general childhood depression
literature, a study of pediatric cancer patients ages
2-20 found no gender differences on an observational
measure of behavioral distress during a painful medical
procedure (Jay et al., 1983). Moreover, no sex
differences were found in hospitalized pediatric
patients ages 5-15 on self-reported depression,
interview-rated depression, or parent-reported
behavioral upset (Saylor et al., 1987). One study of
hospitalized children resulted in a greater number of
boys diagnosed with depression, although this subsample
was small (Kashani, Barbero, & Bolander, 1981). Some
studies have found girls demonstrated more behavioral
distress during painful medical procedures (e.g., Katz
et al., 1980; Melamed & Siegel, 1975). Although the
literature suggests that girls may exhibit more
distress, evidence for overall gender differences in
depressive symptoms for pediatric populations is
limited.


This girl has to have her own way.
DO YOU:
always OR sometimes
have to have to
have your have your
own way own way
4
3
ITEM 15
This girl doesnt have to have her own way.
DO YOU:
hardly ever
have to
have your
own way
OR never
have to
have your
own way
2
1
126


32
addition to the participation of mothers, thirteen
fathers were also involved in the study. Five of the
fathers were spouses of mothers who also participated
in the study whereas eight fathers were the only
parental respondents involved in the study.
Children who manifested mental or physical
deficits which would impede responses on the child
self-report and parent measures were excluded (e.g.,
profound mental retardation, lack of language
development, significant vision or hearing impairment).
Comprehension problems or the inability to see the
pictures, for example, would interfere with the child's
responses on the Child-BUMP. The evaluation of whether
a child met these exclusionary criteria was made
through interaction with the child and from parent
and/or nurse judgments. Informed parental consent and
child verbal assent were independently obtained. Basic
demographic information and information regarding
illness was collected on a summary sheet (See Appendix
B) .
Measures
The Behavioral Upset in Medical PatientsRevised
(BUMP-R; Savior et al., 1987). As detailed in the
description of the pilot study (p.23-25), the BUMP-R is
a 56-item parent rating of the child's behavior


Table 11; Factor Structure of the BUMPR-Hosnital
FACTOR 1
FACTOR 2
FACTOR 3
FACTOR 4
Item
1
.63589
-.04183
-.00338
. 02754
Item
2
.45435
.12336
.36104
. 00857
Item
3
.59320
.02689
.18392
-.04462
Item
4
.65127
.13778
.22307
-.05709
Item
5
.65284
.12495
.17103
-.02268
Item
6
.30260
.37821
.25504
-.06537
Item
7
.18094
.18013
.28743
.01944
Item
8
.22155
.17271
.10444
.66649
Item
9
-.06193
-.00589
.43279
.14313
Item
10
.32859
-.03857
-.10527
.45162
Item
11
.19292
.08428
.67505
-.02277
Item
12
.16359
.22701
.70590
-.03168
Item
13
.53006
.07606
.30855
.02363
Item
14
.57713
.26186
.08055
.39853
Item
15
.51382
.06229
-.02597
-.14439
Item
16
.70806
.08321
-.01942
.19862
Item
17
.12274
.07588
.32630
.48533
Item
18
.72155
.03593
-.16548
.20505
Item
19
.55057
-.10472
-.12761
.29440
Item
20
.21628
.10644
.45006
.00852
Item
21
-.14432
.39522
.02132
.13843
Item
22
.25396
.67377
.14469
-.12223
Item
23
.28992
.64513
-.13552
.15905
Item
24
-.03104
.53951
.20288
-.09263
Item
25
.04713
.56132
.30167
-.11305
Item
26
.24431
.63897
.12487
.01719
Item
27
.03600
.54897
.01670
.08005
Item
28
.23332
.48658
-.27860
.27999


This girl doesnt try to get better in the hospital.
DO YOU:
ITEM 27
This girl tries to get better in the hospital.
DO YOU:
never OR
hardly ever
sometimes
OR always
try to get
try to get
try to get
try to get
better in
better in
better in
better in
the hospital
the hospital
the hospital
the hospital
4
3
2
1
150


APPENDIX A
BEHAVIORAL UPSET IN MEDICAL PATIENTS REVISED (BUMP-R)
PARENT RATING FORM
Subject #: DATE:
Your relationship to child:
Side ope
PLEASE RATE YOOR CHILD'S BEHAVIOR AS YOD HAVE SEEN IT SINCE YODR ARRIVAL AT THE
HOSPITAL. MARK ONE BOX PER QUESTION TO INDICATE YODR CHOICE.
1.
2 .
Is impatient
Cries
u
s
>
e
a
n
r i
0)

ft
H
u

fl
o
m
1
[]
[]
a

%4
o
n
r i
*
M
-t
s
3
e
D
[]
r n
s
>
s
*
iH
A
C 3
r i
3.
Gets angry
[]
l J
[]
l J
[3
l J
[]
4.
Becomes upset easily
[]
[]
n
[3
[]
5.
Is irritable or grouchy
[]
[]
n
[]
[3
6.
Refuses to speak
[]
u
n
[3
C 3
7.
says he or she feels blue or depressed
[]
[]
n
C 3
[3
8.
Has to be reminded what to do
[]
[]
[]
[]
[]
9.
Sleeps unless directed into activity..
M
n
[]
[3
[]
10.
Has to be told to follow hospital routine....
[]
[]
[3
[]
[3
11.
Looks worried, tense
[]
[]
C 3
[3
C3
12.
Looks depressed or sad
[]
c]
C3
(3
C3
13.
Clinging, needs lots of reassurance
[]
[]

[]
C 3
14.
Is uncooperative
M
[]
C3
[ 3
[3
15.
Complains
[]
[]
C3
C3
[]
16.
Stubborn, negativistio.
[]
u
tl

[ 3
17.
Is incredibly passive
[]
[3
[3
C 3
[]
18.
Demanding
[]
[]
[]
C 3
C 3
19.
Manipulative
[]
U
[3
[ 3
C 3
20.
Has sleep problems
[]
[]
[3
C 3
[3
21.
Able to ask for help
u
[]
[3


22.
Tries to be friendly
[]
[]
[]
[ 3
C 3
23.
Accepts advice or instructions easily
[]
[]
[3
Cl
n
24.
Starts conversations
[]
[]
C3
[3
3
25.
Laughs or smiles at funny comments or events.
[]
[]
[ 3
[3
[3
26.
Pleasant to be with
[]
C3
[3
[]
[3
27 .
Shows interest in recovery (takes initiative)
[]
[]
C3
[]
[]
28.
Does what he or she is told
[]
[]
[]
[]
n
94


37
The scale was modeled according to the Harter &
Pike (1983) Pictorial Scale of Perceived Competence and
Social Acceptance for Young Children. In this scale,
children are asked to select from two items with
corresponding pictures which is most like them. Then,
for each item, two possible questions are asked
regarding the frequency with which they exhibit the
behavior. This two-step process facilitates attention
and simplifies the items for young children. A sample
item from the Child-BUMP is as follows:
This boy cries a lot. This boy cries a little.
(If child selects first picture,
ask next two questions)
DO YOU:
always cry OR sometimes cry
(If child selects second picture,
ask next two questions)
DO YOU:
hardly ever cry OR never cry
Each final response is scored on a scale of one to
four, with high scores suggestive of greater behavioral
distress. A total score was obtained by summing the
item scores. Fourteen of the items begin with the most
negative response (4) and thirteen of the items depict
African-American children. Male and female versions of


93
Zeldow, P. B., & Braun, L. (1985). Measuring regression
in hospitalized medical patients: The BUMP scale.
General Hospital Psychiatry, 7, 49-53.


12
teachers, and health personnel (see Katz et al., 1980
for example of nurse ratings in pediatric populations).
Compounding the usual assessment issues, the
evaluation of emotional distress in pediatric
populations is further complicated by the scarcity of
measures appropriate for the assessment of these
children. The application of standard measures of
depressive symptoms, such as the CDI, may not be
suitable for hospitalized children. For example, the
CDI did not differentiate between a sample of chronic
headache sufferers and a non-headache control group
(Wisniewski, Naglieri, & Mulick, 1988). Similarly,
pediatric inpatients attained CDI scores comparable to
gender and age-matched schoolchildren (Saylor, Finch, &
McIntosh, 1988). In fact, one study found that normal
children tended to report more depression on the CDI
than cancer patients, which the researchers suggest
reflects denial of symptoms (Worchel, Nolan, Willson,
Purser, Copeland, & Pfefferbaum, 1988). In contrast,
another study found that children with recurrent
abdominal pain scored significantly higher on the CDI
than healthy children (Walker & Greene, 1989) These
mixed results with the CDI question its utility in
pediatric populations.
Likewise, a clinician-administered structured
interview, the Children's Depression Rating Scale


95
Side Two
PLEASE RATE YOUR CHILD'S BEHAVIOR AS YOU HAVE SEEN IT AT.
WAS HOSPITALIZED. MARK ONE BOX PER QUESTION TO INDICATE
HOME. BEFORE
YOUR CHOICE.
HE OR SHE
1. is impatient
2. Cries
3. Gets angry
4. Becomes upset easily
5. Is irritable or grouchy
6. Refuses to speak
7. Says he or she feels blue or depressed
8. Has to be reminded what to do
9. sleeps unless directed into activity
10. Has to be told to follow hospital routine....
11. Looks worried, tense
12. Looks depressed or sad
13. Clinging, needs lots of reassurance
14. Is uncooperative
15. Complains
16. Stubborn, negativistic
17. Is incredibly passive
18. Demanding
19. Manipulative
20. Has sleep problems
21. Able to ask for help...
22. Tries to be friendly
23. Accepts advice or instructions easily
24. Starts conversations
25. Laughs or smiles at funny comments or events.
26. Pleasant to be with
27. Shows interest in recovery (takes initiative)
28. Does what he or she is told
u
05

a
H
4J
t
H
H
a
>i
9
<8
9
9
9
>
a
4J
P

9
0
%4
9
H
z
00
o
D
<
. []
[]
[]
Cl
Cl
. []
[]
Cl
u
Cl
. []
[J
[]
[]
Cl
. []
[]
Cl
n
[]
[]
[]
Cl
n
[]
[]
[]

n
[]
. []
[]
Cl

Cl
[]
[]
[]
n
n
[]
Cl
[]
n
Cl
[]
[]
[]
u
Cl
[]
[]
[]
n
Cl
. []
Cl
[]
C3
Cl
n
[3
n
n
Cl
[]


u
Cl
ci
[]
n
n
Cl
u

n
n
Cl
ti
C3
n
n
Cl
n
ci
n
t]
Cl
[j
cj
n
n
Cl
n
[]

Cl
[]
[]
[]
n
Cl
[]
[]
n
Cl
Cl
Cl
n
[]
u
Cl
Cl
n
n
Cl
Cl
Cl
n
n
n
[]
[]
u
n
u
Cl
Cl
n
[]
n
Cl
[]
[]

n
Cl
Cl


127
i
O


109


ITEM 16
This girl does a lot of things.
This girl doesnt do a
DO YOU:
DO YOU:
always OR sometimes
do a lot of do a lot of
things things
hardly ever OR
do a lot of
things
1
2
3
lot of things.
never
do a lot of
things
4
128


99
O
o
o


ITEM 3
This girl gets mad a lot.
ARE YOU:
always OR sometimes mad
mad
4
3
This girl doesnt get mad a lot.
ARE YOU:
hardly ever mad OR never mad
2
1
102


28
(2) Differences in depression and behavioral upset
were hypothesized for different diagnostic groups.
(3) Onset of illness, number of prior
hospitalizations, and length of current hospital stay
were expected to be positively related to behavioral
upset and depression.
(4) Number of hours spent with parent was expected
to be negatively correlated with adjustment to
hospitalization.
(5) All outcome measures, including parental
anxiety, were hypothesized to be intercorrelated.
(6) In addition, behavioral distress was expected
to be rated higher in the hospital setting as compared
to home, indicating distress at hospitalization.


I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Professor of Nursing
This dissertation was submitted to the Graduate
Faculty of the College of Health Related Professions
and to the Graduate School and was accepted as partial
fulfillment of the requirements for the degree of
Doctor of Philosophy.
August 1993
Dean, College of Health
Related Professions
Dean, Graduate School


2
criteria for establishing prevalence in a given
population.
Assessment Issues
The research on general childhood depression
contains numerous conflicting results. Much of the
variability in estimating prevalence and in estimating
the influence of variables may be attributed to
difficulties in assessment. The definitions of
depression determine the selection of assessment
techniques (Angold, 1988). Although several studies
utilize diagnostic criteria for depression (see Finch &
Saylor, 1984 for review), the majority of studies
define depression based on scores obtained on a variety
of measures. However, many inventories gain only
temporary popularity and are utilized in only a single
study (Kerr, Holer, & Versi, 1987). Thus, depression
is interpreted differently, which complicates questions
regarding the influence of numerous variables.
With adults, assessment of depression is often
based on self-report or interview data. However,
different assessment methods are conducted at different
ages. For instance, assessment of adolescents, like
adult assessment, often utilizes self-report and
interview methods. However, investigation of affective
disturbance in preschool children is particularly
challenging. A preschool child is not readily able to


66
Table 15: Item Loadings for Factor FourNoncompliance
Item
Loading
Item 8
Has to be reminded what to do
(.67)
Item 17
Is incredibly passive
(.49)
Item 10
Has to be told to follow hospital
routine
(.45)
Item 14
Is uncooperative
(.40)
Internal Consistency Factor 4: Coefficient alpha = .68


44
Table 2: Demographic Differences for the BUMPR-Hospital
N
Mean
l (SD)
t
Sex
Males
29
28.3
(14.2)
.60
Females
41
26.4
(12.7)
Acre Group
4-8
32
29.0
(15.5)
1.01
8-12
38
25.7
(11.0)
Race
White
46
28.1
(13.0)
.36
African-American
19
26.7
(14.9)
Family Composition
2-Parent
44
25.4
(12.8)
1.55
1-Parent
23
30.6
(13.5)
Note; All t values are nonsignificant.


REFERENCES
Achenbach, T. M., McConaughy, S. H., & Howell, C. T.
(1987). Child/adolescent behavioral and emotional
problems: Implications of cross-informant
correlations for situational specificity. Psychology
Bulletin. 101. 213-232.
Ammon Cavanaugh, S. von. (1986). Depression in the
hospitalized inpatient with various medical
illnesses. Psychotherapy and Psychosomatics. 45.
97-104.
Angold, A. (1988). Childhood and adolescent depression:
I. Epidemiological and aetiological aspects. British
Journal of Psychiatry. 152. 601-617.
Behrman, R. P., Vaughan, V. C. Ill, & Nelson, W. E.
(1987). Nelson's textbook of pediatrics (13th Ed.).
Philadelphia: W. C. Saunders Co.
Blotcky, A. D., Raczynski, J. M., Gurwitch, R., &
Smith, K. (1985). Family influences on hopelessness
among children early in the cancer experience.
Journal of Pediatric Psychology. 10. 479-493.
Boyd, J. H., & Weissman, M. M. (1981). Epidemiology of
affective disorders: A re-examination and future
directions. Archives of General Psychiatry. 38,
1039-1046.
Burke, P., Meyer, V., Kocoshis, S., Orenstein, D. M.,
Chandra, R., Nord, D. J., Sauer, J., & Cohen, E.
(1989). Depression and anxiety in pediatric
inflammatory bowel disease and cystic fibrosis.
Journal of the American Academy of Child &
Adolescent Psychiatry. 28, 948-951.
Cunningham, S. J., McGrath, P. J., Ferguson, H. B.,
Humphreys, P., D'Astous, J., Latter, J., Goodman, J.
T., & Firestone, P. (1987). Personality and
behavioural characteristics in pediatric migraine.
Headache. 27. 16-20.
86


BIOGRAPHICAL SKETCH
My parents were born and raised just outside of Havana,
Cuba, and they came to America in the 1960s looking for
a better place for themselves and their children. I
was born in Newark, New Jersey, but I spent most of my
life in Miami, Florida. I graduated from the
University of Miami and decided to pursue my graduate
work at the University of Florida. When I am not
working, I enjoy painting and crafts, writing and
reading, and the outdoors.
156


71
general child population (e.g., Angold, 1988; Kaslow &
Racusin, 1990) as well as with the pilot study results.
With regard to family composition, only the CDI
showed a small, nonsignificant difference, with
children of single-parent households obtaining somewhat
higher scores than children of two-parent households.
This suggests that there may be a tendency for children
to display more depressive symptoms because of the
stressors associated with single-parent homes.
However, given that this group difference was not
significant and that family composition differences
have not been reported in the general or pediatric
population with regard to depressive symptoms, this
finding may be spurious. Overall, the results
confirmed the pilot study findings and supported the
hypothesis that family composition does not
significantly affect behavioral distress in pediatric
populations.
No relationship between socioeconomic status and
distress was expected. However, the PIC-D and the
BUMPR-Home ratings were significantly positively
correlated with SES index. Thus, mothers in lower
socioeconomic levels reported more symptoms of
depression and distress in their children, possibly
reflecting the degree of stress associated with greater
financial hardship. This finding is consistent with


73
may not be significantly associated with the child's
behavioral upset or depression.
With regard to the amount of parent contact during
hospitalization, a negative correlation with child
distress had been anticipated. However, only the
Child-BUMP self-report scale was positively correlated
with the number of hours the parent spent with the
child. In other words, the more contact the child had
with the parent upon hospitalization, the more
behavioral upset the child reported. This finding
supports studies relating maternal presence to
increased negative behavior (e.g., Shaw & Routh, 1982).
Children may have interpreted high contact with a
parent during hospitalization as indicative of cause
for concern, or the children may have felt more
comfortable expressing their distress in their parents'
presence. Alternatively, those parents with more
distressed children may be more likely to spend more
time at their child's bedside in an attempt to
alleviate their distress.
In contrast, the remaining measures of child
distress were not associated with the amount of parent
contact, which supports earlier findings on
hospitalized children (Saylor et al., 1987). An
interesting and unexpected low positive correlation
between amount of parent contact and maternal state


65
Item
Loading
Item
12
Looks depressed or sad
(.71)
Item
11
Looks worried, tense
(.68)
Item
20
Has sleep problems
(.45)
Item
9
Sleeps unless directed into
activity (.43)
Internal Consistency Factor 3: Coefficient alpha = .68


33
corresponding to emotional distress at the hospital and
at home (See Appendix A). Patient behaviors are rated
by parents on a Likert scale ranging from 0 to 4. In
addition, the child's nurse completed the BUMP-R
questions for behavior observed in the hospital.
The Personality Inventory for Children. Depression
Scale (PIC-D; Lachar & Gdowski. 1979; Wirt, Lachar,
Klinedinst, & Seat. 1977). The PIC-D is one of 16
profile scales of the PIC, a comprehensive 600-item,
true-false parent rating measure of psychological
adjustment for children ages 3-16.
The Depression scale contains 46 rationally-
derived items. A sample item is:
My child is usually in good spirits True
False
These items yield ten interpretable factors, with
Brooding/Moodiness and Social Isolation accounting for
56% of the variance. The eight other factors are
Crying Spells, Lack of Energy, Pessimism/Anhedonia,
Concern with Death and Separation, Serious Attitude,
Sensitivity to Criticism, Indecisiveness/Poor
Self-Concept, and Uncommunicativeness. A Total
Depression scale score is obtained.
With respect to validity, the PIC-D correlated
significantly with all scales on the Conners Parent
Questionnaire except the Antisocial Scale and


35
I feel nervous and restless Not at all
Somewhat
Moderately so
Almost always
The following sample item is from the State Anxiety
Scale:
I feel strained Not at all
Somewhat
Moderately so
Almost always
Test-retest correlations for the Trait Anxiety
scale range from .73 to .86 for college students.
Stability on the State Anxiety scale, ranging from .16
to .62, reflects that state anxiety is influenced by
situational factors (Spielberger, 1983). With regard
to internal consistency, the median alpha coefficient
was reported as .93 for State Anxiety and .90 for Trait
Anxiety (Spielberger, 1983). The Trait Anxiety scale
has demonstrated concurrent validity with such measures
as the Taylor Manifest Anxiety Scale (Taylor, 1953) and
the scores on the state anxiety scale have been
correlated with situational stress (see Spielberger,
1983 for review).
Parents were instructed to complete the State
Anxiety items for how they felt at that moment, during
the evaluation. Then they were asked to complete the


47
Table 5: Demographic Differences for the CPI
N
Mean
: (SD)
t
Sex
Males
12
8.3
(5.0)
1.12
Females
21
6.2
(5.2)
Race
White
21
7.0
(5.2)
.28
African-American
10
6.4
(5.2)
Family Composition
2-Parent
21
5.4
(4.3)
1.99*
1-Parent
11
8.8
(5.1)
Note: All t values are nonsignificant.
E = .055


This girl doesnt like waiting.
$ DO YOU:
never like OR sometimes not
waiting like waiting
4
3
ITEM 1
This girl doesnt mind waiting.
DO YOU:
hardly ever OR never mind
mind waiting waiting
2
1


143


17
and depression were significantly higher in children
with recurrent abdominal pain than in healthy controls
(Walker & Greene, 1989). Another study of pediatric
cancer patients demonstrated a significant positive
correlation between parental trait anxiety and the
child's behavioral distress (Jay et al., 1983).
Furthermore, poor parental coping patterns were
associated with the depressive symptom of hopelessness
in children with cancer (Blotcky, Raczynski, Gurwitch,
& Smith, 1985). Thus, a mother's response to her
child's hospitalization may affect the child's
adjustment.
A concept related to maternal anxiety, whether the
parent is present as well as the duration of contact
during hospitalization, may also influence the
emotional adjustment of the hospitalized child. For
example, maternal presence was associated with more
negative behavior from children receiving injections
(Shaw & Routh, 1982). However, in another study, the
amount of time spent with a hospitalized child did not
predict depression or behavioral distress (Saylor et
al., 1987). Assuming that a factor in poor hospital
adjustment lies in separation, Peterson, Mori, and
Carter (1985) argue for the importance of encouraging
parent contact and enlisting parental assistance during
hospitalization. Although parental contact may ease


25
Psychometric data are limited due to the recency
of the measure's development. For the adult version,
internal consistency was reported as .93 and test-
retest reliability over variable intervals was reported
as .66. Four factors were identified from 213
inpatient adults, including behavioral regression, poor
patient-staff relationship, depression and anxiety, and
passivity and withdrawal (Zeldow & Braun, 1985).
Procedure. Participation was solicited the day
following the child's admission to the hospital.
Mothers were instructed to complete the BUMP-R in the
hospital room, rating their child's behavior since the
current hospitalization as well as rating their child's
behavior at home.
Results and Discussion
Analyses indicated that age, SES, and the BUMP-R
scores were normally distributed. However, the number
of previous hospitalizations and the duration of
illness were not normally distributed. Consequently,
analyses utilizing these two illness variables are
based on Spearman rho correlation coefficients.
Whereas no sex differences were found in the
BUMPR-Home rating of distress, a trend for girls (M =
29.8, SD = 11.4) to exhibit more distress in the
hospital than boys (M = 24.9, SD = 13.7) emerged (t =
1.67, p = .09). As expected, no race or family


MULTIMETHOD ASSESSMENT OF DEPRESSION AND BEHAVIORAL
DISTRESS IN A PEDIATRIC POPULATION
By
CHRISTINA M. RODRIGUEZ
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA


48
Table 6: Demographic Differences for the Nurse-BUMP
N
Mean fSD)
t
Sex
Males
17
31.3
(19.6)
.86
Females
15
25.5
(18.1)
Acre Group
4-8
16
32.2
(17.8)
1.08
8-12
16
25.0
(19.6)
Race
White
26
26.8
(17.8)
1.23
African-American
5
38.2
(24.7)
Family Composition
2-Parent
19
25.5
(17.6)
.97
1-Parent
11
32.4
(20.5)
Note: All t values are nonsignificant.


67
15.6) for the BUMPR-Hospital scores (t(46) = .54, g >
.05). Ratings on the BUMPR-Home showed a similar
pattern (t(46) = .57, p > .05), with fathers (M = 25.0,
SD = 12.0) obtaining scores comparable to mothers (M =
23.0, SD = 11.7).
For the 24 fathers involved in the pilot study and
the current study, no significant sex differences for
the children emerged on the BUMPR-Hospital rating
(t(22) = .80, g > .05) or for the BUMPR-Home rating
(t(22) = .34, g > .05). In addition, no age group
differences were found between 4-8 year old children
and 8-12 year old children for the BUMPR-Hospital
scores (t(22) = 1.03, g > .05) or for the BUMPR-Home
scores (t(22) = .43, g > .05). Age was also not
significantly correlated with the BUMPR-Hospital
ratings (r = -.25, g > .05) or with the BUMPR-Home
ratings (r = -.11, g > .05). Previous number of
hospitalizations was not associated with the
BUMPR-Hospital scores (r = .10, g > .05) or with the
BUMPR-Home scores (r = .36, g > .05). However,
duration of illness was correlated with the
BUMPR-Hospital ratings (r = -.49, g < .05) but not with
the BUMPR-Home ratings (r = .03, g > .05).
Similar to the earlier analyses of mothers,
fathers also rated hospital distress (M = 30.7, SD =
15.6) as significantly higher (t(22) = 2.38, g < .05)


57
Table 9: Item-Total Correlations for the BUMPR-Hospital
Item
1
.416
Item
15
. 332
Item
2
.484
Item
16
.563
Item
3
.466
Item
17
.328
Item
4
.568
Item
18
.491
Item
5
.562
Item
19
.333
Item
6
.453
Item
20
.343
Item
7
.290
Item
21
.109
Item
8
.398
Item
22
.518
Item
9
.113
Item
23
.486
Item
10
.242
Item
24
.269
Item
11
.385
Item
25
.376
Item
12
.435
Item
26
.516
Item
13
.511
Item
27
.301
Item
14
. 638
Item
28
. 346


56
Factor Analysis of BUMPR-Hospital
The BUMPR-Hospital scores obtained from the 151
mothers participating in the pilot study and the
current study were then further analyzed in order to
explore the psychometric characteristics of this
measure. Because the BUMPR-Home items were designed to
serve only as comparisons to the hospital items, the
BUMPR-Home scores alone would not address
hospitalization adjustment and thus were not further
analyzed. Correlations between individual items on the
BUMPR-Hospital rating and the total score are reported
in Table 9. Cronbach's coefficient alpha assessing
internal consistency of the total BUMPR-Hospital
measure was .87, suggesting that individual items are
strongly intercorrelated.
Factor analytic procedures were then performed on
the BUMPR-Hospital scores. Principal factors were
derived from the correlation matrix and the diagonal
was replaced by communality estimates. For these
communality estimates, Gorsuch (1983) maintains that,
instead of unities for the diagnonals, a more accurate
estimation of the communality for a given variable
utilizes the squared multiple correlation with all
other variables. To obtain the optimal factor
solution, factor rotation was performed using the
varimax orthogonal transformation, the promax oblique


hospital stay may have revealed differences in
adjustment.
77
Evaluation of the BUMP-R
The second purpose of this study was the
investigation of measures of distress particularly
suitable for pediatric populations. The Behavioral
Upset in Medical PatientsRevised designed by Saylor
and her colleagues (1987) appears to be a promising
parent report measure. The current study did not
reveal significant differences between ratings of
hospital and home distress, which did not support the
results obtained from the pilot study. With the data
from both the pilot and current study combined, a more
evenly distributed sample emerged. Maternal hospital
ratings in this combined sample were significantly
higher than home ratings. Moreover, fathers also
reported more behavioral upset in the hospital as
compared to the home. Therefore, the overall findings
suggest that hospitalization may precipitate emotional
distress.
The current study confirmed the pilot study's
strong positive correlation between the BUMPR-Hospital
scores and the BUMPR-Home scores. When the two samples
were combined, this relationship remained strong for
both mothers' and fathers' ratings. These results


21
Methods
Subjects. The sample consisted of 81 mothers of
hospitalized children (49 boys, 32 girls) from
consecutive pediatric admissions at Shands Hospital in
Gainesville, Florida. Ages of the children ranged from
4 years, 1 month to 11 years, 8 months (M = 6 years, 8
months; SD = 1 year, 11 months). The sample of
children was 73% White and 27% African-American; 64% of
the children lived in two-parent homes and 32% lived in
single-parent homes. The sample was predominantly
lower to middle SES (95%), based on the five levels of
social class as assessed by Hollingshead's (Myers &
Bean, 1968) Two Factor Index of Social Position (Class
1 = 3%; Class 2 = 2%; Class 3 = 17%; Class 4 = 40%;
Class 5 = 38%).
With respect to illness-related variables, 27% of
the children had never been previously hospitalized,
24% had one prior hospitalization, and the remaining
49% had multiple hospitalizations (ranging from 2 to
28), with the number of hospitalizations M = 6.1, SD =
15.8. Duration of illness ranged from diagnosis at
birth to diagnosis upon admission to the hospital, with
47% of the patients diagnosed within the past six
months.
Medical diagnoses obtained from mothers were
subsequently categorized into illness groups derived


CHAPTER 1
INTRODUCTION
General Childhood Depression
Prevalence
After years of controversy, the recent
proliferation of research suggests a growing acceptance
that depression exists in children. Several
comprehensive reviews of the literature estimate that
the diagnosis of major depression in childhood appears
in approximately 2% of 7- to 12-year-olds in the
general population (see Finch & Saylor, 1984; Kaslow &
Racusin, 1990; Kazdin, 1987, 1988 for reviews).
However, this estimate increases to 60% in outpatient
clinical populations (Kashani, Husain, Shekim, Hodges,
Cytryn, & McKnew, 1981). The wide variance in current
estimates of the prevalence of childhood depression are
primarily due to differences in assessment, definition,
and age level (e.g., Angold, 1988; Finch & Saylor,
1984). Prevalence estimates may vary according to age
level because depressive symptoms differ between
prepubertal children and adolescents (Ryan,
Puig-Antich, Ambrosini, Rabinovich, Robinson, Nelson,
Iyengar, & Twomey, 1987). In addition, because
depression is defined and assessed in a variety of ways
(Angold, 1988), different studies use different
1


68
than home distress (M = 25.0, SD = 12.0). Moreover,
the hospital distress score was also signicantly
correlated with the home distress score (r = .66, p <
.001).


18
the child during hospitalization, the conflicting
findings suggest that the role of parent contact is
either unclear or setting-specific.
The effect of diagnosis on adjustment to
hospitalization has not been studied closely in
pediatric populations. The literature on depression in
adult medical patients notes differences in prevalence
of symptoms as a function of medical diagnosis (e.g.,
von Ammon Cavanaugh, 1986). For instance, the highest
self-reported depression scores were found in patients
with gastrointestinal disease, cancer, bone and
connective tissue disease, renal disease, and
neurological disease. Few studies have investigated
this aspect in pediatric samples. One investigation
found no differences in behavioral distress due to type
of cancer diagnosis (Jay et al., 1983). Another study
comparing Crohn's disease, ulcerative colitis, and
cystic fibrosis found significantly greater prevalence
of depression in Crohn's disease compared to cystic
fibrosis (Burke, Meyer, Kocoshis, Orenstein, Chandra,
Nord, Sauer, & Cohen, 1989). However, research has not
fully examined the differential adjustment of children
based on the wide variety of medical diagnoses.
Previous medical experience may also influence the
pediatric patient's emotional adjustment to
hospitalization. A study of hospitalized children


101


155
Now that you're done with the family, find out who the
child's nurse is. Track her down and ask her how long
she has been working with the child. In order to
complete the form, she must have been working with the
child for at least eight hours or approximately one
shift. If she has, write down the estimated number of
hours of contact and give her the form to complete,
reading the instructions.


CHAPTER 3
RESULTS
Initially, this section focuses on the results
obtained from the examination and analysis of the data
collected from the 70 mothers, their children, and the
nurses involved in the current study. Differences in
demographic and illness-related variables were examined
as well as the relationships among the test scores.
Following this presentation, the information gathered
from mothers in this study were compared to the data
collected from the 81 mothers who participated in the
pilot study to highlight significant differences
between studies. Data from the mothers in the current
study and the pilot study were then combined and
analyzed to explore differences in demographic and
illness-related variables on the BUMP-R for this
larger, combined sample. In addition, a factor
analysis of the BUMPR-Hospital ratings was performed
for this combined sample.
Finally, analyses on a sample of fathers was also
conducted. This sample consisted of 13 fathers from
the current study and 11 fathers from the pilot study.
A matched sample of mothers was obtained for
comparison, in which eight of the fathers were compared
40


41
to their spouses who had also participated. For the
remaining 16 fathers, mothers were selected for
comparison based on the child's age and sex. Although
this matched sample cannot address correspondence of
ratings between mothers and fathers, the comparison
addresses whether fathers rate distress in their
children similar to mothers' ratings.
Analyses for Current Study
Descriptive Results & Statistics
Scores used for all analyses were the raw scores
of the BUMP-R, Child-BUMP, and CDI and the standard T-
Scores of the PIC-D and STAI. Inspection of skewness,
kurtosis, and histograms indicated that all of the
outcome measures were normally distributed. However,
the length of current hospitalization, the number of
prior hospitalizations, the duration of illness, and
the number of hours the parent spent with the child
were not normally distributed. Therefore, analyses
utilizing these variables were based on Spearman rho
coefficients. Moreover, given the number of
correlations computed, the significance level was
reduced to alpha=.01 for correlational analyses in
order to control for Type I errors.
Descriptive statistics on each test variable are
displayed in Table 1. No normative data is available


UNIVERSITY OF FLORIDA
3 1262 08554 8294


131


63
Table 12: Item Loadings for Factor OneNegativity/
Agitation
Item Loading
Item 18
Demanding
(.72)
Item 16
Stubborn, negativistic
(.71)
Item 4
Becomes upset easily
(.65)
Item 5
Is irritable or grouchy
(.65)
Item 1
Is impatient
(.64)
Item 3
Gets angry
(.59)
Item 14
Is uncooperative
(.58)
Item 19
Manipulative
(.55)
Item 13
Clinging, needs lots of reassurance
(.53)
Item 15
Complains
(.51)
Item 2
Cries
(.45)
Internal Consistency Factor 1: Coefficient alpha = .86


54
Given that these findings were not replicated in the
present study, a comparison of the two samples of
subjects was performed.
Distribution of race, family composition, SES,
duration of illness, number of prior hospitalizations,
and diagnostic grouping were comparable between
samples. However, the pilot study was composed of
predominantly males whereas the current study consisted
of primarily females, a significant gender difference
(X2 = 5.46, p = .01). Furthermore, the children in the
pilot study were significantly younger than the
children in the current study (F(150) = 23.86, p <
.0001).
Analyses of Combined Sample
The data obtained from maternal ratings on the
BUMP-R from both samples were combined and analyzed,
resulting in a more evenly distributed sample with
respect to sex and age. This combined sample consisted
of 151 mothers of 78 boys and 73 girls. Ages of the
children ranged from 4 years, 1 month to 12 years, 11
months (M = 7 years, 6 months, SD = 2 years, 5 months).
With this combined sample, no significant sex
differences were found in the mothers' ratings of
hospital distress (t(149) = .80, p > .05) or in the
ratings of home distress (t(149) = .70, p > .05).
Moreover, children aged 4-8 years obtained ratings


46
Table 4: Demographic Differences for the Child-BUMP
N
Mean (SD)
t
Sex
Males
29
52.0
(8.8)
.48
Females
41
51.0
(8.7)
Acre GrouD
4-8
32
50.7
(10.2)
. 65
8-12
38
52.0
(7.3)
Race
White
46
52.6
(8.6)
1.75
African-American
19
48.5
(8.5)
Family Composition
2-Parent
44
51.5
(8.3)
. 05
1-Parent
23
51.4
(9.1)
Note: All t values are nonsignificant.


CHAPTER 2
METHODS
Subjects
Seventy mothers and their children were recruited
from consecutive pediatric admissions at Shands
Teaching Hospital. The sample of children consisted of
29 males and 41 females between the ages of 4 years, 2
months and 12 years, 11 months (M = 8 years, 6 months,
SD = 2 years, 8 months). Racial composition was 66%
White, 27% African-American, and 7% Other. With regard
to school grade level, there were 10% of the children
not attending school, 4% in preschool programs, 12% in
kindergarten, 16% in the first grade, 11% in the second
grade, 10% in the third grade, 10% in the fourth grade,
13% in the fifth grade, 10% in the sixth grade, and 4%
in the seventh grade.
The family composition of the sample consisted of
66% of the children living in two-parent homes and 34%
of the children living in single-parent homes.
Although primarily lower to middle SES (90%), the
sample included all five levels of social class as
measured by Hollingshead's (Myers & Bean, 1968)
Two-Factor Index of Social Position (Class 1 = 3%;
Class 2 = 7%; Class 3 = 16%; Class 4 = 37%; Class 5 = 37%).
29


This girl doesnt look worried.
DO YOU:
never OR hardly ever
look look worried
worried
1
2
ITEM 10
This girl usually looks kind of worried.
DO YOU:
sometimes look OR always look
worried worried
3
4


113


72
the relationship found in a dental setting (Wright &
Alpern, 1971). Given that the remaining measures used
in this study that target distress in pediatric
populations did not manifest this relationship, perhaps
SES is not as influential in this group of children.
Maternal anxiety was predicted to be positively
correlated with child's distress. The findings
revealed that maternal reports of state and trait
anxiety were generally not related to the child's
behavioral upset. The only relationship that emerged
involved the PIC-D and maternal trait anxiety in that
mothers' reports of greater anxiety during the child's
hospital stay was associated with their reports of more
depressive symptoms in their children. This result may
be attributed to source bias because anxious mothers
may expect their children to display adjustment
difficulties upon hospitalization. The absence of
strong support for the relationship between children's
distress and maternal anxiety contrasts with several
previous studies (e.g., Blotcky et al., 1985; Jay et
al., 1983; Walker & Greene, 1989). Perhaps maternal
anxiety was not strongly related to child distress in
the present study because those previous studies
involved specific diagnostic groups (e.g., cancer,
recurrent abdominal pain). When pediatric patients
with diverse diagnoses are involved, parental anxiety


38
the scale are identical except for the sex of the child
in the picture and as phrased in the question.
Procedure
Parents of eligible patients were informed of the
study and participation was solicited by trained
undergraduate research assistants supervised by the
experimenter. Research assistants underwent a training
program involving several steps. Assistants were first
given a written description of general data collection
procedures (See Appendix D) that they rehearsed prior
to attending a 2-hour training session. During the
training session, the assistants watched a videotape
simulation of the experimenter administering the
measures to two children, a 5-year-old male and an
11-year-old female. Following the videotape, the
assistants role-played and discussed solving potential
difficulties that could be encountered during data
collection. After the training session, each assistant
observed the experimenter administering the measures to
at least one subject. Finally, the experimenter
observed each research assistant giving the measures
until the assistant demonstrated proficiency in the
data collection procedures.
Families were administered the questionnaires in
the hospital rooms the day following the child's


ITEM 5
This girl wont talk.
DO YOU:
not talk OR only sometimes
at all not talk
4
3
This girl doesnt mind talking.
DO YOU:
usually not OR
mind talking
never mind
talking


141


7
Racusin, 1990). Developmental differences in the
manifestation of depression may also lead to different
age groups exhibiting different features of depression
(Kazdin, 1987). For example, although Ryan et al.
(1987) found little variation in depressive
symptomatology between children and adolescents,
results did identify several features on which the age
groups differed (e.g., lethality of suicide attempts,
hopelessness, symptoms of anxiety).
In contrast to studies of school-aged and
adolescent children, depression in the lower end of the
age range, children under the age of 6, is studied
infrequently. Skepticism regarding depressive features
in preschoolers prevails despite clinical reports of
sadness and suicide attempts in children as young as
three years (e.g., Pfeffer & Trad, 1988). Although
research using preschool populations has been limited,
one study utilizing a multimethod assessment of normal
2- to 7-year-olds found 9 of 109 (8%) children
displaying depressive symptoms (Kashani, Holcomb, &
Orvaschel, 1986). Therefore, age differences in
prevalence may represent measurement variance given
that different assessment devices are administered to
different age groups (Kazdin, 1988).
Other demographic characteristics have been less
well studied. With regard to ethnicity, no clear


58
transformation, as well as the quartimax orthogonal
transformation. Comparison of the rotation methods
revealed that the promax oblique rotation yielded
factor structures of substantial factor complexity and
thus complicated interpretability. Thus, orthogonal
solutions were preferred over the promax oblique
solution which assumes intercorrelated factors. The
quartimax orthogonal rotation was selected in favor of
the varimax orthogonal rotation because Gorsuch (1983)
advises that the varimax method is inappropriate when
the measure has high internal consistency. The
quartimax rotation assumes that a single general factor
accounts for a substantial amount of the variance,
which is implied with high internal consistency,
whereas the varimax rotation does not assume one large
factor.
The number of factors extracted was based on three
criteria. As Gorsuch (1983) described, Guttman
recommended examination of the characteristic roots
(eigenvalues) for factor extraction. The number of
eigenvalues greater than one estimates the number of
factors. Based on this criterion, four factors were
extracted (See Table 10), which accounted for 85% of
the variance.


I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Sheila M. Eyberg', Chair
Professor of
Clinical and Health Psychology
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Stephen R. Boggs, Cochair
Assistant Professor of
Clinical and Health Psychology
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy. ^ (]/
James R. Rodric
Assistant Professor of
Clinical and Health Psychology
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully, adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Assistant Professor of
Clinical and Health Psychology
Michael E. Geisser


51
outcome measures. In contrast to the pilot results and
the hypothesis, there was no significant difference
between the BUMPR-Hospital and BUMPR-Home score (t(70)
= 1.73, p > .05).
Correlational Analyses
The correlation matrix in Table 8 indicates that,
as expected, many of the outcome variables were
intercorrelated. Consistent with the pilot study, the
BUMPR-Hospital ratings were significantly positively
correlated with the BUMPR-Home ratings. In addition,
the BUMPR-Hospital scores were also significantly
positively associated with the PIC-D, Child-BUMP, and
CDI. However, the parent ratings of hospital distress
on the BUMP-R was only marginally positively correlated
with the Nurse-BUMP. The parent report of distress on
the BUMPR-Home form was significantly positively
correlated with the PIC-D and CDI. Moreover, the
parent ratings on the PIC-D were significantly
positively correlated with the CDI and parental trait
anxiety. Parental trait and state anxiety were also
positively correlated. With regard to demographic
variables, SES was significantly positively correlated
with the BUMPR-Home rating and the PIC-D scores.
The BUMPR-Hospital ratings and the Nurse-BUMP
ratings were not significantly correlated with parent
report of state or trait anxiety. In addition, the


22
from Nelson's Textbook of Pediatrics (Behrman, Vaughan,
& Nelson, 1987) with 79% of the sample falling into
one of six groups. Twenty percent of those children
classified were diagnosed with cardiovascular or
respiratory illnesses (e.g., pneumonia, tetralogy of
fallot, cystic fibrosis). Fourteen percent were
hospitalized for immunity, allergy, or related diseases
(e.g., asthma, reflex sympathetic dystrophy). Another
17% were experiencing illnesses interfering with the
digestive system (e.g., Hirshsprung's disease,
inflammatory bowel disease, appendicitis). Nineteen
percent of the patients were diagnosed with urinary
problems (e.g., reflux, nephrotic syndrome, kidney
infection). An additional 17% were classified with
diseases of the nervous system (e.g., spina bifida,
Hallervorden spatz, hydrocephalus). Diagnoses had not
been identified for another 13% of the children. The
seventeen patients not categorized into the above
groups had been diagnosed with several different
illnesses (e.g., burns, autism, leg fracture) and did
not conform with the six illness groups.
Furthermore, descriptions of the clinical course
of illnesses were obtained form Nelson's Textbook of
Pediatrics (Behrman et al., 1987) to identify diagnoses
based on chronicity. Sixty-eight children in the
sample were thus classified, with 28% of these


153
If the child can participate, then say:)
We are also hoping to get more information before your
child is discharged. If your child is still here for
at least another day we'll try to catch you before you
leave and ask you and your child to do the
questionnaires again. We hope that with this project
we can learn which children are at risk to develop
problems in the hospital. We are also trying to find
better tests to use to make it easier to help
hospitalized children in the future. All information
will be kept confidential. Would you be interested in
participating?
(Encourage them to participate as much as possible.
Make sure to check with child that they are also
interested in participating.
If they agree, then give them the informed consent
form, skimming and highlighting the sections of the
form. On the last page, ask the parent to sign it and,
if the child is able and is between 7-17, have them
sign it. Then you sign as witness. Give them a copy
of the consent form and tell them that, should they
have any questions, they can contact me at the phone
number on the form.
Assign the family a subject number, write this number
on the consent form and remaining forms, and administer
the background information questions.
At this point, assess the parent's literacy. If the
parent has less than a tenth grade education, read the
questions aloud immediately after getting the
background information form. If the parent has a
higher educational level, then give them the three
questionnaires, explaining the directions for each one.
Then, go to the first questionnaire and say:)
Okay. Why don't we start with the first question.
Which answer seems best for how your child has reacted
since they arrived?
(Watch them complete the first three or four questions
and observe if they are experiencing any difficulty.
If they take too long, say in a easy voice:)
Well, it's probably just faster/easier for me to just
read them and I'll check off your answers. Does ...


This girl usually does what others want.
DO YOU:
always
do what
others
want
OR sometimes
do what
others
want
1
2
This girl doesnt do what others want.
DO YOU:
hardly ever OR
do what
others
want
never
do what
others
want


9
encompasses a broad range of symptoms, including
general subjective distress, dysphoric mood, behavioral
manifestations of emotional distress, and anxiety.
This wide range of symptoms reveals the difficulties
involved due to definitional variability. Many of the
symptoms studied are features typically associated with
depression, whereas other symptoms are externalizing
responses associated with general emotional or
psychological distress. Several studies have
demonstrated a relationship between externalizing
behavior problems and depression (e.g., Leon et al.,
1980). Therefore, adjustment problems in pediatric
populations are examined in a variety of ways.
Historically, psychologists have been interested
in children's adjustment to medical procedures and
hospitalization. A child's short-term reaction to
medical intervention may vary widely. These depressive
reactions can include restlessness, apathy, and sleep
and appetite disturbances (Jessner, Blom, & Waldfogel,
1952), panic and crying (Prugh, Staub, Sands,
Kirschbaum, & Lenihan, 1953), as well as anger and
aggression (Jensen, 1955). Additionally, a large scale
follow-up study compared 1000 children hospitalized
before age 5 to a nonhospitalized control group
(Douglas, 1975). Results indicated that the
hospitalized children exhibited numerous difficulties


Table 13: Item Loadings for Factor TwoAmiability
Item Loading
Item 22
Tries to be friendly
(.67)
Item 23
Accepts advice or instructions easily
(.64)
Item 26
Pleasant to be with
(.64)
Item 25
Laughs or smiles at funny comments
or events
(.56)
Item 27
Shows interest in recovery (takes
initiative)
(.55)
Item 24
Starts conversation
(.54)
Item 28
Does what he or she is told
(.49)
Item 21
Able to ask for help
(.40)
Internal Consistency Factor 2: Coefficient alpha = .79


MULTIMETHOD ASSESSMENT OF DEPRESSION AND BEHAVIORAL
DISTRESS IN A PEDIATRIC POPULATION
By
CHRISTINA M. RODRIGUEZ
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA

ACKNOWLEDGMENTS
First and foremost, I must thank two individuals
who helped transform an idea into reality: Steve Boggs,
for his guidance, incredible patience, unwavering
support, and long hours; and Sheila Eyberg, as a female
role model of unparalleled professionalism as well as
for her steadfast enthusiasm and faith. Also, I thank
Jim Rodrigue, Mike Geisser, and Faye Gary-Harris, whose
excitement about this project energized me and whose
suggestions, comments, and questions facilitated the
direction and critical evaluation of the project. I
thank all of these committee members for helping me
pull this together.
Secondly, I thank my mother not only for her
invaluable assistance in creating the Child-BUMP
pictures but also for instilling in me a deep
compassion for human suffering that led me to the field
of clinical psychology. I also thank Becky, Brigette,
Elena, Jennifer, Melodye, and Randi, my close friends
and fellow graduate school survivors, for the emotional
support that sustained me and for redefining the depths
of friendship. I also appreciate the help of other
friends, faculty, and staff at the Dept, of Clinical
ii

and Health Psychology, whose daily contributions to my
well-being and training made accomplishing this project
meaningful. I thank the group of research assistants
who helped with data collection and turned what
initially seemed to be an insurmountable daily task
into an organized demonstration of teamwork. Finally,
I thank the parents and nurses who took the energy from
their pressing concerns in order to help the sons and
daughters of tomorrow.
iii

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS ii
LIST OF TABLES vi
ABSTRACT V
CHAPTERS
1 INTRODUCTION 1
General Childhood Depression 1
Prevalence 1
Assessment Issues 2
Risk Factors 4
Pediatric Population 8
Definition Issues 8
Prevalence 10
Assessment Issues 11
Risk Factors 14
Pilot Study 2 0
Introduction and Rationale 2 0
Methods 21
Results and Discussion 25
Purpose of Study and Hypotheses 27
2 METHODS 29
Subjects 29
Measures 3 2
Procedure 38
3 RESULTS 4 0
Analyses for Current Study 41
Descriptive Results & Statistics 41
Analyses of Background Data 4 3
Correlational Analyses 51
iv

Analyses of Current Study and Pilot Study
Combined 53
Comparison of Samples 53
Analysis of Combined Sample 54
Factor Analysis of BUMPR-Hospital 56
Analyses of Fathers for Combined Sample 61
4 DISCUSSION 69
Background Variables Affecting Distress 69
Evaluation of the BUMP-R 77
Evaluation of the Child-BUMP 80
Implications 81
REFERENCES 8 6
APPENDICES
A BEHAVIORAL UPSET IN MEDICAL PATIENTSREVISED
(BUMP-R) 94
B BACKGROUND INFORMATION SHEET 9 6
C BEHAVIORAL UPSET IN MEDICAL PATIENTS-CHILD
SELF-REPORT VERSION (CHILD-BUMP) 97
D DATA COLLECTION PROCEDURES TRAINING GUIDE 152
BIOGRAPHICAL SKETCH 156
v

LIST OF TABLES
TABLE Page
1Means and Standard Deviations of the Outcome
Measures 42
2 Demographic Differences for the BUMPR-Hospital.. 44
3 Demographic Differences for the PIC-Depression.. 45
4 Demographic Differences for the Child-BUMP 46
5 Demographic Differences for the CDI 47
6 Demographic Differences for the Nurse-BUMP 48
7 Spearman Correlations Between Illness-Related
Variables and Outcome Measures 50
8 Correlations Among Outcome Measures and
Demographic Variables 52
9 Item-Total Correlations for the BUMPR-Hospital.. 57
10 Eigenvalues for the BUMPR-Hospital Factor
Analysis 59
11 Factor Structure for the BUMPR-Hospital 62
12 Item Loadings for Factor OneNegativity/
Agitation 63
13 Item Loadings for Factor TwoAmiability 64
14 Item Loadings for Factor ThreeDysphoria 65
15Item Loadings for Factor FourNoncompliance.... 66
vi

Abstract of Dissertation Presented to the
Graduate School of the University of Florida in
Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
MULTIMETHOD ASSESSMENT OF DEPRESSION AND BEHAVIORAL
DISTRESS IN A PEDIATRIC POPULATION
By
Christina M. Rodriguez
August 1993
Chair: Sheila Eyberg, Ph.D.
Cochair: Stephen R. Boggs, Ph.D.
Major Department: Clinical and Health Psychology
Growing acceptance of the existence of childhood
depression is apparent in recent literature. Despite
this scrutiny, research on childhood depression
contains mixed results largely because of assessment
issues. Studies obtain information from various
sources (e.g., parents, children) and utilize different
definitions of depression. Assessment of preschool
children is further complicated because of skepticism
about depression at this age and because of the
scarcity of measures appropriate for this age group.
The study of depression in pediatric psychology
translates into adjustment in medical settings.
Although pediatric populations seem more likely to
exhibit depressive symptoms, measures suitable for
vii

these children are scarce. The utility of a new parent
report measure, the Behavioral Upset in Medical
Patients-Revised (BUMP-R), was examined in a pilot
study of 81 mothers, which found increased behavioral
distress upon hospitalization. Research on variables
influencing adjustment has found that gender, race,
socioeconomic status, and family composition are not
significant factors. Greater distress may appear with
younger children, maternal anxiety, and limited parent
contact during hospitalization. Effects of diagnosis,
duration of illness, length of hospitalization, and
previous hospitalizations are unclear.
The current study explored variables influencing
depressive symptoms in hospitalized children aged 4-12.
Parent ratings of child distress were compared to nurse
ratings and to children's responses to a pictorial
measure designed in this study for hospitalized
preschoolers. The day following hospital admission, an
assessment battery containing standard measures of
depression and measures of distress for hospitalized
children was administered to 70 mothers and their
children. Thirty-two nurse ratings were also obtained.
Results indicated that demographic and illness-
related variables were not risk factors for hospital
adjustment difficulties. Based on parent ratings,
children exhibiting behavioral distress at home may be
viii

more likely to experience adjustment problems upon
hospitalization. Concerning assessment, the BUMP-R
demonstrated internal consistency and concurrent
validity. A factor analysis of mothers from the
current study combined with the pilot study revealed
four factors identified as Negativity/Agitation,
Amiability, Dysphoria, and Noncompliance. The child
pictorial measure also demonstrated internal
consistency and correspondence with mothers' ratings.
Therefore, both measures are promising for use with
hospitalized children.
ix

CHAPTER 1
INTRODUCTION
General Childhood Depression
Prevalence
After years of controversy, the recent
proliferation of research suggests a growing acceptance
that depression exists in children. Several
comprehensive reviews of the literature estimate that
the diagnosis of major depression in childhood appears
in approximately 2% of 7- to 12-year-olds in the
general population (see Finch & Saylor, 1984; Kaslow &
Racusin, 1990; Kazdin, 1987, 1988 for reviews).
However, this estimate increases to 60% in outpatient
clinical populations (Kashani, Husain, Shekim, Hodges,
Cytryn, & McKnew, 1981). The wide variance in current
estimates of the prevalence of childhood depression are
primarily due to differences in assessment, definition,
and age level (e.g., Angold, 1988; Finch & Saylor,
1984). Prevalence estimates may vary according to age
level because depressive symptoms differ between
prepubertal children and adolescents (Ryan,
Puig-Antich, Ambrosini, Rabinovich, Robinson, Nelson,
Iyengar, & Twomey, 1987). In addition, because
depression is defined and assessed in a variety of ways
(Angold, 1988), different studies use different
1

2
criteria for establishing prevalence in a given
population.
Assessment Issues
The research on general childhood depression
contains numerous conflicting results. Much of the
variability in estimating prevalence and in estimating
the influence of variables may be attributed to
difficulties in assessment. The definitions of
depression determine the selection of assessment
techniques (Angold, 1988). Although several studies
utilize diagnostic criteria for depression (see Finch &
Saylor, 1984 for review), the majority of studies
define depression based on scores obtained on a variety
of measures. However, many inventories gain only
temporary popularity and are utilized in only a single
study (Kerr, Holer, & Versi, 1987). Thus, depression
is interpreted differently, which complicates questions
regarding the influence of numerous variables.
With adults, assessment of depression is often
based on self-report or interview data. However,
different assessment methods are conducted at different
ages. For instance, assessment of adolescents, like
adult assessment, often utilizes self-report and
interview methods. However, investigation of affective
disturbance in preschool children is particularly
challenging. A preschool child is not readily able to

3
distinguish emotions (Stone & Lamenek, 1990). Of those
socioemotional measures available for assessment of
preschool children, self-report measures are rare
(Lichtenstein, Dreger, & Cattell, 1986). Consequently,
preschool assessment relies heavily on parent report.
For school age children, a larger variety of assessment
techniques for depression are utilized, including
self-report questionnaires, interview rating scales,
parent ratings, and observational measures (see
examples of observational measures in dental
populations, Johnson & Baldwin, 1968 and Wright &
Alpern, 1971; in medical populations, see Jay, Ozolins,
Elliot, & Caldwell, 1983 and Katz, Kellerman, & Siegel,
1980). Rating scales are completed by a variety of
sources, including parents, peers, and teachers.
Given the variability in sources, correspondence
among raters becomes a significant concern. A
comprehensive review of correspondence issues in
measurement of both behavioral and emotional childhood
problems found relatively low correspondence among
raters (Achenbach, McConaughy, & Howell, 1987).
Several other studies have documented low correlations
between self-reported depression and parent ratings of
depression (e.g., Kazdin, 1989; Kazdin, French, Unis, &
Esveldt-Dawson, 1983; see Kaslow & Racusin, 1990;
Kazdin, 1987, 1988 for reviews). A few studies have

4
demonstrated adequate correlations between raters
(e.g., in normal children, Leon, Kendall, & Garber,
1980; in pediatric population, Eason, Finch, Brasted, &
Saylor, 1985). Despite the general finding of low or
poor correspondence, most researchers in the area of
depression conclude that multiple sources are required
in order to clarify the diagnostic picture (e.g.,
Kaslow & Racusin, 1990; Kazdin, 1988). In fact,
Kazdin, Colbus, & Rodgers (1986) describe the
application of discriminant analyses which combines a
battery of measures to maximize the classification,
i.e., diagnosis, of depression in children. Therefore,
in spite of the poor correspondence between raters,
current research encourages assessment of childhood
depression from multiple sources.
Risk Factors
In addition to variability due to assessment
methods, a number of variables may affect the
prevalence of childhood depression, including gender
and age of the child and demographic characteristics,
such as race, socioeconomic status, and family
composition. A fair amount of research has
investigated the effect of sex on depression. In
childhood depression, no gender differences emerge for
children ages 6-12 (Angold, 1988; Kaslow & Racusin,
1990). The prevalence, however, increases in females

5
throughout adolescence and approaching adulthood
(Kazdin, 1987, 1988). Consequently, the sex ratio of
depression for females to males nears 2:1 by adulthood
(Boyd & Weissmann, 1981).
Nevertheless, for prepubertal children, several
studies utilizing various assessment instruments have
corroborated the 1:1 sex ratio. Among these studies,
scores on a clinician-administered interview rating
scale revealed no sex differences (Shanahan,
Zolkowski-Wynne, Coury, Collins, & O'Shea, 1987). In
addition, normative data on over 1400 schoolchildren
grades 2-8 for one of the most widely-used self-report
depression scales, the Children's Depression Inventory
(CDI; Kovacs, 1983, 1985), found negligible gender
differences, and thus separate norms were not
recommended (Finch, Saylor, & Edwards, 1985). This
finding was substantiated when no gender differences on
the CDI were found in a smaller sample of 166 children
grades 3-6 (Reynolds, Anderson, & Bartell, 1985). A
study of parent-reported childhood depression also
reflects this equal prevalence for children grades 3-6
(Leon et al., 1980; Reynolds et al., 1985).
Furthermore, no statistically significant sex
differences were found on a measure of peer-nominated
depression (Lefkowitz & Tesiny, 1980).

6
Only a few studies have demonstrated sex
differences in depressive symptomatology for
prepubertal children. For instance, minor differences
were found in which girls scored higher than boys on a
different self-report measure (Children's Depression
Scale; Reynolds et al., 1985). A study of 8- to 13-
year-olds also found a trend for girls to report more
depressive symptoms than boys on the CDI (Seligman,
Peterson, Kaslow, Tanenbaum, Alloy, & Abramson, 1984).
These conflicting findings may reflect variability in
assessment or in populations, or they may indicate that
observers respond differentially based on the child's
sex (Saylor, Finch, Spirito, & Bennett, 1984). Despite
these mixed results, the bulk of the literature
continues to support the 1:1 sex ratio for depression
in prepubertal children.
As mentioned above, an interaction between sex and
age appears, leading to an increased incidence of
depression in adolescent girls (Kazdin, 1987, 1988). A
significant interaction effect between gender and age
was found in a study of over 1000 schoolchildren, with
adolescent girls reporting the most unhappiness in a
structured interview (Webb & VanDevere, 1985).
Although depressive disorder is relatively uncommon in
prepubertal children, the prevalence of depression
rises with increasing age (Angold, 1988; Kaslow &

7
Racusin, 1990). Developmental differences in the
manifestation of depression may also lead to different
age groups exhibiting different features of depression
(Kazdin, 1987). For example, although Ryan et al.
(1987) found little variation in depressive
symptomatology between children and adolescents,
results did identify several features on which the age
groups differed (e.g., lethality of suicide attempts,
hopelessness, symptoms of anxiety).
In contrast to studies of school-aged and
adolescent children, depression in the lower end of the
age range, children under the age of 6, is studied
infrequently. Skepticism regarding depressive features
in preschoolers prevails despite clinical reports of
sadness and suicide attempts in children as young as
three years (e.g., Pfeffer & Trad, 1988). Although
research using preschool populations has been limited,
one study utilizing a multimethod assessment of normal
2- to 7-year-olds found 9 of 109 (8%) children
displaying depressive symptoms (Kashani, Holcomb, &
Orvaschel, 1986). Therefore, age differences in
prevalence may represent measurement variance given
that different assessment devices are administered to
different age groups (Kazdin, 1988).
Other demographic characteristics have been less
well studied. With regard to ethnicity, no clear

8
racial differences have been found in depressed
prepubertal children (Angold, 1988; Kaslow & Racusin,
1990). Although a few studies have suggested that
depression may be associated with lower socioeconomic
status (see Angold, 1988 for review), depressive
symptomatology was not found to be related to father's
education or family income (Kandel & Davies, 1982).
Finally, no differences in socioeconomic status or
family composition (i.e., child living in single-parent
home, two-parent home) were found between depressed or
nondepressed clinic children (Kaslow, Rehm, Pollack, &
Siegel, 1988). The effects of these demographic
variables on childhood depression, however, are seldom
explored.
Pediatric Populations
Definition Issues
Given the relationship between stressful life
events and depressive symptoms (e.g., Mullins, Siegel,
& Hodges, 1985), pediatric psychology has evidenced a
growing interest in depressive symptoms in medical
populations. Few studies have examined the prevalence
of major depression in pediatric populations. In
contrast, several studies have focused on adjustment in
medical settings. The depressive symptoms associated
with poor adjustment in pediatric populations

9
encompasses a broad range of symptoms, including
general subjective distress, dysphoric mood, behavioral
manifestations of emotional distress, and anxiety.
This wide range of symptoms reveals the difficulties
involved due to definitional variability. Many of the
symptoms studied are features typically associated with
depression, whereas other symptoms are externalizing
responses associated with general emotional or
psychological distress. Several studies have
demonstrated a relationship between externalizing
behavior problems and depression (e.g., Leon et al.,
1980). Therefore, adjustment problems in pediatric
populations are examined in a variety of ways.
Historically, psychologists have been interested
in children's adjustment to medical procedures and
hospitalization. A child's short-term reaction to
medical intervention may vary widely. These depressive
reactions can include restlessness, apathy, and sleep
and appetite disturbances (Jessner, Blom, & Waldfogel,
1952), panic and crying (Prugh, Staub, Sands,
Kirschbaum, & Lenihan, 1953), as well as anger and
aggression (Jensen, 1955). Additionally, a large scale
follow-up study compared 1000 children hospitalized
before age 5 to a nonhospitalized control group
(Douglas, 1975). Results indicated that the
hospitalized children exhibited numerous difficulties

10
corresponding to early hospitalization, including
conduct problems, delinquency, academic difficulty, and
unstable work history. Another study of pediatric
surgery patients found that, without intervention,
children displayed more behavioral difficulties nearly
one month after hospitalization (Melamed & Siegel,
1975). This early research demonstrated that children
often adjust poorly to hospitalization or medical
procedures.
Prevalence
The study of poor adjustment to hospitalization is
susceptible to similar difficulties in estimating
prevalence as those difficulties in general childhood
depression. Different age groups, definitions, and
assessment techniques are utilized. Although one study
found nearly 90% of pediatric surgery patients
displayed behavioral difficulties following
hospitalization (Prugh et al., 1953), most estimates of
depressive symptoms in pediatric populations are more
conservative. In children with physical handicaps or
chronic illness, poor adjustment is reported as ranging
from 13-26% based on parent report, approximately two
times the rate found in healthy children (Wallander,
Varni, Babani, Banis, & Wilcox, 1988). In a review of
studies using diagnostic criteria (derived from DSM-III
criteria), estimates of depressive symptoms in both

11
inpatient and outpatient pediatric samples ranged from
7-40%, with the wide variability possibly due to
different age ranges and different medical populations
(Finch & Saylor, 1984). In a study of 7- to 12-year-
old hospitalized children, 38% exhibited dysphoric mood
(Kashani, Barbero, & Bolander, 1981). An examination
of a pediatric psychology service in a children's
hospital found 19% of consultations were referred for
depression or suicide attempts and 12% of the
consultations were referred for adjustment to chronic
illness (Olson, Holden, Friedman, Faust, Kenning, &
Mason, 1988). Thus, pediatric populations may be more
likely to display depressive symptoms than the general
child population.
Assessment Issues
As mentioned above, the assessment of pediatric
patients' adjustment involves symptoms from overlapping
constructs and may include symptoms of major
depression, dysphoric mood, behavioral concomitants of
emotional upset, and symptoms related to anxiety.
Assessment in pediatric populations shares difficulties
similar to the assessment of depression in the general
population. Problems with correspondence among raters
appear in the pediatric psychology literature as well
as in the general childhood depression literature.
Rating scales are typically administered to parents,

12
teachers, and health personnel (see Katz et al., 1980
for example of nurse ratings in pediatric populations).
Compounding the usual assessment issues, the
evaluation of emotional distress in pediatric
populations is further complicated by the scarcity of
measures appropriate for the assessment of these
children. The application of standard measures of
depressive symptoms, such as the CDI, may not be
suitable for hospitalized children. For example, the
CDI did not differentiate between a sample of chronic
headache sufferers and a non-headache control group
(Wisniewski, Naglieri, & Mulick, 1988). Similarly,
pediatric inpatients attained CDI scores comparable to
gender and age-matched schoolchildren (Saylor, Finch, &
McIntosh, 1988). In fact, one study found that normal
children tended to report more depression on the CDI
than cancer patients, which the researchers suggest
reflects denial of symptoms (Worchel, Nolan, Willson,
Purser, Copeland, & Pfefferbaum, 1988). In contrast,
another study found that children with recurrent
abdominal pain scored significantly higher on the CDI
than healthy children (Walker & Greene, 1989) These
mixed results with the CDI question its utility in
pediatric populations.
Likewise, a clinician-administered structured
interview, the Children's Depression Rating Scale

13
(CDRS; Poznanski, Cook, & Carroll, 1979), appears to be
as limited as the CDI. A study of migraine patients
found no difference in CDRS scores compared to a
control group of children (Cunningham, McGrath,
Ferguson, Humphreys, D'Astous, Latter, Goodman, &
Firestone, 1987). A closer examination of the CDRS
administered to pediatric cancer patients found
significant overlap between depressive symptoms and
impairment due to illness (Heilgenstein & Jacobsen,
1988). Thus, the authors found that measures which
include somatic symptoms may overestimate the presence
of depression.
Instruments which target emotional upset in
pediatric populations are few in number. The
Observation Scale of Behavioral Distress (Jay et al.,
1983) interprets behavioral distress as behaviors
indicative of anxiety and pain. Similarly, an
observational pain rating scale for children aged 2-6
includes some ''depression-like'' items (Gauvain-Piquard,
Rodary, Rezvani, & Lemerle, 1987). However, both
observation rating scales were based on cancer patients
and both scales feature an emphasis on pain behavior
rather than on depressive symptoms. In contrast, a
parent rating scale, the Behavioral Upset in Medical
PatientsRevised (BUMP-R; Saylor, Pallmeyer, Finch,
Eason, Trieber, & Folger, 1987), focuses on emotional

14
distress and was devised for hospitalized children with
a variety of diagnoses.
Risk Factors
Factors which influence the prevalence of
childhood depression in the general population may also
affect pediatric populations. With respect to gender
differences, results have been mixed. Consistent with
findings in the general childhood depression
literature, a study of pediatric cancer patients ages
2-20 found no gender differences on an observational
measure of behavioral distress during a painful medical
procedure (Jay et al., 1983). Moreover, no sex
differences were found in hospitalized pediatric
patients ages 5-15 on self-reported depression,
interview-rated depression, or parent-reported
behavioral upset (Saylor et al., 1987). One study of
hospitalized children resulted in a greater number of
boys diagnosed with depression, although this subsample
was small (Kashani, Barbero, & Bolander, 1981). Some
studies have found girls demonstrated more behavioral
distress during painful medical procedures (e.g., Katz
et al., 1980; Melamed & Siegel, 1975). Although the
literature suggests that girls may exhibit more
distress, evidence for overall gender differences in
depressive symptoms for pediatric populations is
limited.

15
The effect of age on the adjustment of pediatric
patients contrasts with findings in the general
childhood depression literature. In the general
population, the prevalence of depressive symptomatology
increases through childhood and adolescence. However,
in a study conducted with cancer patients, age was the
strongest predictor of distress during medical
procedures, with younger children exhibiting greater
distress (Jay et al., 1983). Another study of a
pediatric cancer population aged 1-17 also found
younger children expressed more distress during a
painful medical procedure (Katz et al., 1980).
Additional support for the inverse relationship between
age and distress was found when younger hospitalized
children were rated by parents as more distressed
(Saylor et al., 1987). Furthermore, an interaction
between sex and age may occur. In contrast to the
trend in general child depression for adolescent
females to demonstrate more depressive symptoms,
younger females may exhibit more distress in pediatric
settings (e.g., Melamed & Siegel, 1975). Overall,
younger pediatric patients appear more likely to
display behaviors suggestive of depressive symptoms and
emotional distress.
Information on the influence of other demographic
variables (such as race, socioeconomic status, or

16
family composition) on adjustment in pediatric
populations is even more limited than that found in
research on general childhood depression. Of the few
studies available, appropriate behavior in a dental
setting was associated with socioeconomic status, with
upper SES children exhibiting less negative behavior
(Wright & Alpern, 1971). Influence of race and family
composition are not known and thus may correspond to
data obtained on general childhood depression.
Consequently, future research should explore the
influence of these demographic variables on the
emotional adjustment of pediatric patients and whether
these influences correspond to those found in general
childhood depression.
Several other variables related to pediatric
populations may also affect adjustment to
hospitalization. These variables include maternal
anxiety, parental presence in the hospital, diagnosis,
prior medical experience, onset of illness, and length
of hospitalization. With regard to maternal anxiety,
mothers' self-report of anxiety was positively related
to their children's negative behavior during dental
visits (Johnson & Baldwin, 1968; Wright & Alpern,
1971). Moreover, mother's state anxiety was positively
correlated with a physiological measure of anxiety in
pediatric patients (Vardaro, 1978). Maternal anxiety

17
and depression were significantly higher in children
with recurrent abdominal pain than in healthy controls
(Walker & Greene, 1989). Another study of pediatric
cancer patients demonstrated a significant positive
correlation between parental trait anxiety and the
child's behavioral distress (Jay et al., 1983).
Furthermore, poor parental coping patterns were
associated with the depressive symptom of hopelessness
in children with cancer (Blotcky, Raczynski, Gurwitch,
& Smith, 1985). Thus, a mother's response to her
child's hospitalization may affect the child's
adjustment.
A concept related to maternal anxiety, whether the
parent is present as well as the duration of contact
during hospitalization, may also influence the
emotional adjustment of the hospitalized child. For
example, maternal presence was associated with more
negative behavior from children receiving injections
(Shaw & Routh, 1982). However, in another study, the
amount of time spent with a hospitalized child did not
predict depression or behavioral distress (Saylor et
al., 1987). Assuming that a factor in poor hospital
adjustment lies in separation, Peterson, Mori, and
Carter (1985) argue for the importance of encouraging
parent contact and enlisting parental assistance during
hospitalization. Although parental contact may ease

18
the child during hospitalization, the conflicting
findings suggest that the role of parent contact is
either unclear or setting-specific.
The effect of diagnosis on adjustment to
hospitalization has not been studied closely in
pediatric populations. The literature on depression in
adult medical patients notes differences in prevalence
of symptoms as a function of medical diagnosis (e.g.,
von Ammon Cavanaugh, 1986). For instance, the highest
self-reported depression scores were found in patients
with gastrointestinal disease, cancer, bone and
connective tissue disease, renal disease, and
neurological disease. Few studies have investigated
this aspect in pediatric samples. One investigation
found no differences in behavioral distress due to type
of cancer diagnosis (Jay et al., 1983). Another study
comparing Crohn's disease, ulcerative colitis, and
cystic fibrosis found significantly greater prevalence
of depression in Crohn's disease compared to cystic
fibrosis (Burke, Meyer, Kocoshis, Orenstein, Chandra,
Nord, Sauer, & Cohen, 1989). However, research has not
fully examined the differential adjustment of children
based on the wide variety of medical diagnoses.
Previous medical experience may also influence the
pediatric patient's emotional adjustment to
hospitalization. A study of hospitalized children

19
found that the total number of days spent in previous
hospitalizations predict higher parent ratings of their
child's emotional distress during the current
hospitalization (Saylor et al., 1987). In contrast,
Jay et al. (1983) suggest that children may habituate
to painful medical procedures because behavioral
distress was negatively correlated with number of
previous medical procedures. With regard to dental
visits, no relationship was found between behavior
during the visit and history of unpleasant medical
experiences (Johnson & Baldwin, 1968). Thus, these
mixed results indicate that the relationship between
prior medical experience and emotional adjustment is
not well understood.
Onset of illness has also not been extensively
investigated. Time since diagnosis was significantly
negatively correlated with behavioral distress during a
medical procedure (Jay et al., 1983), further
supporting the idea that children habituate to aversive
medical procedures. Nevertheless, chronicity of
illness as a variable in hospital adjustment has not
been studied. A related concept, length of hospital
stay, has also not received much research attention.
In an adult medical population, length of
hospitalization was not related to psychological
adjustment (Levenson, Hamer, Silverman, Rossiter,

1986-1987). Both variables, onset of illness and
length of hospitalization, warrant further
investigation.
20
Pilot Study
Introduction and Rationale
Given the many unanswered issues regarding
adjustment, a pilot study was conducted to examine the
applicability of a new measure in studying variables
influencing distress in a hospitalized pediatric
population aged 4-12. Few studies have investigated
depressive symptomatology in pediatric populations of
preschool and pre-literate children. Moreover,
measures specifically designed to assess distress in
medical settings are limited.
The pilot study focused on the application of a
new parent rating scale of behaviors associated with
depression and anxiety (Saylor et al., 1987). This
rating scale involves the evaluation of the frequency
of specific behaviors, not a parental interpretation of
emotional distress. Moreover, this brief questionnaire
does not require direct observation by a trained
clinician. Thus, preliminary findings regarding the
influence of background variables on adjustment to
hospitalization were gathered.

21
Methods
Subjects. The sample consisted of 81 mothers of
hospitalized children (49 boys, 32 girls) from
consecutive pediatric admissions at Shands Hospital in
Gainesville, Florida. Ages of the children ranged from
4 years, 1 month to 11 years, 8 months (M = 6 years, 8
months; SD = 1 year, 11 months). The sample of
children was 73% White and 27% African-American; 64% of
the children lived in two-parent homes and 32% lived in
single-parent homes. The sample was predominantly
lower to middle SES (95%), based on the five levels of
social class as assessed by Hollingshead's (Myers &
Bean, 1968) Two Factor Index of Social Position (Class
1 = 3%; Class 2 = 2%; Class 3 = 17%; Class 4 = 40%;
Class 5 = 38%).
With respect to illness-related variables, 27% of
the children had never been previously hospitalized,
24% had one prior hospitalization, and the remaining
49% had multiple hospitalizations (ranging from 2 to
28), with the number of hospitalizations M = 6.1, SD =
15.8. Duration of illness ranged from diagnosis at
birth to diagnosis upon admission to the hospital, with
47% of the patients diagnosed within the past six
months.
Medical diagnoses obtained from mothers were
subsequently categorized into illness groups derived

22
from Nelson's Textbook of Pediatrics (Behrman, Vaughan,
& Nelson, 1987) with 79% of the sample falling into
one of six groups. Twenty percent of those children
classified were diagnosed with cardiovascular or
respiratory illnesses (e.g., pneumonia, tetralogy of
fallot, cystic fibrosis). Fourteen percent were
hospitalized for immunity, allergy, or related diseases
(e.g., asthma, reflex sympathetic dystrophy). Another
17% were experiencing illnesses interfering with the
digestive system (e.g., Hirshsprung's disease,
inflammatory bowel disease, appendicitis). Nineteen
percent of the patients were diagnosed with urinary
problems (e.g., reflux, nephrotic syndrome, kidney
infection). An additional 17% were classified with
diseases of the nervous system (e.g., spina bifida,
Hallervorden spatz, hydrocephalus). Diagnoses had not
been identified for another 13% of the children. The
seventeen patients not categorized into the above
groups had been diagnosed with several different
illnesses (e.g., burns, autism, leg fracture) and did
not conform with the six illness groups.
Furthermore, descriptions of the clinical course
of illnesses were obtained form Nelson's Textbook of
Pediatrics (Behrman et al., 1987) to identify diagnoses
based on chronicity. Sixty-eight children in the
sample were thus classified, with 28% of these

23
receiving diagnoses for acute illnesses and 72%
receiving diagnoses for chronic illnesses. The
remaining thirteen diagnoses not categorized on
chronicity represent children with ambiguous or unknown
illnesses.
In addition to data collected on mothers, eleven
fathers also participated in the pilot study. Three of
these fathers completed the measure along with their
spouses, whereas the remaining eight fathers were the
only parent available for participation in the study.
Children who exhibited mental or physical
limitations which would interfere with parental report
of distress were not included in the sample. For
instance, children with a developmental delay or
hearing impairment would complicate parental response
to items regarding conversational ability (e.g.,
"Refuses to speak") or on items requiring comprehension
of requests (e.g., "Accepts advice or instructions
easily"). Determination of whether the child met these
exclusionary criteria was made through interaction with
the child and soliciting the judgment of the parent
and/or nurse.
Measure. The Behavioral Upset in Medical
PatientsRevised (BUMP-R; Saylor et al., 1987) is a
56-item parent rating of the child's behavior
corresponding to emotional distress at the hospital and

24
at home (See Appendix A). This scale is a revision of
the adult version (Zeldow & Braun, 1985) consisting of
a 32-item checklist of behaviors that nonpsychiatric
patients may exhibit in hospital settings. Patient
behaviors were rated by nurses on a Likert scale
ranging from 0 to 4. This scale range indicates the
frequency of behavior, with 0 representing "never and
4 representing "always."
Saylor et al. (1987) revised the scale for use
with children by having five judges independently
evaluate which items were inappropriate for children.
Those items deemed inappropriate by a majority of
judges were deleted, yielding a 28-item scale. A
sample item is:
Looks depressed or sad Never
Sometimes
Often
Usually
Always
Scoring of the BUMP-R parallels the original adult
version. However, instead of ratings by nurses,
parents initially rate the child's behavior in the
hospital followed by ratings of the same behaviors at
home. Therefore, the BUMP-R provides a parent rating
of the child's behavioral upset in the hospital and at
home, prior to hospitalization.

25
Psychometric data are limited due to the recency
of the measure's development. For the adult version,
internal consistency was reported as .93 and test-
retest reliability over variable intervals was reported
as .66. Four factors were identified from 213
inpatient adults, including behavioral regression, poor
patient-staff relationship, depression and anxiety, and
passivity and withdrawal (Zeldow & Braun, 1985).
Procedure. Participation was solicited the day
following the child's admission to the hospital.
Mothers were instructed to complete the BUMP-R in the
hospital room, rating their child's behavior since the
current hospitalization as well as rating their child's
behavior at home.
Results and Discussion
Analyses indicated that age, SES, and the BUMP-R
scores were normally distributed. However, the number
of previous hospitalizations and the duration of
illness were not normally distributed. Consequently,
analyses utilizing these two illness variables are
based on Spearman rho correlation coefficients.
Whereas no sex differences were found in the
BUMPR-Home rating of distress, a trend for girls (M =
29.8, SD = 11.4) to exhibit more distress in the
hospital than boys (M = 24.9, SD = 13.7) emerged (t =
1.67, p = .09). As expected, no race or family

26
composition differences were found. There were also no
significant differences in the BUMPR-Hospital ratings
due to diagnosis based on the six categories of illness
(F(63) = 1.12, p > .05). Only marginal differences for
chronicity of illness (t(66) = 1.73, p = .09) were
found on the BUMPR-Hospital, with children diagnosed
with acute illnesses (M = 32.1, SD = 14.4) scoring
higher than those diagnosed with chronic illnesses (M =
26.3, SD = 11.6).
Age was significantly correlated with hospital
distress (r = -.24, p < .05), with younger children
obtaining higher distress ratings than older children.
Behavioral distress in the hospital or at home was not
significantly related to SES, duration of illness, or
number of previous hospitalizations.
Ratings of hospital distress were significantly
associated with home distress (r = .48, p < .0001). In
addition, maternal ratings of their child's behavioral
upset in the hospital (M = 26.8, SD = 13.0) were
significantly higher (t(79) = 2.67, p < .01) than
ratings of distress behaviors at home (M = 23.1, SD =
11.3), suggesting that mothers observed increased
behavioral distress following hospitalization.

27
Purpose of Study and Hypotheses
The purpose of the current investigation was two
fold. First, variables affecting behavioral upset in a
hospitalized population aged 4-12 were further
explored. Second, issues regarding assessment
modalities were studied. Specifically, differences
among raters were investigated as well as a comparison
of measures designed to assess distress in medical
populations with measures more commonly used in the
assessment of depression. As part of the assessment
battery, a newly designed self-report scale appropriate
for pre-literate children in hospital settings was
included. The usefulness of this new measure for
hospitalized children was of particular interest.
Assessment of depression was obtained following
hospital admission via self-reported behavioral
distress and depression, nurse report of the child's
behavioral distress, and parental report of their
child's depression and behavioral distress as well as
the parent's own anxiety in the hospital setting. No
significant differences in the outcome measures due to
sex, race, socioeconomic status, or family composition
(i.e., single v. two-parent homes) were anticipated.
The following hypotheses were tested.
(1) Age and grade differences were expected, with
younger children exhibiting greater emotional distress.

28
(2) Differences in depression and behavioral upset
were hypothesized for different diagnostic groups.
(3) Onset of illness, number of prior
hospitalizations, and length of current hospital stay
were expected to be positively related to behavioral
upset and depression.
(4) Number of hours spent with parent was expected
to be negatively correlated with adjustment to
hospitalization.
(5) All outcome measures, including parental
anxiety, were hypothesized to be intercorrelated.
(6) In addition, behavioral distress was expected
to be rated higher in the hospital setting as compared
to home, indicating distress at hospitalization.

CHAPTER 2
METHODS
Subjects
Seventy mothers and their children were recruited
from consecutive pediatric admissions at Shands
Teaching Hospital. The sample of children consisted of
29 males and 41 females between the ages of 4 years, 2
months and 12 years, 11 months (M = 8 years, 6 months,
SD = 2 years, 8 months). Racial composition was 66%
White, 27% African-American, and 7% Other. With regard
to school grade level, there were 10% of the children
not attending school, 4% in preschool programs, 12% in
kindergarten, 16% in the first grade, 11% in the second
grade, 10% in the third grade, 10% in the fourth grade,
13% in the fifth grade, 10% in the sixth grade, and 4%
in the seventh grade.
The family composition of the sample consisted of
66% of the children living in two-parent homes and 34%
of the children living in single-parent homes.
Although primarily lower to middle SES (90%), the
sample included all five levels of social class as
measured by Hollingshead's (Myers & Bean, 1968)
Two-Factor Index of Social Position (Class 1 = 3%;
Class 2 = 7%; Class 3 = 16%; Class 4 = 37%; Class 5 = 37%).
29

30
Based on diagnoses reported by mothers, 77% of the
children were divided into six diagnostic groups based
on Nelson's Textbook of Pediatrics (Behrman et al.,
1987). Of those diagnoses categorized, 19% of the
children were hospitalized for cardiovascular or
respiratory difficulties (e.g., cystic fibrosis,
cardiac myopathy, coronary artery disease). Twenty-two
percent of the children were diagnosed with immunity,
allergy, or related diseases (e.g., HIV, asthma,
juvenile rheumatoid arthritis) and 20% percent of the
children were experiencing difficulties involving the
digestive system (e.g., cleft lip and palate,
alpha-antitrypsin deficiency, gastroenteritis). An
additional 13% of the children were diagnosed with
urinary difficulties (e.g., nephrotic syndrome, urinary
tract infection, bladder infection). Thirteen percent
of the patients had been hospitalized for diseases
affecting the nervous system (e.g., spina bifida,
seizure disorders). For another 13% of the children,
diagnosis had not been determined (e.g., fever of
unknown origin). The sixteen children not classified
into diagnostic groups had received a variety of
diagnoses, preventing assignment to a group of adeguate
number for analyses. These unclassified diagnoses
included viral meningitis, diabetes, snakebite, and
sickle cell anemia.

31
In addition, 62 of the diagnoses were categorized
on chronicity based on Nelson's Textbook of Pediatrics
(Behrman et al., 1987) description of pathogenesis.
Thirty-five percent of diagnoses were described as
acute illnesses and 65% of diagnoses were identified as
chronic illnesses. The eight children not classified
on chronicity had undiagnosed illnesses.
Forty percent of the sample of children had never
been previously hospitalized, 12% of the children had
been hospitalized once before, whereas 43% had been
hospitalized on multiple occasions (ranging from 2 to
20 prior hospitalizations), with prior hospitalizations
M = 3.2, SD = 5.1. The time spent hospitalized during
the study ranged from 1 day to 29 days (M = 5.0, SD =
4.8) with 54% hospitalized for three days or less. The
duration of illness associated with their diagnosis
ranged from newly diagnosed to diagnosis at birth, with
41% diagnosed within one month, and an additional 20%
within six months. The average number of hours the
parent spent with the child in the hospital in a 24
hour period ranged from 3 hours to 24 hours a day, with
67% of parents reporting they spent 24 hours a day with
their child and 17% spending between 18 and 23 hours
with their child.
Thirty-two of the children involved in the study
also received behavior ratings from their nurses. In

32
addition to the participation of mothers, thirteen
fathers were also involved in the study. Five of the
fathers were spouses of mothers who also participated
in the study whereas eight fathers were the only
parental respondents involved in the study.
Children who manifested mental or physical
deficits which would impede responses on the child
self-report and parent measures were excluded (e.g.,
profound mental retardation, lack of language
development, significant vision or hearing impairment).
Comprehension problems or the inability to see the
pictures, for example, would interfere with the child's
responses on the Child-BUMP. The evaluation of whether
a child met these exclusionary criteria was made
through interaction with the child and from parent
and/or nurse judgments. Informed parental consent and
child verbal assent were independently obtained. Basic
demographic information and information regarding
illness was collected on a summary sheet (See Appendix
B) .
Measures
The Behavioral Upset in Medical PatientsRevised
(BUMP-R; Savior et al., 1987). As detailed in the
description of the pilot study (p.23-25), the BUMP-R is
a 56-item parent rating of the child's behavior

33
corresponding to emotional distress at the hospital and
at home (See Appendix A). Patient behaviors are rated
by parents on a Likert scale ranging from 0 to 4. In
addition, the child's nurse completed the BUMP-R
questions for behavior observed in the hospital.
The Personality Inventory for Children. Depression
Scale (PIC-D; Lachar & Gdowski. 1979; Wirt, Lachar,
Klinedinst, & Seat. 1977). The PIC-D is one of 16
profile scales of the PIC, a comprehensive 600-item,
true-false parent rating measure of psychological
adjustment for children ages 3-16.
The Depression scale contains 46 rationally-
derived items. A sample item is:
My child is usually in good spirits True
False
These items yield ten interpretable factors, with
Brooding/Moodiness and Social Isolation accounting for
56% of the variance. The eight other factors are
Crying Spells, Lack of Energy, Pessimism/Anhedonia,
Concern with Death and Separation, Serious Attitude,
Sensitivity to Criticism, Indecisiveness/Poor
Self-Concept, and Uncommunicativeness. A Total
Depression scale score is obtained.
With respect to validity, the PIC-D correlated
significantly with all scales on the Conners Parent
Questionnaire except the Antisocial Scale and

34
correlated with the Children's Depression Inventory
(Leon et al., 1980). Moreover, the Depression scale
correlated with social withdrawal, depression,
uncommunicativeness, and social subscales of the Child
Behavior Checklist (Kelly, 1982). Test-retest
reliability was reported as ranging from .80 to .94 and
internal consistency (coefficient alpha) of .86 (Wirt
et al., 1984).
The State Trait Anxiety Inventory (STAI, Form Y)
(Spielberqer, 1983). The STAI consists of two
self-report scales measuring state anxiety and trait
anxiety. The form used in this study is a revision of
an earlier version (Spielberger, Gorsuch, & Lushene,
1970). In contrast to the earlier version which had
items related to depression, Form Y measures feelings
of anxiety discriminated from symptoms of depression.
The State Anxiety scale consists of 20 statements about
how the subject feels "right now, at this moment." The
Trait Anxiety scale consists of 20 statements about how
the subject generally feels. Subjects are asked to
rate the intensity of feelings on a Likert scale
ranging from (1) not at all, (2) somewhat, (3)
moderately so, (4) almost always. The following sample
item is from the Trait Anxiety scale:

35
I feel nervous and restless Not at all
Somewhat
Moderately so
Almost always
The following sample item is from the State Anxiety
Scale:
I feel strained Not at all
Somewhat
Moderately so
Almost always
Test-retest correlations for the Trait Anxiety
scale range from .73 to .86 for college students.
Stability on the State Anxiety scale, ranging from .16
to .62, reflects that state anxiety is influenced by
situational factors (Spielberger, 1983). With regard
to internal consistency, the median alpha coefficient
was reported as .93 for State Anxiety and .90 for Trait
Anxiety (Spielberger, 1983). The Trait Anxiety scale
has demonstrated concurrent validity with such measures
as the Taylor Manifest Anxiety Scale (Taylor, 1953) and
the scores on the state anxiety scale have been
correlated with situational stress (see Spielberger,
1983 for review).
Parents were instructed to complete the State
Anxiety items for how they felt at that moment, during
the evaluation. Then they were asked to complete the

36
Trait Anxiety items for their general feeling since
their child was hospitalized.
The Children's Depression Inventory (CPI) (Kovacs.
1983, 1985; Finch et al., 1985). The CDI is a 27-item
self-report depression scale appropriate for children
ages 8 and older. Each item consists of three
statements representing graded levels of severity of a
depressive symptom and the child selects one of the
three statements. A sample item is:
I feel like crying everyday.
I feel crying many days.
I feel like crying once in a while.
The choices are valued from 0 to 2, with high scores
indicating depression. The internal consistency of the
CDI (coefficient alphas) was reported as ranging from
.70 to .94 (Kovacs, 1985; Saylor et al., 1984).
Test-retest reliability ranged from correlations of .38
to .87 (Saylor et al., 1984).
The Behavioral Upset in Medical Patients-Child
Self-Report Version (Child-BUMP). The Child-BUMP is a
27-item pictorial scale designed for use in this study
(See Appendix C). All questions from the BUMP-R for
the hospital setting (with one exception) were
rephrased in language understandable for children ages
4-12. The questionnaire was rephrased by three
independent sources and the simplest items selected.

37
The scale was modeled according to the Harter &
Pike (1983) Pictorial Scale of Perceived Competence and
Social Acceptance for Young Children. In this scale,
children are asked to select from two items with
corresponding pictures which is most like them. Then,
for each item, two possible questions are asked
regarding the frequency with which they exhibit the
behavior. This two-step process facilitates attention
and simplifies the items for young children. A sample
item from the Child-BUMP is as follows:
This boy cries a lot. This boy cries a little.
(If child selects first picture,
ask next two questions)
DO YOU:
always cry OR sometimes cry
(If child selects second picture,
ask next two questions)
DO YOU:
hardly ever cry OR never cry
Each final response is scored on a scale of one to
four, with high scores suggestive of greater behavioral
distress. A total score was obtained by summing the
item scores. Fourteen of the items begin with the most
negative response (4) and thirteen of the items depict
African-American children. Male and female versions of

38
the scale are identical except for the sex of the child
in the picture and as phrased in the question.
Procedure
Parents of eligible patients were informed of the
study and participation was solicited by trained
undergraduate research assistants supervised by the
experimenter. Research assistants underwent a training
program involving several steps. Assistants were first
given a written description of general data collection
procedures (See Appendix D) that they rehearsed prior
to attending a 2-hour training session. During the
training session, the assistants watched a videotape
simulation of the experimenter administering the
measures to two children, a 5-year-old male and an
11-year-old female. Following the videotape, the
assistants role-played and discussed solving potential
difficulties that could be encountered during data
collection. After the training session, each assistant
observed the experimenter administering the measures to
at least one subject. Finally, the experimenter
observed each research assistant giving the measures
until the assistant demonstrated proficiency in the
data collection procedures.
Families were administered the questionnaires in
the hospital rooms the day following the child's

39
hospital admission. This delay allowed approximately
24 hours of behavior for both the parents and children
to assess reaction to hospitalization (time since
hospitalization ranged from 17 to 35 hours). Following
the collection of demographic information, parents were
instructed to complete the BUMP-R and PIC-D based on
their child's behavior and the STAI on their own
feelings since their child's hospitalization. As the
parents completed their forms, the research assistant
read aloud the Child-BUMP to all children and read the
CDI to children ages 8-12. To reward cooperation, the
child was allowed to select a small prize (e.g.,
sticker, toy car, puzzle book). After the data was
collected from the parent and child, the nurse
completed the BUMPR-Hospital form provided that the
nurse had interacted with the child for a minimum of
eight hours.

CHAPTER 3
RESULTS
Initially, this section focuses on the results
obtained from the examination and analysis of the data
collected from the 70 mothers, their children, and the
nurses involved in the current study. Differences in
demographic and illness-related variables were examined
as well as the relationships among the test scores.
Following this presentation, the information gathered
from mothers in this study were compared to the data
collected from the 81 mothers who participated in the
pilot study to highlight significant differences
between studies. Data from the mothers in the current
study and the pilot study were then combined and
analyzed to explore differences in demographic and
illness-related variables on the BUMP-R for this
larger, combined sample. In addition, a factor
analysis of the BUMPR-Hospital ratings was performed
for this combined sample.
Finally, analyses on a sample of fathers was also
conducted. This sample consisted of 13 fathers from
the current study and 11 fathers from the pilot study.
A matched sample of mothers was obtained for
comparison, in which eight of the fathers were compared
40

41
to their spouses who had also participated. For the
remaining 16 fathers, mothers were selected for
comparison based on the child's age and sex. Although
this matched sample cannot address correspondence of
ratings between mothers and fathers, the comparison
addresses whether fathers rate distress in their
children similar to mothers' ratings.
Analyses for Current Study
Descriptive Results & Statistics
Scores used for all analyses were the raw scores
of the BUMP-R, Child-BUMP, and CDI and the standard T-
Scores of the PIC-D and STAI. Inspection of skewness,
kurtosis, and histograms indicated that all of the
outcome measures were normally distributed. However,
the length of current hospitalization, the number of
prior hospitalizations, the duration of illness, and
the number of hours the parent spent with the child
were not normally distributed. Therefore, analyses
utilizing these variables were based on Spearman rho
coefficients. Moreover, given the number of
correlations computed, the significance level was
reduced to alpha=.01 for correlational analyses in
order to control for Type I errors.
Descriptive statistics on each test variable are
displayed in Table 1. No normative data is available

42
Table 1: Means and Standard Deviations of the Outcome
Measures
N
Mean
(SD)
BUMPR-Hospital
70
27.2
(13.3)
BUMPR-Home
70
24.7
(11.2)
PIC-D
70
60.9
(15.0)
STAI-State
70
55.3
(12.2)
STAI-Trait
65
52.2
(10.4)
Child-BUMP
70
51.5
(8.4)
CDI
33
6.9
(5.1)
Nurse-BUMP
32
28.6
(18.8)
for the BUMP-R scores or for the new Child-BUMP. The
sample mean on the PIC-D (M = 60.9, SD = 15.0) was one
standard deviation above the normative sample mean,
suggesting that mothers reported more depressive
symptomatology in this population than in the general
population. However, sample means on parent report of
anxiety on both the STAI State scale (M = 55.3, SD =
12.2) and the STAI Trait scale (M = 52.2, SD = 10.4)
were within normal limits. In addition, the current
sample mean on the CDI (M = 6.9, SD =5.1) is
comparable to the mean reported for newly diagnosed
diabetics (Kovacs, 1983) and is below the recommended
cut-off for diagnosis of depression.

43
The new Child-BUMP measure appeared to function as
designed, with the children readily grasping the two-
step response process. Four-year-old children
occasionally had difficulty with the measure if the
parent had suggested that the child had comprehension
problems. Repeating items was often helpful for
children, particularly for those with more limited
attention spans. Cronbach's coefficient alpha for the
Child-BUMP was .76, suggesting fairly strong internal
consistency. For the Nurse-BUMP, internal consistency
was .93, indicating that the individual items are very
strongly intercorrelated.
Analyses of Background Data
An examination of the demographic variables sex,
age group, race, and family composition indicated no
significant differences for the outcome measures (See
Tables 2-6). No differences had been anticipated based
on sex, race, or family composition. However, a
marginal relationship was found between the CDI scores
and family composition (See Table 5), with children of
single parent homes reporting more depressive symptoms
than children in two-parent homes (t = 1.99, p = .055).
Contrary to the expectation that younger children (ages
4-8) would exhibit more distress than older children
(ages 8-12), no differences between age groups were
found on any of the outcome measures. Analyses also

44
Table 2: Demographic Differences for the BUMPR-Hospital
N
Mean
l (SD)
t
Sex
Males
29
28.3
(14.2)
.60
Females
41
26.4
(12.7)
Acre Group
4-8
32
29.0
(15.5)
1.01
8-12
38
25.7
(11.0)
Race
White
46
28.1
(13.0)
.36
African-American
19
26.7
(14.9)
Family Composition
2-Parent
44
25.4
(12.8)
1.55
1-Parent
23
30.6
(13.5)
Note; All t values are nonsignificant.

45
Table 3: Demographic Differences for the PIC-Depression
N Mean (SD) t
Males
29
62.1
(13.9)
.59
Females
41
60.0
(15.9)
Acre Group
4-8
32
63.2
(18.2)
1.20
8-12
38
58.9
(11.6)
Race
White
46
60.8
(16.2)
. 04
African-American
19
60.6
(13.2)
Family Composition
2-Parent
44
58.3
(12.8)
1.70
1-Parent
23
64.9
(18.7)
Note: All t values are nonsignificant.

46
Table 4: Demographic Differences for the Child-BUMP
N
Mean (SD)
t
Sex
Males
29
52.0
(8.8)
.48
Females
41
51.0
(8.7)
Acre GrouD
4-8
32
50.7
(10.2)
. 65
8-12
38
52.0
(7.3)
Race
White
46
52.6
(8.6)
1.75
African-American
19
48.5
(8.5)
Family Composition
2-Parent
44
51.5
(8.3)
. 05
1-Parent
23
51.4
(9.1)
Note: All t values are nonsignificant.

47
Table 5: Demographic Differences for the CPI
N
Mean
: (SD)
t
Sex
Males
12
8.3
(5.0)
1.12
Females
21
6.2
(5.2)
Race
White
21
7.0
(5.2)
.28
African-American
10
6.4
(5.2)
Family Composition
2-Parent
21
5.4
(4.3)
1.99*
1-Parent
11
8.8
(5.1)
Note: All t values are nonsignificant.
E = .055

48
Table 6: Demographic Differences for the Nurse-BUMP
N
Mean fSD)
t
Sex
Males
17
31.3
(19.6)
.86
Females
15
25.5
(18.1)
Acre Group
4-8
16
32.2
(17.8)
1.08
8-12
16
25.0
(19.6)
Race
White
26
26.8
(17.8)
1.23
African-American
5
38.2
(24.7)
Family Composition
2-Parent
19
25.5
(17.6)
.97
1-Parent
11
32.4
(20.5)
Note: All t values are nonsignificant.

49
revealed no significant interaction between sex and age
group or between sex and race.
An examination of the Spearman coefficients (See
Table 7) indicated that the length of the current
hospital visit and the number of prior hospitalizations
were not associated with any of the measures of
behavioral distress. Duration of illness was
significantly negatively correlated with only parental
state anxiety. Moreover, the relationship between the
Child-BUMP scores and hours spent with the parent was
positively correlated, a direction opposite of that
predicted. Thus, the more time a parent spent with the
child, the more distress the child reported on the
pictorial measure. Although no other significant
relationships with time spent with the parent emerged,
a modest negative correlation with parental state
anxiety was noted.
With regard to the six diagnostic groups, there
were no significant differences on the BUMPR-Hospital
scores (F(53) = .87, p > .05), on the CDI (F(26) =
1.92, p >.05), or on the Nurse-BUMP (F(24) = 1.68, p >
.05). Only a marginal difference in diagnostic groups
was found with the PIC-D scores (F(53) = 2.11, p = .08)
and with the Child-BUMP scores (F(53) = 2.13, p = .08).
Based on the classification of chronicity of illness,
no significant differences appeared on any of the

Table 7: Spearman Correlations Between Illness-Related Variables and Outcome Measures
Length of
hospitalization
r (D)
Number of Prior
hospitalizations
r (n)
Duration of
illness
r (n)
Number of Hours
spent w/ parent
r (n)
BUMPR-Hospital
.14 (70)
.04 (70)
.02 (64)
. 16
(70)
BUMPR-Home
.13 (70)
.21 (70)
.14 (64)
. 05
(70)
PIC-Depression
.04 (70)
.11 (70)
.00 (64)
. 05
(70)
Child-BUMP
-.14 (70)
-.04 (70)
-.17 (64)
.29
(70)*
CDI
.00 (33)
.09 (33)
-.08 (31)
. 01
(33)
Nurse-BUMP
-.15 (32)
-.15 (32)
-.29 (30)
. 04
(32)
STAI-State
.20 (70)
-.15 (70)
-.30 (64)*
-.26
(70)a
STAI-Trait
.03 (65)
.09 (65)
-.17 (59)
-.13
(65)
* E < -01
a Because the significance level was reduced to .01, the marginal relationship was found at
only p < .05.

51
outcome measures. In contrast to the pilot results and
the hypothesis, there was no significant difference
between the BUMPR-Hospital and BUMPR-Home score (t(70)
= 1.73, p > .05).
Correlational Analyses
The correlation matrix in Table 8 indicates that,
as expected, many of the outcome variables were
intercorrelated. Consistent with the pilot study, the
BUMPR-Hospital ratings were significantly positively
correlated with the BUMPR-Home ratings. In addition,
the BUMPR-Hospital scores were also significantly
positively associated with the PIC-D, Child-BUMP, and
CDI. However, the parent ratings of hospital distress
on the BUMP-R was only marginally positively correlated
with the Nurse-BUMP. The parent report of distress on
the BUMPR-Home form was significantly positively
correlated with the PIC-D and CDI. Moreover, the
parent ratings on the PIC-D were significantly
positively correlated with the CDI and parental trait
anxiety. Parental trait and state anxiety were also
positively correlated. With regard to demographic
variables, SES was significantly positively correlated
with the BUMPR-Home rating and the PIC-D scores.
The BUMPR-Hospital ratings and the Nurse-BUMP
ratings were not significantly correlated with parent
report of state or trait anxiety. In addition, the

Table 8: Correlations Among Outcome Measures and Demographic Variables
BUMPR-Hosp
r (n)
BUMPR-Home
E (n)
PIC-D
r (n)
Child-BUMP
E (n)
CDI Nurse-BUMP
r (n) r (n)
STAI-State
r (n)
STAI-Trait
E (n)
BUMPR-Home
.53
(70)
PIC-D
.45
(70)
.49 (70)
Child-BUMP
.36
(70)*
.10 (70)
.14 (70)
CDI
.43
(33)
.52 (33)
.54 (33)
.30 (33)
Nurse-BUMP
.39
(32)*
.02 (32)
.04 (32)
.32 (32)
-.17 (14)
STAI-State
.15
(70)
.07 (70)
.22 (70)
-.05 (70)
-.03 (33)
-.16 (32)
STAI-Trait
.14
(65)
.21 (65)
.40 (65)
-.02 (65)
.26 (31)
-.19 (29)
.64 (65)
AGE
-.15
(70)
.04 (70)
-.17 (70)
.04 (70)
.36 (33)* -
-.23 (32)
-
.03 (70)
-.08 (70)
GRADE
-.20
(70)
-.04 (70)
-.22 (70)
-.01 (70)
.30 (33)
-.25 (32)
-
.02 (70)
-.07 (65)
SES
-.02
(70)
.31 (70)*
.36 (70)
.16 (70)
.27 (33)
.29 (32)
.00 (70)
.23 (65)
* n < .01
" a < .001
' Because the significance level was reduced to .01, this marginal relationship was found at only p < .05.

53
parent report on the BUMPR-Home form was not
significantly correlated with the Child-BUMP, Nurse-
BUMP, or parental state or trait anxiety. Similarly,
the PIC-D was not significantly related with the Child-
BUMP, Nurse-BUMP, or parental state anxiety. The
Child-BUMP was also not significantly associated with
the CDI, Nurse-BUMP, or parental state or trait
anxiety. The CDI was not significantly correlated with
the Nurse-BUMP or parental state or trait anxiety.
Moreover, contrary to the hypothesis that younger
children would demonstrate more distress, age and grade
were not significantly correlated with the outcome
variables. This finding does not replicate the modest
correlation of age and BUMPR-Hospital scores revealed
in the pilot study. Lastly, SES was not significantly
related to scores on the BUMPR-Hospital, Child-BUMP,
CDI, Nurse-BUMP, or parental state or trait anxiety.
Analyses of Current Study and
Pilot Study Combined
Comparison of Samples
Based on the results obtained from the pilot
study, a difference had been anticipated in the current
study between maternal ratings of distress at
hospitalization and maternal ratings of distress at
home. Moreover, the effects of age and gender were not
observed in the current study compared to the pilot.

54
Given that these findings were not replicated in the
present study, a comparison of the two samples of
subjects was performed.
Distribution of race, family composition, SES,
duration of illness, number of prior hospitalizations,
and diagnostic grouping were comparable between
samples. However, the pilot study was composed of
predominantly males whereas the current study consisted
of primarily females, a significant gender difference
(X2 = 5.46, p = .01). Furthermore, the children in the
pilot study were significantly younger than the
children in the current study (F(150) = 23.86, p <
.0001).
Analyses of Combined Sample
The data obtained from maternal ratings on the
BUMP-R from both samples were combined and analyzed,
resulting in a more evenly distributed sample with
respect to sex and age. This combined sample consisted
of 151 mothers of 78 boys and 73 girls. Ages of the
children ranged from 4 years, 1 month to 12 years, 11
months (M = 7 years, 6 months, SD = 2 years, 5 months).
With this combined sample, no significant sex
differences were found in the mothers' ratings of
hospital distress (t(149) = .80, p > .05) or in the
ratings of home distress (t(149) = .70, p > .05).
Moreover, children aged 4-8 years obtained ratings

55
comparable to the children aged 8-12 years on the
BUMPR-Hospital (t(149) = .91, p > .05) and on the
BUMPR-Home (t(149) = 1.19, p > .05). Age evidenced
only a small relationship (r = -.17, p < .05) with
ratings of behavioral distress in the hospital.
There were no differences in the BUMPR-Hospital
scores (F(117) = 1.55, p > .05) or in the BUMPR-Home
scores (F(117) = 1.45, p > .05) for the six diagnostic
groups. The combined sample demonstrated no
differences for chronicity of illness on the ratings of
hospital distress (t(128) = .78, p > .05) or the
ratings of home distress (t(128) = .29, p > .05). No
significant relationship was found between the number
of prior hospitalizations and the BUMPR-Hospital scores
(r = .06, p > .05) or the BUMPR-Home scores (r = .06,
p > .05). Similarly, duration of illness was unrelated
to the BUMPR-Hospital ratings (r = .06, p > .05) or the
BUMPR-Home ratings (r = .05, p > .05).
With this combined sample, maternal ratings of
distress following hospitalization (M = 27.0, SD =
13.1) were significantly higher (t(151) = 3.15, p <
.01) than ratings of distress at home (M = 23.8, SD =
11.2). Consistent with the results of both the pilot
study and the current study, ratings of distress in the
hospital were significantly correlated with ratings of
distress in the home (r = .50, p < .0001).

56
Factor Analysis of BUMPR-Hospital
The BUMPR-Hospital scores obtained from the 151
mothers participating in the pilot study and the
current study were then further analyzed in order to
explore the psychometric characteristics of this
measure. Because the BUMPR-Home items were designed to
serve only as comparisons to the hospital items, the
BUMPR-Home scores alone would not address
hospitalization adjustment and thus were not further
analyzed. Correlations between individual items on the
BUMPR-Hospital rating and the total score are reported
in Table 9. Cronbach's coefficient alpha assessing
internal consistency of the total BUMPR-Hospital
measure was .87, suggesting that individual items are
strongly intercorrelated.
Factor analytic procedures were then performed on
the BUMPR-Hospital scores. Principal factors were
derived from the correlation matrix and the diagonal
was replaced by communality estimates. For these
communality estimates, Gorsuch (1983) maintains that,
instead of unities for the diagnonals, a more accurate
estimation of the communality for a given variable
utilizes the squared multiple correlation with all
other variables. To obtain the optimal factor
solution, factor rotation was performed using the
varimax orthogonal transformation, the promax oblique

57
Table 9: Item-Total Correlations for the BUMPR-Hospital
Item
1
.416
Item
15
. 332
Item
2
.484
Item
16
.563
Item
3
.466
Item
17
.328
Item
4
.568
Item
18
.491
Item
5
.562
Item
19
.333
Item
6
.453
Item
20
.343
Item
7
.290
Item
21
.109
Item
8
.398
Item
22
.518
Item
9
.113
Item
23
.486
Item
10
.242
Item
24
.269
Item
11
.385
Item
25
.376
Item
12
.435
Item
26
.516
Item
13
.511
Item
27
.301
Item
14
. 638
Item
28
. 346

58
transformation, as well as the quartimax orthogonal
transformation. Comparison of the rotation methods
revealed that the promax oblique rotation yielded
factor structures of substantial factor complexity and
thus complicated interpretability. Thus, orthogonal
solutions were preferred over the promax oblique
solution which assumes intercorrelated factors. The
quartimax orthogonal rotation was selected in favor of
the varimax orthogonal rotation because Gorsuch (1983)
advises that the varimax method is inappropriate when
the measure has high internal consistency. The
quartimax rotation assumes that a single general factor
accounts for a substantial amount of the variance,
which is implied with high internal consistency,
whereas the varimax rotation does not assume one large
factor.
The number of factors extracted was based on three
criteria. As Gorsuch (1983) described, Guttman
recommended examination of the characteristic roots
(eigenvalues) for factor extraction. The number of
eigenvalues greater than one estimates the number of
factors. Based on this criterion, four factors were
extracted (See Table 10), which accounted for 85% of
the variance.

59
Table 10:
Eigenvalues for the BUMPR-Hospital Factor
Analysis
Eigenvalues
Factor 1
6.07
Factor 2
2.50
Factor 3
1.78
Factor 4
1.09
The next criteria involved a plot of the
eigenvalues, known as the scree test. Gorsuch (1983)
describes that the predominant factors will account for
a significant portion of the variance. Thus, the
number of factors can be extracted where the plot of
the eigenvalues levels off. The scree test shown in
Figure 1 supports extraction of four factors.
Lastly, interpretability was used as the final
criterion. Comparisons to three factor and five factor
solutions confirmed that the quartimax four-factor
solution retained the greatest number of items and
resulted in the most parsimonious factor structure.
Moreover, the quartimax four factor rotation yielded
the most readily interpretable solution.
Item loadings of .40 were considered significant,
which is more conservative than the traditional .30
cutoff and would result in stronger factors. Only Item

Figure 1: Scree Test of Eigenvalues for Factor Analysis of
BUMP-R Hospital Ratings
60
/ +
6 +
5 +
4 +
3 +
2 +
6
7
8 9
0 1
2 3 4
5 67 89 01 23 45
+ + + + + + + + + +--
6 8 10 12 14 16 18 20 22 24 26
28
Number
VO +

61
14 (Is uncooperative) loaded on more than one factor,
and two items did not have significant loadings on any
of the four factors (Item 6: Refuses to speak; Item 7:
Says he or she feels blue or depressed; See Table 11
for factor structure). Using the .40 item loading
criterion, the first factor was extracted with 11
items, identified as Negativity/Agitation (See Table
12), with an internal consistency coefficient of .86.
This first factor accounted for 41.2% of the variance.
The second factor of nine items, labelled Amiability
(See Table 13), encompassed sociability and
agreeability. This factor attained an internal
consistency coefficient of .79 and accounted for 26.3%
of the variance. The third factor with four items was
named Dysphoria (See Table 14), including vegetative
symptoms of depression and emotional discomfort, with
internal consistency at .68. The third factor
accounted for 19.9% of the variance. The final factor
was identified as Noncompliance (See Table 15), with
four items and an internal consistency of .68. The
last factor accounted for 12.5% of the variance.
Analysis of Fathers for Combined Sample
To examine similarities between parental raters,
fathers were compared to a matched sample of mothers.
Results indicated that fathers' ratings (M = 30.7, SD =
15.6) were similar to mothers' ratings (M = 28.3, SD =

Table 11; Factor Structure of the BUMPR-Hosnital
FACTOR 1
FACTOR 2
FACTOR 3
FACTOR 4
Item
1
.63589
-.04183
-.00338
. 02754
Item
2
.45435
.12336
.36104
. 00857
Item
3
.59320
.02689
.18392
-.04462
Item
4
.65127
.13778
.22307
-.05709
Item
5
.65284
.12495
.17103
-.02268
Item
6
.30260
.37821
.25504
-.06537
Item
7
.18094
.18013
.28743
.01944
Item
8
.22155
.17271
.10444
.66649
Item
9
-.06193
-.00589
.43279
.14313
Item
10
.32859
-.03857
-.10527
.45162
Item
11
.19292
.08428
.67505
-.02277
Item
12
.16359
.22701
.70590
-.03168
Item
13
.53006
.07606
.30855
.02363
Item
14
.57713
.26186
.08055
.39853
Item
15
.51382
.06229
-.02597
-.14439
Item
16
.70806
.08321
-.01942
.19862
Item
17
.12274
.07588
.32630
.48533
Item
18
.72155
.03593
-.16548
.20505
Item
19
.55057
-.10472
-.12761
.29440
Item
20
.21628
.10644
.45006
.00852
Item
21
-.14432
.39522
.02132
.13843
Item
22
.25396
.67377
.14469
-.12223
Item
23
.28992
.64513
-.13552
.15905
Item
24
-.03104
.53951
.20288
-.09263
Item
25
.04713
.56132
.30167
-.11305
Item
26
.24431
.63897
.12487
.01719
Item
27
.03600
.54897
.01670
.08005
Item
28
.23332
.48658
-.27860
.27999

63
Table 12: Item Loadings for Factor OneNegativity/
Agitation
Item Loading
Item 18
Demanding
(.72)
Item 16
Stubborn, negativistic
(.71)
Item 4
Becomes upset easily
(.65)
Item 5
Is irritable or grouchy
(.65)
Item 1
Is impatient
(.64)
Item 3
Gets angry
(.59)
Item 14
Is uncooperative
(.58)
Item 19
Manipulative
(.55)
Item 13
Clinging, needs lots of reassurance
(.53)
Item 15
Complains
(.51)
Item 2
Cries
(.45)
Internal Consistency Factor 1: Coefficient alpha = .86

Table 13: Item Loadings for Factor TwoAmiability
Item Loading
Item 22
Tries to be friendly
(.67)
Item 23
Accepts advice or instructions easily
(.64)
Item 26
Pleasant to be with
(.64)
Item 25
Laughs or smiles at funny comments
or events
(.56)
Item 27
Shows interest in recovery (takes
initiative)
(.55)
Item 24
Starts conversation
(.54)
Item 28
Does what he or she is told
(.49)
Item 21
Able to ask for help
(.40)
Internal Consistency Factor 2: Coefficient alpha = .79

65
Item
Loading
Item
12
Looks depressed or sad
(.71)
Item
11
Looks worried, tense
(.68)
Item
20
Has sleep problems
(.45)
Item
9
Sleeps unless directed into
activity (.43)
Internal Consistency Factor 3: Coefficient alpha = .68

66
Table 15: Item Loadings for Factor FourNoncompliance
Item
Loading
Item 8
Has to be reminded what to do
(.67)
Item 17
Is incredibly passive
(.49)
Item 10
Has to be told to follow hospital
routine
(.45)
Item 14
Is uncooperative
(.40)
Internal Consistency Factor 4: Coefficient alpha = .68

67
15.6) for the BUMPR-Hospital scores (t(46) = .54, g >
.05). Ratings on the BUMPR-Home showed a similar
pattern (t(46) = .57, p > .05), with fathers (M = 25.0,
SD = 12.0) obtaining scores comparable to mothers (M =
23.0, SD = 11.7).
For the 24 fathers involved in the pilot study and
the current study, no significant sex differences for
the children emerged on the BUMPR-Hospital rating
(t(22) = .80, g > .05) or for the BUMPR-Home rating
(t(22) = .34, g > .05). In addition, no age group
differences were found between 4-8 year old children
and 8-12 year old children for the BUMPR-Hospital
scores (t(22) = 1.03, g > .05) or for the BUMPR-Home
scores (t(22) = .43, g > .05). Age was also not
significantly correlated with the BUMPR-Hospital
ratings (r = -.25, g > .05) or with the BUMPR-Home
ratings (r = -.11, g > .05). Previous number of
hospitalizations was not associated with the
BUMPR-Hospital scores (r = .10, g > .05) or with the
BUMPR-Home scores (r = .36, g > .05). However,
duration of illness was correlated with the
BUMPR-Hospital ratings (r = -.49, g < .05) but not with
the BUMPR-Home ratings (r = .03, g > .05).
Similar to the earlier analyses of mothers,
fathers also rated hospital distress (M = 30.7, SD =
15.6) as significantly higher (t(22) = 2.38, g < .05)

68
than home distress (M = 25.0, SD = 12.0). Moreover,
the hospital distress score was also signicantly
correlated with the home distress score (r = .66, p <
.001).

CHAPTER 4
DISCUSSION
Background Variables Affecting Distress
The first purpose of the present study was to
examine variables influencing depression and adjustment
in hospitalized children aged 4-12. The findings
indicated no significant gender differences on any of
the distress or depression measures. This result is
consistent with the bulk of the literature (e.g., Jay
et al., 1983; Saylor et al., 1987). However, it did
not support the trend observed in the pilot study for
girls to obtain higher BUMPR-Hospital scores than boys.
A comparison of the pilot sample with the current
sample revealed significantly more females in the
latter sample. When the two samples were combined, no
significant sex differences were found on the BUMPR-
Hospital form, for mothers or fathers. In fact, the
two studies cited earlier that suggested that girls
exhibit more distress (Katz et al., 1980; Melamed &
Siegel, 1975) consisted of samples with proportionately
more boys. Consequently, when the sample is more
evenly represented by males and females, any sex
differences in depressive symptoms for pediatric
populations may disappear.
69

70
Contrary to the hypothesis, younger children did
not obtain higher distress scores in the current study.
Many previous studies have found an inverse
relationship between age and depressive symptoms in
pediatric populations (e.g., Jay et al., 1983; Katz et
al., 1980; Saylor et al., 1987). Moreover, age was
found to be negatively correlated with BUMPR-Hospital
scores in the pilot study. Thus, the current study did
not confirm the negative correlation between age and
distress found in the pilot study. Again, a comparison
revealed that children in the pilot study were
significantly younger than those in the current sample.
With both samples combined, a low but significant
negative correlation was found between age and mothers'
BUMPR-Hospital ratings. The magnitude of this age
correlation is comparable to that found in the study
that introduced the BUMP-R (Saylor et al., 1987).
Thus, younger children may show more distress upon
hospitalization, although the relationship does not
appear strong.
As anticipated, no significant differences were
found for race on any of the outcome measures. The
influence of ethnicity on distress in pediatric
populations has not been studied. However, the absence
of racial differences in distress found in the current
study is consistent with research in depression for the

71
general child population (e.g., Angold, 1988; Kaslow &
Racusin, 1990) as well as with the pilot study results.
With regard to family composition, only the CDI
showed a small, nonsignificant difference, with
children of single-parent households obtaining somewhat
higher scores than children of two-parent households.
This suggests that there may be a tendency for children
to display more depressive symptoms because of the
stressors associated with single-parent homes.
However, given that this group difference was not
significant and that family composition differences
have not been reported in the general or pediatric
population with regard to depressive symptoms, this
finding may be spurious. Overall, the results
confirmed the pilot study findings and supported the
hypothesis that family composition does not
significantly affect behavioral distress in pediatric
populations.
No relationship between socioeconomic status and
distress was expected. However, the PIC-D and the
BUMPR-Home ratings were significantly positively
correlated with SES index. Thus, mothers in lower
socioeconomic levels reported more symptoms of
depression and distress in their children, possibly
reflecting the degree of stress associated with greater
financial hardship. This finding is consistent with

72
the relationship found in a dental setting (Wright &
Alpern, 1971). Given that the remaining measures used
in this study that target distress in pediatric
populations did not manifest this relationship, perhaps
SES is not as influential in this group of children.
Maternal anxiety was predicted to be positively
correlated with child's distress. The findings
revealed that maternal reports of state and trait
anxiety were generally not related to the child's
behavioral upset. The only relationship that emerged
involved the PIC-D and maternal trait anxiety in that
mothers' reports of greater anxiety during the child's
hospital stay was associated with their reports of more
depressive symptoms in their children. This result may
be attributed to source bias because anxious mothers
may expect their children to display adjustment
difficulties upon hospitalization. The absence of
strong support for the relationship between children's
distress and maternal anxiety contrasts with several
previous studies (e.g., Blotcky et al., 1985; Jay et
al., 1983; Walker & Greene, 1989). Perhaps maternal
anxiety was not strongly related to child distress in
the present study because those previous studies
involved specific diagnostic groups (e.g., cancer,
recurrent abdominal pain). When pediatric patients
with diverse diagnoses are involved, parental anxiety

73
may not be significantly associated with the child's
behavioral upset or depression.
With regard to the amount of parent contact during
hospitalization, a negative correlation with child
distress had been anticipated. However, only the
Child-BUMP self-report scale was positively correlated
with the number of hours the parent spent with the
child. In other words, the more contact the child had
with the parent upon hospitalization, the more
behavioral upset the child reported. This finding
supports studies relating maternal presence to
increased negative behavior (e.g., Shaw & Routh, 1982).
Children may have interpreted high contact with a
parent during hospitalization as indicative of cause
for concern, or the children may have felt more
comfortable expressing their distress in their parents'
presence. Alternatively, those parents with more
distressed children may be more likely to spend more
time at their child's bedside in an attempt to
alleviate their distress.
In contrast, the remaining measures of child
distress were not associated with the amount of parent
contact, which supports earlier findings on
hospitalized children (Saylor et al., 1987). An
interesting and unexpected low positive correlation
between amount of parent contact and maternal state

74
anxiety emerged. Therefore, mothers reported greater
anxiety the more time they spent with their child.
This may reflect parent fatigue or it may suggest that
concern for their child's illness may influence both
amount of contact and anxiety ratings.
Overall, this pattern of results regarding parent
contact does not support the Peterson et al. (1985)
position promoting parental contact in order to avoid
difficulties following separation upon hospitalization.
One salient limitation in drawing conclusions about
parent contact from the current study stems from
subject participation being contingent upon parental
presence. Obviously, those patients who spent little
time with their hospitalized child were particularly
likely to be unavailable for study participation.
Thus, distress in children who could not participate in
this study may indeed be greater with limited parent
contact. In addition, parents themselves reported the
number of hours spent with their child and they may
have exaggerated the amount of time. Indeed, there was
minimal statistical variability in reported parent
contact, and, given the skewed distribution and the
possible influence of outliers, results regarding
amount of time the parent spent with the child should
be interpreted cautiously.

75
A number of illness-related variables were also
examined. Diagnostic group differences in depression
and behavioral distress were not found in the current
study. For this relatively understudied aspect of
adjustment, categorization by type of illness and by
chronicity of illness failed to support significant
group differences, contrary to expectation. These
findings are comparable to those in the pilot study and
in a study of cancer patients (Jay et al., 1983).
However, given the wide variety of diagnoses obtained
in the current sample, categorization of illness was
difficult and group sizes were possibly too small for
adequate comparison. An ideal study would compare
several clearly definable illness groups which might
then yield differences in behavioral upset and
depression.
The number of prior hospitalizations was expected
to be positively associated with distress, although the
results did not confirm this hypothesis. Previous
medical experience was not correlated with behavioral
distress in the current study or in the pilot study.
Results from previous research have been mixed, with
some investigations suggesting greater distress for
frequently hospitalized patients (e.g., Saylor et al.,
1987) and other investigations suggesting habituation
and decreased upset with more hospital visits (e.g.,

76
Jay et al., 1983). The current findings suggest that
previous medical history may be unrelated to the
child's reaction to hospitalization. The current study
included children with a wide variety of diagnoses and
medical histories whereas the Jay et al. (1983) study
involved a sample of children with cancer. Perhaps the
quality of previous medical experience (i.e., negative
or positive perceptions) is more influential in a
child's adjustment to hospitalization than frequency of
prior hospitalizations.
Contrary to the hypothesis, duration of illness
and length of hospital stay were not positively
correlated with the child's emotional distress.
Although previous research has not studied these
variables closely, children with chronic illness and
longer hospitalizations were expected to obtain higher
scores on measures of depression and behavioral upset.
However, both the current study and pilot study did not
find these relationships. One interesting negative
association between maternal state anxiety and duration
of illness emerged. The concept of habituation may
account for this relationship, with mothers reporting
less anxiety as they become accustomed to the chronic
nature of their child's illness. With regard to length
of hospitalization, most children were hospitalized for
short periods and greater variability in length of

hospital stay may have revealed differences in
adjustment.
77
Evaluation of the BUMP-R
The second purpose of this study was the
investigation of measures of distress particularly
suitable for pediatric populations. The Behavioral
Upset in Medical PatientsRevised designed by Saylor
and her colleagues (1987) appears to be a promising
parent report measure. The current study did not
reveal significant differences between ratings of
hospital and home distress, which did not support the
results obtained from the pilot study. With the data
from both the pilot and current study combined, a more
evenly distributed sample emerged. Maternal hospital
ratings in this combined sample were significantly
higher than home ratings. Moreover, fathers also
reported more behavioral upset in the hospital as
compared to the home. Therefore, the overall findings
suggest that hospitalization may precipitate emotional
distress.
The current study confirmed the pilot study's
strong positive correlation between the BUMPR-Hospital
scores and the BUMPR-Home scores. When the two samples
were combined, this relationship remained strong for
both mothers' and fathers' ratings. These results

78
corroborate the correlation found by Saylor et al.
(1987). These authors suggested that premorbid
psychological functioning, as measured by the BUMPR-
Home, predicts a child's adjustment to the stress of
hospitalization. Consequently, children reportedly
experiencing greater emotional distress at home may be
more susceptible to adjustment difficulties upon
hospitalization.
The BUMPR-Hospital and BUMPR-Home scores were also
positively correlated with the other parent report
measure, the PIC-D. This indicated that the BUMP-R
taps into the construct of depression similarly to an
instrument traditionally used for healthy children.
Likewise, the BUMPR-Hospital and BUMPR-Home ratings
were positively correlated with the children's self-
reported depression on the CDI. Given that these
correlations with measures commonly used in the general
child population are only moderate, the BUMP-R appears
to be assessing some unique aspect of child behavior.
Indeed, that the BUMPR-Home ratings correlated more
strongly with the PIC-D and CDI than does the BUMPR-
Hospital would suggest that the latter is more
sensitive to hospitalization behavior.
The BUMPR-Hospital rating was significantly
positively correlated with the Child-BUMP but only
marginally positively correlated with the Nurse-BUMP.

79
On the other hand, the BUMPR-Home was not correlated
with either the Child-BUMP or Nurse-BUMP. This further
supports that the BUMPR-Hospital specifically measures
adjustment in a hospital setting, not depressive
symptoms in general.
Conclusions regarding correspondence between
mothers and nurses are limited given the small sample
of 32 nurses available to complete the BUMPR-Hospital.
However, from the marginal relationship found between
mothers' and nurses' ratings and the strong internal
consistency of the Nurse-BUMP, nurses may be able to
reliably assess behavioral distress in a hospitalized
child after as little as eight hours of contact.
Fathers were seldom available to participate in
the study or they indicated that they preferred the
mother complete the measures. Thus, this study did not
address interrater agreement between fathers and
mothers for the BUMPR-Hospital form. However, the
sample of fathers (comprised of fathers participating
conjointly with mothers and fathers participating
alone) was compared to a sample of mothers matched for
the child's age and sex. This comparison indicated
that fathers provided BUMP-R ratings comparable to
mothers. Therefore, fathers appear to be as reliable
raters on the BUMP-R as mothers.

80
For the combined samples of the pilot and current
studies, Cronbach's coefficient alpha for the BUMPR-
Hospital rating was .87. Given this strong internal
consistency, the most suitable and parsimonious
solution in the factor analysis of the BUMPR-Hospital
scores was based on the quartimax rotation. Four
factors emerged, with the largest number of items
representing Negativity/Agitation. The second factor
was labelled as Amiability, the third factor as
Dysphoria, and the fourth factor as Noncompliance.
This factor structure appears to adequately capture
those dimensions involved in adjustment to
hospitalization.
Evaluation of Child-BUMP
A self-report pictorial measure, the Child-BUMP,
was designed for the current study to tap emotional
distress in hospitalized children as young as four
years old. Based on a two-step selection process,
children identify pictures reflecting their level of
distress since hospitalization. Most children appeared
to understand the instrument, particularly with
repetition of items. Cronbach's coefficient alpha for
the Child-BUMP was computed as .76, which indicates
moderately strong internal consistency.

81
The Child-BUMP was moderately significantly
correlated with the parent report on the BUMPR-Hospital
form. Thus, children and mothers demonstrated some
agreement regarding the child's emotional distress,
although they appear to hold different perspectives
given that the correlation was moderate. As would be
expected, the Child-BUMP, which assesses behavioral
upset in the hospital, did not correlate significantly
with the mothers' BUMPR-Home ratings, giving some
initial indication of discriminative validity of the
measure. In addition, the Child-BUMP was not
associated with the parent's report on the PIC-D,
possibly because the PIC-D was not intended for use in
medical populations. Similarly, the CDI was not
significantly correlated with the Child-BUMP. This may
be attributed to the physical symptoms and school-
related items appearing on the CDI which would not be
appropriate for hospitalized, physically-ill children.
Finally, the Child-BUMP was not significantly
correlated with the Nurse-BUMP, which suggests low
agreement between these two sources.
Implications
The present study ventured into several relatively
unexplored areas in pediatric psychology, including
perspectives from various sources, incorporating self-

82
reports from preschool children, utilizing a
heterogeneous sample of children, and examining a
comprehensive array of potential risk factors. The
findings from the current study highlight the need for
continued research to determine which children are at
risk for developing adjustment difficulties upon
hospitalization. Most of the background variables
examined did not clarify which children are most likely
to exhibit distress in the hospital.
Of the variables investigated, maternal anxiety,
parent contact, diagnosis, and previous medical
experience specifically require further examination.
Alternative means of assessing maternal anxiety (e.g.,
clinician-administered interview) may help resolve
confusion regarding the influence of mothers' emotional
reactions on child adjustment. With regard to parent
contact, health personnel could log the amount of time
the parent spends with the child to ensure accuracy of
reporting. As suggested earlier, future studies could
determine specific diagnostic groups to examine a
priori and evaluate distress in these groups in order
to obtain sufficient group sizes. In addition,
investigation of previous medical experience should
include the child's subjective evaluation of prior
encounters with medical personnel.

83
Premorbid psychological functioning, as measured
by the BUMPR-Home rating, was strongly related to
distress upon hospitalization. Therefore, a child's
emotional and behavioral difficulties outside the
hospital is predictive of emotional upset upon
hospitalization. Future research should assess
children prior to hospitalization (e.g., for scheduled
procedures) to determine which children are likely to
experience adjustment difficulties. For children
entering the hospital for unscheduled procedures, a
brief assessment battery shortly after hospital
admission could be investigated to facilitate
identification of children who will likely experience
difficulties in the hospital.
One intriguing direction for pediatric psychology
research moves toward evaluating the influence of the
context of health care delivery on adjustment. For
instance, studies should more closely examine the
impact of a family's access to health care. Parental
experience of the availability of health care for their
children may affect their attitudes towards medical
intervention. Many families may also encounter poor
continuity and follow-up by health care providers.
Availability may also affect a family's health care
orientation, i.e., their attitude towards health care
and medical professionals. Belief systems also likely

84
influence the perceived acceptability of medical
intervention. In addition, social support systems may
contribute to health care orientation as well as
provide assistance to parent and child during health
care crises. These variables may affect the parent and
child's perspective and adjustment when encountering a
given hospitalization.
With regard to assessment of pediatric
populations, overall, the BUMP-R appears to be a useful
instrument in the assessment of emotional distress in
hospitalized children, with concurrent validity, high
internal consistency, and appropriateness for various
raters. Further investigation of the BUMP-R should
evaluate the consistency of distress ratings over time
as well as comparing scores of hospitalized children
with outpatient pediatric samples in order to ascertain
the specific effect of hospitalization. The current
multimethod investigation utilized self-report, parent-
report, and nurse-report. Correspondence between
raters should be further explored, focusing on
interrater reliabilities between mothers, fathers,
children, and health personnel (e.g., nurses, doctors).
In addition, the Child-BUMP warrants further
investigation regarding its utility for pre-literate
hospitalized children. Items should be further
examined to ensure the suitability for these young

85
children. Both child and parent report of behavioral
distress could also be compared to clinician ratings of
depression and adjustment difficulties to further
assess the validity of the BUMP-R and Child-BUMP. The
BUMP-R and Child-BUMP appear to be promising
instruments for assessing preschool children, an
understudied age group because of the unique challenges
of research with very young children.

REFERENCES
Achenbach, T. M., McConaughy, S. H., & Howell, C. T.
(1987). Child/adolescent behavioral and emotional
problems: Implications of cross-informant
correlations for situational specificity. Psychology
Bulletin. 101. 213-232.
Ammon Cavanaugh, S. von. (1986). Depression in the
hospitalized inpatient with various medical
illnesses. Psychotherapy and Psychosomatics. 45.
97-104.
Angold, A. (1988). Childhood and adolescent depression:
I. Epidemiological and aetiological aspects. British
Journal of Psychiatry. 152. 601-617.
Behrman, R. P., Vaughan, V. C. Ill, & Nelson, W. E.
(1987). Nelson's textbook of pediatrics (13th Ed.).
Philadelphia: W. C. Saunders Co.
Blotcky, A. D., Raczynski, J. M., Gurwitch, R., &
Smith, K. (1985). Family influences on hopelessness
among children early in the cancer experience.
Journal of Pediatric Psychology. 10. 479-493.
Boyd, J. H., & Weissman, M. M. (1981). Epidemiology of
affective disorders: A re-examination and future
directions. Archives of General Psychiatry. 38,
1039-1046.
Burke, P., Meyer, V., Kocoshis, S., Orenstein, D. M.,
Chandra, R., Nord, D. J., Sauer, J., & Cohen, E.
(1989). Depression and anxiety in pediatric
inflammatory bowel disease and cystic fibrosis.
Journal of the American Academy of Child &
Adolescent Psychiatry. 28, 948-951.
Cunningham, S. J., McGrath, P. J., Ferguson, H. B.,
Humphreys, P., D'Astous, J., Latter, J., Goodman, J.
T., & Firestone, P. (1987). Personality and
behavioural characteristics in pediatric migraine.
Headache. 27. 16-20.
86

87
Douglas, J. W. B. (1975). Early hospital admission and
later disturbances of behaviour and learning.
Developmental Medicine and Child Neurology, 17,
456-480.
Eason, L. J. Finch, A. J., Jr., Brasted, W., & Saylor,
C. F. (1985). The assessment of depression and
anxiety in hospitalized pediatric patients. Child
Psychiatry and Human Development. 16, 57-64.
Finch, A. J., Jr., & Saylor, C. F. (1984). An overview
of child depression. In W. J. Burns & J. V. Lavigne
(Eds.), Progress in pediatric psychology (pp.
201-239). Orlando, FL: Grue & Stratton.
Finch, A. J., Jr., Saylor, C. F., Edwards, G. L.
(1985). Children's Depression Inventory: Sex and
grade norms for normal children. Journal of
Consulting and Clinical Psychology. 53., 424-425.
Gauvain-Piquard, A., Rodary, C., Rezvani, A., &
Lemerle, J. (1987). Pain in children aged 2-6 years:
A new observational rating scale elaborated in a
pediatric oncology unitpreliminary report. Pain.
31, 177-188.
Gorsuch, R. L. (1983). Factor analysis (2nd Ed.).
Hillsdale, NJ: Lawrence Erlbaum Assoc.
Harter S., & Pike, R. (1983). Procedural manual to
accompany the Pictorial Scale of Perceived
Competence and Social Acceptance for Young Children.
University of Denver.
Heilgenstein, E., & Jacobsen, P. B. (1988).
Differentiating depression in medically ill children
and adolescents. Journal of the American Academy of
Child & Adolescent Psychiatry. 27, 716-719.
Jay, S. M., Ozolins, M., Elliott, C. H., & Caldwell, S.
(1983). Assessment of children's distress during
painful medical procedures. Health Psychology. 2,
133-147.
Jensen, R. A. (1955). The hospitalized child: Round
Table, 1954. American Journal of Orthopsychiatry.
25, 293-318.
Jessner, L., Blom, G. E., & Waldfogel, S. (1952).
Emotional implications of tonsillectomy and
adenoidectomy of children. Psychoanalytic Study of
the Child. 7, 126-169.

88
Johnson, R., & Baldwin, D. C. (1968). Relationship of
maternal anxiety to the behavior of young children
undergoing dental extraction. Journal of Dental
Research. 47, 801-805.
Kandel, D. B., & Davies, M. (1982). Epidemiology of
depressive mood in adolescents: An empirical study.
Archives of General Psychiatry. 39., 1205-1212.
Kashani, J. H., Barbero, G. J., & Bolander, F. D.
(1981). Depression in hospitalized pediatric
patients. Journal of the American Academy of Child
Psychiatry, 20. 123-134.
Kashani, J. H., Holcomb, W. R., & Orvaschel, H. (1986).
Depression and depressive symptoms in preschool
children from the general population. American
Journal of Psychiatry. 143. 1138-1143.
Kashani, J. H., Husain, A., Shekim, W. O., Hodges, K.
K., Cytryn, L., & McKnew, D. H. (1981). Current
perspectives on childhood depression: An overview.
American Journal of Psychiatry, 138. 143-153.
Kaslow, N. J., & Racusin, G. R. (1990). Childhood
depression: Current status and future directions. In
A. S. Bellack, M. Hersen, & A. E. Kazdin (Eds.),
International handbook of behavior modification and
therapy (2nd ed., pp. 649-667). New York: Plenum
Press.
Kaslow, N. J., Rehm, L. P., Pollack, S. L., & Siegel,
A. W. (1988). Attributional style and self-control
behavior in depressed and nondepressed children and
their parents. Journal of Abnormal Child Psychology,
16, 163-175.
Katz, E. R., Kellerman, J., & Siegel, S. (1980).
Behavioral distress in children with cancer
undergoing medical procedures: Developmental
considerations. Journal of Consulting and Clinical
Psychology, 48. 356-365.
Kazdin, A. E. (1987). Assessment of childhood
depression: Current issues and strategies.
Behavioral Assessment. 9, 291-319.
Kazdin, A. E. (1988). Childhood depression. In E. J.
Mash & L. G. Terdal (Eds.), Behavioral assessment of
childhood disorders (2nd ed., pp. 157-195). New
York: Guilford Press.

89
Kazdin, A. E. (1989). Identifying depression in
children: A comparison of alternative selection
criteria. Journal of Abnormal Child Psychology, 17,
437-454.
Kazdin, A. E., Colbus, D., Rodgers, A. (1986).
Assessment of depression and diagnosis of depressive
disorder among psychiatrically disturbed children.
Journal of Abnormal Child Psychology. 14, 499-515.
Kazdin, A. E., French, N. H., Unis, A. S., &
Esveldt-Dawson, K. (1983). Assessment of childhood
depression: Correspondence of child and parent
ratings. Journal of the American Academy of Child
Psychiatry. 22, 157-164.
Kerr, M. M., Holer, T. S., Versi, M. (1987).
Methodological issues in childhood depression: A
review of the literature. American Journal of
Orthopsychiatry. 57, 193-198.
Kovacs, M. (1983). The Children's Depression Inventory:
A self-rated depression scale for school-aged
youngsters. Unpublished manuscript, University of
Pittsburg School of Medicine.
Kovacs, M. (1985). The Children's Depression Inventory
(CDI). Psvchopharmacoloav Bulletin. 21. 995-998.
Lachar, D., & Gdowski, C. L. (1979). Actuarial
assessment of child and adolescent personality: An
interpretive guide for the Personality Inventory for
Children profile. Los Angeles, CA: Western
Psychological Services.
Lefkowitz, M. M., & Tesiny, E. P. (1980). Assessment of
childhood depression. Journal of Consulting and
Clinical Psychology, 48, 43-50.
Leon, G. R., Kendall, P. C., Garber, J. (1980).
Depression in children: Parent, teacher, and child
perspectives. Abnormal Child Psychology. 8, 221-235.
Levenson, J. L., Hamer, R., Silverman, J. J., Rossiter,
L. F. (1986-87). Psychopathology in medical
inpatients and its relationship to length of
hospital stay: A pilot study. International Journal
of Psychiatry in Medicine, 16, 231-236.

90
Lichtenstein, D., Dreger, R. M., & Cattell, R. B.
(1986). Factor structure and standardization of the
Preschool Personality Questionnaire. Journal of
Social Behavior and Personality. 1, 165-181.
Melamed, B. G., & Siegel, L. J. (1975). Reduction of
anxiety in children facing hospitalization and
surgery by use of filmed modeling. Journal of
Consulting and Clinical Psychology. 43. 511-521.
Mullins, L. L., Siegel, L. J., & Hodges, K. (1985).
Cognitive Problem-Solving and life-event correlates
of depressive symptoms in children. Journal of
Abnormal Child Psychology. 13, 305-314.
Myers, J., & Bean, L. (1968). A decade later: A follow
up of social class and mental illness. New York:
Wiley & Sons.
Olson, R. A., Holden, E. W., Friedman, A., Faust, J.,
Kenning, M., & Mason, P. J. (1988). Psychological
consultation in a children's hospital: An evaluation
of services. Journal of Pediatric Psychology. 13,
479-492.
Peterson, L., Mori, L., & Carter, P. (1985). The role
of the family in children's responses to stressful
medical procedures. Journal of Clinical Child
Psychology. 14, 98-104.
Pfeffer, C. R., & Trad, P. V. (1988). Sadness and
suicidal tendencies in preschool children.
Developmental and behavioral pediatrics. 9, 86-88.
Poznanski, E. D., Cook, S. C., & Carroll, B. J. (1979).
A depression rating scale for children. Pediatrics.
M, 442-450.
Prugh, D. G., Staub, E. M., Sands, H. H., Kirschbaum,
R. M., & Lenihan, E. A. (1953). A study of the
emotional reactions of children and families to
hospitalization and illness. American Journal of
Orthopsychiatry. 23. 70-106.
Reynolds, W. M., Anderson, G., & Bartell, N. (1985).
Measuring depression in children: A multimethod
assessment investigation. Journal of Abnormal Child
Psychology. 13, 513-526.

91
Ryan, N. D., Puig-Antich, J., Ambrosini, P.,
Rabinovich, H., Robinson, D., Nelson, B., Iyengar,
S., & Twomey, J. (1987). The clinical picture of
major depression in children and adolescents.
Archives of General Psychiatry. 44, 854-861.
Saylor, C. F., Finch, A. J., Jr., & McIntosh, J. A.
(1988). Self-reported depression in psychiatric,
pediatric, and normal populations. Child Psychiatry
& Human Development. 18., 250-254.
Saylor, C. F., Finch, A. J., Jr., Spirito, A., &
Bennett, B. (1984). The Children's Depression
Inventory: A systematic evaluation of psychometric
properties. Journal of Consulting and Clinical
Psychology. 52, 955-967.
Saylor, C. F., Pallmeyer, T. P., Finch, A. J., Jr.,
Eason, L., Trieber, F., & Folger, C. (1987).
Predictors of psychological distress in hospitalized
pediatric patients. Journal of the American Academy
of Child and Adolescent Psychiatry. 26, 232-236.
Seligman, M. E. P., Peterson, C., Kaslow, N. J.,
Tanenbaum, R. L., Alloy, L. B., & Abramson, L. Y.
(1984). Attributional style and depressive symptoms
among children. Journal of Abnormal Psychology. 93.,
235-238.
Shanahan, K. M., Zolkowski-Wynne, J., Coury D. L.,
Collins, E. W., O'Shea, J. S. (1987). The Children's
Depression Rating Scale for normal and depressed
outpatients. Clinical Pediatrics. 26. 245-247.
Shaw, E. G., & Routh, D. K. (1982). Effect of mother
presence on children's reaction to aversive
procedures. Journal of Pediatric Psychology. 7,
33-42.
Spielberger, C. D. (1983) Manual for the State-Trait
Anxiety Inventory (Form Y, Self-Evaluation
Questionnaire). Palo Alto, CA: Consulting
Psychologists Press.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E.
(1970). Manual for the State-Trait Anxiety Inventory
(Self-Evaluation Questionnaire). Palo Alto, CA:
Consulting Psychologists Press.

92
Stone, W. L., & Lamenek, K. L. (1990). Developmental
issues in children's self-reports. In A. M. Lagreca
(Ed.), Through the eves of a child (pp. 18-56).
Boston: Allyn & Bacon.
Taylor, J. A. (1953). A personality scale of manifest
anxiety. Journal of Abnormal and Social Psychology.
48, 285-290.
Vardaro, J. A. (1978). Preadmission anxiety and
mother-child relationships. Journal of the
Association for the Care of Children in Hospitals.
7, 8-15.
Walker, L. S., & Greene, J. W. (1989). Children with
recurrent abdominal pain and their parents: More
somatic complaints, anxiety, and depression than
other patient families? Journal of Pediatric
Psychology. 14. 231-243.
Wallander, J. L., Varni, J. W., Babani, L., Banis, H.
T., & Wilcox, K. T. (1988). Children with chronic
physical disorders: Maternal reports of their
psychological adjustment. Journal of Pediatric
Psychology. 13, 197-212.
Webb, T. E., & VanDevere, C. A. (1985). Sex differences
in the expression of depression: A developmental
interaction effect. Sex Roles. 12, 91-95.
Wirt, R. D., Lachar, D., Klinedinst, J. K., & Seat, P.
D. (1984) Multidimensional description of child
personality: A manual for the Personality Inventory
for Children. Los Angeles, CA: Western Psychological
Services.
Wisniewski, J. J., Naglieri, J. A., & Mulick, J. A.
(1988). Psychometric properties of a Children's
Psychosomatic Symptom Checklist. Journal of
Behavioral Medicine. 11, 497-507.
Worchel, F. F., Nolan, B. F., Willson, V. L., Purser,
J. S., Copeland, D. R., & Pfefferbaum, B. (1988).
Assessment of depression in children with cancer.
Journal of Pediatric Psychology. 13, 101-112.
Wright, G. Z., & Alpern, G. D. (1971). Variables
influencing children's cooperative behavior at the
first dental visit. Journal of Dentistry for
Children. 38, 124-128.

93
Zeldow, P. B., & Braun, L. (1985). Measuring regression
in hospitalized medical patients: The BUMP scale.
General Hospital Psychiatry, 7, 49-53.

APPENDIX A
BEHAVIORAL UPSET IN MEDICAL PATIENTS REVISED (BUMP-R)
PARENT RATING FORM
Subject #: DATE:
Your relationship to child:
Side ope
PLEASE RATE YOOR CHILD'S BEHAVIOR AS YOD HAVE SEEN IT SINCE YODR ARRIVAL AT THE
HOSPITAL. MARK ONE BOX PER QUESTION TO INDICATE YODR CHOICE.
1.
2 .
Is impatient
Cries
u
s
>
e
a
n
r i
0)

ft
H
u

fl
o
m
1
[]
[]
a

%4
o
n
r i
*
M
-t
s
3
e
D
[]
r n
s
>
s
*
iH
A
C 3
r i
3.
Gets angry
[]
l J
[]
l J
[3
l J
[]
4.
Becomes upset easily
[]
[]
n
[3
[]
5.
Is irritable or grouchy
[]
[]
n
[]
[3
6.
Refuses to speak
[]
u
n
[3
C 3
7.
says he or she feels blue or depressed
[]
[]
n
C 3
[3
8.
Has to be reminded what to do
[]
[]
[]
[]
[]
9.
Sleeps unless directed into activity..
M
n
[]
[3
[]
10.
Has to be told to follow hospital routine....
[]
[]
[3
[]
[3
11.
Looks worried, tense
[]
[]
C 3
[3
C3
12.
Looks depressed or sad
[]
c]
C3
(3
C3
13.
Clinging, needs lots of reassurance
[]
[]

[]
C 3
14.
Is uncooperative
M
[]
C3
[ 3
[3
15.
Complains
[]
[]
C3
C3
[]
16.
Stubborn, negativistio.
[]
u
tl

[ 3
17.
Is incredibly passive
[]
[3
[3
C 3
[]
18.
Demanding
[]
[]
[]
C 3
C 3
19.
Manipulative
[]
U
[3
[ 3
C 3
20.
Has sleep problems
[]
[]
[3
C 3
[3
21.
Able to ask for help
u
[]
[3


22.
Tries to be friendly
[]
[]
[]
[ 3
C 3
23.
Accepts advice or instructions easily
[]
[]
[3
Cl
n
24.
Starts conversations
[]
[]
C3
[3
3
25.
Laughs or smiles at funny comments or events.
[]
[]
[ 3
[3
[3
26.
Pleasant to be with
[]
C3
[3
[]
[3
27 .
Shows interest in recovery (takes initiative)
[]
[]
C3
[]
[]
28.
Does what he or she is told
[]
[]
[]
[]
n
94

95
Side Two
PLEASE RATE YOUR CHILD'S BEHAVIOR AS YOU HAVE SEEN IT AT.
WAS HOSPITALIZED. MARK ONE BOX PER QUESTION TO INDICATE
HOME. BEFORE
YOUR CHOICE.
HE OR SHE
1. is impatient
2. Cries
3. Gets angry
4. Becomes upset easily
5. Is irritable or grouchy
6. Refuses to speak
7. Says he or she feels blue or depressed
8. Has to be reminded what to do
9. sleeps unless directed into activity
10. Has to be told to follow hospital routine....
11. Looks worried, tense
12. Looks depressed or sad
13. Clinging, needs lots of reassurance
14. Is uncooperative
15. Complains
16. Stubborn, negativistic
17. Is incredibly passive
18. Demanding
19. Manipulative
20. Has sleep problems
21. Able to ask for help...
22. Tries to be friendly
23. Accepts advice or instructions easily
24. Starts conversations
25. Laughs or smiles at funny comments or events.
26. Pleasant to be with
27. Shows interest in recovery (takes initiative)
28. Does what he or she is told
u
05

a
H
4J
t
H
H
a
>i
9
<8
9
9
9
>
a
4J
P

9
0
%4
9
H
z
00
o
D
<
. []
[]
[]
Cl
Cl
. []
[]
Cl
u
Cl
. []
[J
[]
[]
Cl
. []
[]
Cl
n
[]
[]
[]
Cl
n
[]
[]
[]

n
[]
. []
[]
Cl

Cl
[]
[]
[]
n
n
[]
Cl
[]
n
Cl
[]
[]
[]
u
Cl
[]
[]
[]
n
Cl
. []
Cl
[]
C3
Cl
n
[3
n
n
Cl
[]


u
Cl
ci
[]
n
n
Cl
u

n
n
Cl
ti
C3
n
n
Cl
n
ci
n
t]
Cl
[j
cj
n
n
Cl
n
[]

Cl
[]
[]
[]
n
Cl
[]
[]
n
Cl
Cl
Cl
n
[]
u
Cl
Cl
n
n
Cl
Cl
Cl
n
n
n
[]
[]
u
n
u
Cl
Cl
n
[]
n
Cl
[]
[]

n
Cl
Cl

APPENDIX B
DEMOGRAPHIC INFORMATION SHEET
ID# DATE
RESPONDENT MOTHER FATHER
SEX MALE FEMALE
DOB
RACE WHITE BLACK HISPANIC OTHER
GRADE NONE PRESCHOOL KIND GRADE
FAMILY COMPOSITION MOTHER & FATHER
MOTHER ONLY
FATHER ONLY
MOTHER & STEPFATHER
FATHER & STEPMOTHER
RELATIVES
OTHER
MOTHER'S OCCUPATION
MOTHER'S EDUCATIONAL LEVEL (YEARS)
FATHER'S OCCUPATION
FATHER'S EDUCATIONAL LEVEL (YEARS)
DIAGNOSIS
DATE OF DIAGNOSIS / /
NUMBER OF PRIOR HOSPITALIZATIONS (APPROX)
ESTIMATED NUMBER HOURS/DAY SPENT WITH CHILD
96

APPENDIX C
BEHAVIORAL UPSET IN MEDICAL PATIENTS
CHILD SELF-REPORT VERSION (CHILD-BUMP)

This girl doesnt like waiting.
$ DO YOU:
never like OR sometimes not
waiting like waiting
4
3
ITEM 1
This girl doesnt mind waiting.
DO YOU:
hardly ever OR never mind
mind waiting waiting
2
1

99
O
o
o

ITEM 2
This girl cries a little.
DO YOU:
never cry OR hardly ever cry
1
2
This girl cries a lot.
DO YOU:
sometimes cry OR always cry
3
4
100

101

ITEM 3
This girl gets mad a lot.
ARE YOU:
always OR sometimes mad
mad
4
3
This girl doesnt get mad a lot.
ARE YOU:
hardly ever mad OR never mad
2
1
102


ITEM 4
This girl usually isnt grouchy.
ARE YOU:
never OR hardly ever
grouchy grouchy
1
9
This girl is usually grouchy.
ARE YOU:
sometimes OR
grouchy
always
grouchy

105

ITEM 5
This girl wont talk.
DO YOU:
not talk OR only sometimes
at all not talk
4
3
This girl doesnt mind talking.
DO YOU:
usually not OR
mind talking
never mind
talking

107

ITEM 6
This girl doesnt say shes sad.
DO YOU:
never say OR hardly ever say
youre sad youre sad
1
2
This girl says shes sad.
DO YOU:
sometimes say OR always say
youre sad youre sad
3
4
108

109

ITEM 7
This girl has to be told many times what to do.
DO YOU:
always OR
have to be
told many
times what
to do
sometimes have
to be told
many times
what to do
4
3
This girl doesnt have to be told many times
what to do.
DO YOU:
hardly ever OR
have to be
told many times
what to do
never have
to be told
many times
what to do

Ill

ITEM 8
This girl would rather do other things.
WOULD YOU:
always OR sometimes
rather do rather do
other other
things things
1
2
This girl would rather sleep than do other things.
WOULD YOU:
sometimes
rather sleep
than do other
things
OR always
rather sleep
than do other
things

113

ITEM 9
This girl has to be told to follow
hospital rules.
DO YOU:
always OR
have to be
told to
follow
hospital
rules
sometimes
have to be
told to
follow
hospital
rules
4
3
This girl doesnt have to be told to follow
hospital rules.
DO YOU:
hardly ever have OR never have to
to be told to be told to
follow hospital follow hospital
rules rules
114

115

This girl doesnt look worried.
DO YOU:
never OR hardly ever
look look worried
worried
1
2
ITEM 10
This girl usually looks kind of worried.
DO YOU:
sometimes look OR always look
worried worried
3
4


ITEM 11
This girl usually looks kind of sad.
DO YOU:
always OR sometimes
look sad look sad
4
3
This girl doesnt usually look sad.
DO YOU:
hardly ever OR
look sad
never
look sad

119

This girl usually does what others want.
DO YOU:
always
do what
others
want
OR sometimes
do what
others
want
1
2
This girl doesnt do what others want.
DO YOU:
hardly ever OR
do what
others
want
never
do what
others
want

121

ITEM 13
This girl needs her parents to make her
feel better.
This girl doesnt need her parents to make her
feel better.
DO YOU:
DO YOU:
always OR sometimes
need your need your
parents to parents to
make you make you
feel better feel better
hardly ever OR never
need your need your
parents to parents to
make you make you
feel better feel better
4 3
2 1
122

123

This girl doesnt talk about things that bug her.
DO YOU:
ITEM 14
never talk
about
things that
bug you
1
OR hardly ever
talk about
things that
bug you
2
This girl talks about a lot of things that bug her.
DO YOU:
sometimes talk OR always talk
about things about things
that bug you that bug you
3
4
124

125

This girl has to have her own way.
DO YOU:
always OR sometimes
have to have to
have your have your
own way own way
4
3
ITEM 15
This girl doesnt have to have her own way.
DO YOU:
hardly ever
have to
have your
own way
OR never
have to
have your
own way
2
1
126

127
i
O

ITEM 16
This girl does a lot of things.
This girl doesnt do a
DO YOU:
DO YOU:
always OR sometimes
do a lot of do a lot of
things things
hardly ever OR
do a lot of
things
1
2
3
lot of things.
never
do a lot of
things
4
128

129

This girl tricks people to get what she wants,
wants.
DO YOU:
always OR sometimes
trick trick
people people
4
3
ITEM 17
This girl doesnt trick people to get what she
DO YOU:
hardly ever
trick
people
2
OR never
trick
people
1
130

131

ITEM 18
This girl doesnt want a lot of things.
This girl wants a lot of
things.
DO YOU:
DO YOU:
never OR
hardly ever
sometimes OR
never
want a lot
want a lot
want a lot
want a lot
of things
of things
of things
of things
1
2
3 4
132

133

This girl doesnt sleep well.
DO YOU:
never
sleep
well
OR hardly ever
sleep
well
ITEM 19
This girl sleeps well.
DO YOU:
usually OR always
sleep sleep
well well
2 l
134

135

This girl can ask others for help.
DO YOU:
always ask OR sometimes
others for ask others
help for help
1 2
ITEM 20
This girl cant ask others for help.
DO YOU:
hardly ever OR never ask
ask others others for
for help help
3
4

137

ITEM 21
This girl doesnt try to make friends.
DO YOU:
never try OR hardly ever
to make try to make
friends friends
4
3
This girl tries to make friends.
DO YOU:
sometimes try OR always try
to make friends to make friends
2
1
138

139

This girl listens when others try to help her.
DO YOU:
ITEM 22
This doesnt listen when others try to help her.
DO YOU:
always OR
sometimes
hardly ever
OR never listen
listen when
listen when
listen when
when others
others try
others try
others try
try to help
to help
to help
to help
1
2
3
4
140

141

ITEM 23
This girl starts talking to other people.
DO YOU:
This girl waits for other people to start talking to her.
DO YOU:
always
OR sometimes
sometimes
wait for
wait for
start talking
other
other
to other people
people to
people to
start
start
talking
talking
4
3
2
always
start talking
to other people
142

143

This girl laughs and smiles at funny things.
DO YOU:
always OR
laugh and
smile at
funny things
sometimes
laugh and
smile at
funny things
1
2
ITEM 24
This girl doesnt laugh and smile at funny things.
DO YOU:
hardly ever
laugh and
smile at
funny things
OR never
laugh and
smile at
funny things
3
4
144

145

ITEM 25
This girl doesnt do what she is told to do. This girl does what she
DO YOU: DO YOU:
never do OR
what you
are told
to do
hardly ever do
what you
are told
to do
sometimes do OR
what you
are told
to do
4
3
2
is told to do.
always do
what you
are told
to do
1
146

147

ITEM 26
This girl is nice
to be with.
This girl isnt
nice to be with.
ARE YOU:
ARE YOU:
always OR
sometimes
hardly ever
OR never
nice to be
nice to be
nice to be
nice t
with
with
with
with
1
2
3
4
148

149

This girl doesnt try to get better in the hospital.
DO YOU:
ITEM 27
This girl tries to get better in the hospital.
DO YOU:
never OR
hardly ever
sometimes
OR always
try to get
try to get
try to get
try to get
better in
better in
better in
better in
the hospital
the hospital
the hospital
the hospital
4
3
2
1
150

151

APPENDIX D
DATA COLLECTION PROCEDURES TRAINING GUIDE
(Make sure to present all of the following points:)
Hi. My name is I'm conducting a study with the
clinical and health psychology department. We're
looking at how children react to being hospitalized.
We'd like to ask you to participate in this study, if
you agree, I WILL first ask you a few general questions
about your child and their living situation. Then I
will give you some questionnaires about your child's
reactions and about your own feelings since coming to
the hospital. These questionnaires will take about
30-45 minutes to complete. Also, we'll ask your
child's nurse to fill out a brief questionnaire about
their opinion of your child's reaction. While you're
completing the questionnaires, we'll ask your child to
answer some questions about how they have been feeling
during their stay.
(At this point, you must make an assessment of the
child's ability to participate. Determine:
A. is the child sleeping?
1. if yes, does parent mind if we wake them up?
a. if yes, can we come back in an hour?
B. is the child sedated?
1. if yes, are they too sedated to participate?
a. if yes, will it wear off in an hour?
C. is the child developmentally delayed?
1. if yes, show some questions to parent and
ask would child be able to understand them?
D. is the child hearing or sight impaired?
1. if yes, can child see/hear well enough to
participate?
If the final answers to questions A, B, C, and D are
no, then we won't be able to evaluate child. Thank
parent for time and cooperation, explaining that we
must have data from both the child and the parent.
152

153
If the child can participate, then say:)
We are also hoping to get more information before your
child is discharged. If your child is still here for
at least another day we'll try to catch you before you
leave and ask you and your child to do the
questionnaires again. We hope that with this project
we can learn which children are at risk to develop
problems in the hospital. We are also trying to find
better tests to use to make it easier to help
hospitalized children in the future. All information
will be kept confidential. Would you be interested in
participating?
(Encourage them to participate as much as possible.
Make sure to check with child that they are also
interested in participating.
If they agree, then give them the informed consent
form, skimming and highlighting the sections of the
form. On the last page, ask the parent to sign it and,
if the child is able and is between 7-17, have them
sign it. Then you sign as witness. Give them a copy
of the consent form and tell them that, should they
have any questions, they can contact me at the phone
number on the form.
Assign the family a subject number, write this number
on the consent form and remaining forms, and administer
the background information questions.
At this point, assess the parent's literacy. If the
parent has less than a tenth grade education, read the
questions aloud immediately after getting the
background information form. If the parent has a
higher educational level, then give them the three
questionnaires, explaining the directions for each one.
Then, go to the first questionnaire and say:)
Okay. Why don't we start with the first question.
Which answer seems best for how your child has reacted
since they arrived?
(Watch them complete the first three or four questions
and observe if they are experiencing any difficulty.
If they take too long, say in a easy voice:)
Well, it's probably just faster/easier for me to just
read them and I'll check off your answers. Does ...

154
(and quickly read the first question not already
completed. If you're at ease, the parent will think
this is normal procedure and won't be uncomfortable.
Once the parent questionnaires are underway, introduce
the study to the child, saying the following more or
lessbe sure to adapt wording to child's age level:)
As I told your mom (and dad), I'm going around talking
to the kids in the hospital asking them how they feel
about being here. I'll be asking you some serious
questions which will be kept secret. We want to try to
find out what makes kids have a hard time in the
hospital so that we can help other kids when they have
to come here. There's no right or wrong answers to the
questions. I just want to find out how you feel since
you came to the hospital. When we're done, I have some
prizes you can choose from to thank you for helping us.
Okay? First, I'll show you two pictures and ask you to
pick which picture is most like how you have been
feeling in the hospital...
(Continue by reading the questions, periodically
reminding them that they are to chose the picture which
is most like them in the hospital.
If the child is between 8 and 12, then administer the
CDI. The revised directions should be :)
Now I'm going to read to you some groups of sentences.
For each group, pick one sentence that describes you
best since you got to the hospital. Remember, there is
no right or wrong answer. Just pick the sentence that
best describes the way you have been since you got
here. Here is an example of how it works.
(Then read the sample question and continue with the
remaining items. Again, remind them periodically that
they are to chose the sentence that describes them
since they got to the hospital.
When finished with the CDI, select two appropriate toys
for the child and ask them to choose one. Thank them
for their help.
Check to see if parent is done with questionnaires. If
not, decide if you want to wait in the room or come
back later to pick them up. If done, thank them
profusely. Remind them that someone will be back to
collect information before the child is discharged if
they are in the hospital for at least another day.
Thank them again on your way out.

155
Now that you're done with the family, find out who the
child's nurse is. Track her down and ask her how long
she has been working with the child. In order to
complete the form, she must have been working with the
child for at least eight hours or approximately one
shift. If she has, write down the estimated number of
hours of contact and give her the form to complete,
reading the instructions.

BIOGRAPHICAL SKETCH
My parents were born and raised just outside of Havana,
Cuba, and they came to America in the 1960s looking for
a better place for themselves and their children. I
was born in Newark, New Jersey, but I spent most of my
life in Miami, Florida. I graduated from the
University of Miami and decided to pursue my graduate
work at the University of Florida. When I am not
working, I enjoy painting and crafts, writing and
reading, and the outdoors.
156

I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Sheila M. Eyberg', Chair
Professor of
Clinical and Health Psychology
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Stephen R. Boggs, Cochair
Assistant Professor of
Clinical and Health Psychology
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy. ^ (]/
James R. Rodric
Assistant Professor of
Clinical and Health Psychology
I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully, adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Assistant Professor of
Clinical and Health Psychology
Michael E. Geisser

I certify that I have read this study and that in
my opinion it conforms to acceptable standards of
scholarly presentation and is fully adequate, in scope
and quality, as a dissertation for the degree of Doctor
of Philosophy.
Professor of Nursing
This dissertation was submitted to the Graduate
Faculty of the College of Health Related Professions
and to the Graduate School and was accepted as partial
fulfillment of the requirements for the degree of
Doctor of Philosophy.
August 1993
Dean, College of Health
Related Professions
Dean, Graduate School

UNIVERSITY OF FLORIDA
3 1262 08554 8294



137


APPENDIX B
DEMOGRAPHIC INFORMATION SHEET
ID# DATE
RESPONDENT MOTHER FATHER
SEX MALE FEMALE
DOB
RACE WHITE BLACK HISPANIC OTHER
GRADE NONE PRESCHOOL KIND GRADE
FAMILY COMPOSITION MOTHER & FATHER
MOTHER ONLY
FATHER ONLY
MOTHER & STEPFATHER
FATHER & STEPMOTHER
RELATIVES
OTHER
MOTHER'S OCCUPATION
MOTHER'S EDUCATIONAL LEVEL (YEARS)
FATHER'S OCCUPATION
FATHER'S EDUCATIONAL LEVEL (YEARS)
DIAGNOSIS
DATE OF DIAGNOSIS / /
NUMBER OF PRIOR HOSPITALIZATIONS (APPROX)
ESTIMATED NUMBER HOURS/DAY SPENT WITH CHILD
96


13
(CDRS; Poznanski, Cook, & Carroll, 1979), appears to be
as limited as the CDI. A study of migraine patients
found no difference in CDRS scores compared to a
control group of children (Cunningham, McGrath,
Ferguson, Humphreys, D'Astous, Latter, Goodman, &
Firestone, 1987). A closer examination of the CDRS
administered to pediatric cancer patients found
significant overlap between depressive symptoms and
impairment due to illness (Heilgenstein & Jacobsen,
1988). Thus, the authors found that measures which
include somatic symptoms may overestimate the presence
of depression.
Instruments which target emotional upset in
pediatric populations are few in number. The
Observation Scale of Behavioral Distress (Jay et al.,
1983) interprets behavioral distress as behaviors
indicative of anxiety and pain. Similarly, an
observational pain rating scale for children aged 2-6
includes some ''depression-like'' items (Gauvain-Piquard,
Rodary, Rezvani, & Lemerle, 1987). However, both
observation rating scales were based on cancer patients
and both scales feature an emphasis on pain behavior
rather than on depressive symptoms. In contrast, a
parent rating scale, the Behavioral Upset in Medical
PatientsRevised (BUMP-R; Saylor, Pallmeyer, Finch,
Eason, Trieber, & Folger, 1987), focuses on emotional


CHAPTER 4
DISCUSSION
Background Variables Affecting Distress
The first purpose of the present study was to
examine variables influencing depression and adjustment
in hospitalized children aged 4-12. The findings
indicated no significant gender differences on any of
the distress or depression measures. This result is
consistent with the bulk of the literature (e.g., Jay
et al., 1983; Saylor et al., 1987). However, it did
not support the trend observed in the pilot study for
girls to obtain higher BUMPR-Hospital scores than boys.
A comparison of the pilot sample with the current
sample revealed significantly more females in the
latter sample. When the two samples were combined, no
significant sex differences were found on the BUMPR-
Hospital form, for mothers or fathers. In fact, the
two studies cited earlier that suggested that girls
exhibit more distress (Katz et al., 1980; Melamed &
Siegel, 1975) consisted of samples with proportionately
more boys. Consequently, when the sample is more
evenly represented by males and females, any sex
differences in depressive symptoms for pediatric
populations may disappear.
69


129


Abstract of Dissertation Presented to the
Graduate School of the University of Florida in
Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
MULTIMETHOD ASSESSMENT OF DEPRESSION AND BEHAVIORAL
DISTRESS IN A PEDIATRIC POPULATION
By
Christina M. Rodriguez
August 1993
Chair: Sheila Eyberg, Ph.D.
Cochair: Stephen R. Boggs, Ph.D.
Major Department: Clinical and Health Psychology
Growing acceptance of the existence of childhood
depression is apparent in recent literature. Despite
this scrutiny, research on childhood depression
contains mixed results largely because of assessment
issues. Studies obtain information from various
sources (e.g., parents, children) and utilize different
definitions of depression. Assessment of preschool
children is further complicated because of skepticism
about depression at this age and because of the
scarcity of measures appropriate for this age group.
The study of depression in pediatric psychology
translates into adjustment in medical settings.
Although pediatric populations seem more likely to
exhibit depressive symptoms, measures suitable for
vii


34
correlated with the Children's Depression Inventory
(Leon et al., 1980). Moreover, the Depression scale
correlated with social withdrawal, depression,
uncommunicativeness, and social subscales of the Child
Behavior Checklist (Kelly, 1982). Test-retest
reliability was reported as ranging from .80 to .94 and
internal consistency (coefficient alpha) of .86 (Wirt
et al., 1984).
The State Trait Anxiety Inventory (STAI, Form Y)
(Spielberqer, 1983). The STAI consists of two
self-report scales measuring state anxiety and trait
anxiety. The form used in this study is a revision of
an earlier version (Spielberger, Gorsuch, & Lushene,
1970). In contrast to the earlier version which had
items related to depression, Form Y measures feelings
of anxiety discriminated from symptoms of depression.
The State Anxiety scale consists of 20 statements about
how the subject feels "right now, at this moment." The
Trait Anxiety scale consists of 20 statements about how
the subject generally feels. Subjects are asked to
rate the intensity of feelings on a Likert scale
ranging from (1) not at all, (2) somewhat, (3)
moderately so, (4) almost always. The following sample
item is from the Trait Anxiety scale:


81
The Child-BUMP was moderately significantly
correlated with the parent report on the BUMPR-Hospital
form. Thus, children and mothers demonstrated some
agreement regarding the child's emotional distress,
although they appear to hold different perspectives
given that the correlation was moderate. As would be
expected, the Child-BUMP, which assesses behavioral
upset in the hospital, did not correlate significantly
with the mothers' BUMPR-Home ratings, giving some
initial indication of discriminative validity of the
measure. In addition, the Child-BUMP was not
associated with the parent's report on the PIC-D,
possibly because the PIC-D was not intended for use in
medical populations. Similarly, the CDI was not
significantly correlated with the Child-BUMP. This may
be attributed to the physical symptoms and school-
related items appearing on the CDI which would not be
appropriate for hospitalized, physically-ill children.
Finally, the Child-BUMP was not significantly
correlated with the Nurse-BUMP, which suggests low
agreement between these two sources.
Implications
The present study ventured into several relatively
unexplored areas in pediatric psychology, including
perspectives from various sources, incorporating self-


19
found that the total number of days spent in previous
hospitalizations predict higher parent ratings of their
child's emotional distress during the current
hospitalization (Saylor et al., 1987). In contrast,
Jay et al. (1983) suggest that children may habituate
to painful medical procedures because behavioral
distress was negatively correlated with number of
previous medical procedures. With regard to dental
visits, no relationship was found between behavior
during the visit and history of unpleasant medical
experiences (Johnson & Baldwin, 1968). Thus, these
mixed results indicate that the relationship between
prior medical experience and emotional adjustment is
not well understood.
Onset of illness has also not been extensively
investigated. Time since diagnosis was significantly
negatively correlated with behavioral distress during a
medical procedure (Jay et al., 1983), further
supporting the idea that children habituate to aversive
medical procedures. Nevertheless, chronicity of
illness as a variable in hospital adjustment has not
been studied. A related concept, length of hospital
stay, has also not received much research attention.
In an adult medical population, length of
hospitalization was not related to psychological
adjustment (Levenson, Hamer, Silverman, Rossiter,


ITEM 21
This girl doesnt try to make friends.
DO YOU:
never try OR hardly ever
to make try to make
friends friends
4
3
This girl tries to make friends.
DO YOU:
sometimes try OR always try
to make friends to make friends
2
1
138


APPENDIX C
BEHAVIORAL UPSET IN MEDICAL PATIENTS
CHILD SELF-REPORT VERSION (CHILD-BUMP)


39
hospital admission. This delay allowed approximately
24 hours of behavior for both the parents and children
to assess reaction to hospitalization (time since
hospitalization ranged from 17 to 35 hours). Following
the collection of demographic information, parents were
instructed to complete the BUMP-R and PIC-D based on
their child's behavior and the STAI on their own
feelings since their child's hospitalization. As the
parents completed their forms, the research assistant
read aloud the Child-BUMP to all children and read the
CDI to children ages 8-12. To reward cooperation, the
child was allowed to select a small prize (e.g.,
sticker, toy car, puzzle book). After the data was
collected from the parent and child, the nurse
completed the BUMPR-Hospital form provided that the
nurse had interacted with the child for a minimum of
eight hours.


ITEM 2
This girl cries a little.
DO YOU:
never cry OR hardly ever cry
1
2
This girl cries a lot.
DO YOU:
sometimes cry OR always cry
3
4
100


133


75
A number of illness-related variables were also
examined. Diagnostic group differences in depression
and behavioral distress were not found in the current
study. For this relatively understudied aspect of
adjustment, categorization by type of illness and by
chronicity of illness failed to support significant
group differences, contrary to expectation. These
findings are comparable to those in the pilot study and
in a study of cancer patients (Jay et al., 1983).
However, given the wide variety of diagnoses obtained
in the current sample, categorization of illness was
difficult and group sizes were possibly too small for
adequate comparison. An ideal study would compare
several clearly definable illness groups which might
then yield differences in behavioral upset and
depression.
The number of prior hospitalizations was expected
to be positively associated with distress, although the
results did not confirm this hypothesis. Previous
medical experience was not correlated with behavioral
distress in the current study or in the pilot study.
Results from previous research have been mixed, with
some investigations suggesting greater distress for
frequently hospitalized patients (e.g., Saylor et al.,
1987) and other investigations suggesting habituation
and decreased upset with more hospital visits (e.g.,


ITEM 8
This girl would rather do other things.
WOULD YOU:
always OR sometimes
rather do rather do
other other
things things
1
2
This girl would rather sleep than do other things.
WOULD YOU:
sometimes
rather sleep
than do other
things
OR always
rather sleep
than do other
things


31
In addition, 62 of the diagnoses were categorized
on chronicity based on Nelson's Textbook of Pediatrics
(Behrman et al., 1987) description of pathogenesis.
Thirty-five percent of diagnoses were described as
acute illnesses and 65% of diagnoses were identified as
chronic illnesses. The eight children not classified
on chronicity had undiagnosed illnesses.
Forty percent of the sample of children had never
been previously hospitalized, 12% of the children had
been hospitalized once before, whereas 43% had been
hospitalized on multiple occasions (ranging from 2 to
20 prior hospitalizations), with prior hospitalizations
M = 3.2, SD = 5.1. The time spent hospitalized during
the study ranged from 1 day to 29 days (M = 5.0, SD =
4.8) with 54% hospitalized for three days or less. The
duration of illness associated with their diagnosis
ranged from newly diagnosed to diagnosis at birth, with
41% diagnosed within one month, and an additional 20%
within six months. The average number of hours the
parent spent with the child in the hospital in a 24
hour period ranged from 3 hours to 24 hours a day, with
67% of parents reporting they spent 24 hours a day with
their child and 17% spending between 18 and 23 hours
with their child.
Thirty-two of the children involved in the study
also received behavior ratings from their nurses. In


4
demonstrated adequate correlations between raters
(e.g., in normal children, Leon, Kendall, & Garber,
1980; in pediatric population, Eason, Finch, Brasted, &
Saylor, 1985). Despite the general finding of low or
poor correspondence, most researchers in the area of
depression conclude that multiple sources are required
in order to clarify the diagnostic picture (e.g.,
Kaslow & Racusin, 1990; Kazdin, 1988). In fact,
Kazdin, Colbus, & Rodgers (1986) describe the
application of discriminant analyses which combines a
battery of measures to maximize the classification,
i.e., diagnosis, of depression in children. Therefore,
in spite of the poor correspondence between raters,
current research encourages assessment of childhood
depression from multiple sources.
Risk Factors
In addition to variability due to assessment
methods, a number of variables may affect the
prevalence of childhood depression, including gender
and age of the child and demographic characteristics,
such as race, socioeconomic status, and family
composition. A fair amount of research has
investigated the effect of sex on depression. In
childhood depression, no gender differences emerge for
children ages 6-12 (Angold, 1988; Kaslow & Racusin,
1990). The prevalence, however, increases in females


24
at home (See Appendix A). This scale is a revision of
the adult version (Zeldow & Braun, 1985) consisting of
a 32-item checklist of behaviors that nonpsychiatric
patients may exhibit in hospital settings. Patient
behaviors were rated by nurses on a Likert scale
ranging from 0 to 4. This scale range indicates the
frequency of behavior, with 0 representing "never and
4 representing "always."
Saylor et al. (1987) revised the scale for use
with children by having five judges independently
evaluate which items were inappropriate for children.
Those items deemed inappropriate by a majority of
judges were deleted, yielding a 28-item scale. A
sample item is:
Looks depressed or sad Never
Sometimes
Often
Usually
Always
Scoring of the BUMP-R parallels the original adult
version. However, instead of ratings by nurses,
parents initially rate the child's behavior in the
hospital followed by ratings of the same behaviors at
home. Therefore, the BUMP-R provides a parent rating
of the child's behavioral upset in the hospital and at
home, prior to hospitalization.


119


3
distinguish emotions (Stone & Lamenek, 1990). Of those
socioemotional measures available for assessment of
preschool children, self-report measures are rare
(Lichtenstein, Dreger, & Cattell, 1986). Consequently,
preschool assessment relies heavily on parent report.
For school age children, a larger variety of assessment
techniques for depression are utilized, including
self-report questionnaires, interview rating scales,
parent ratings, and observational measures (see
examples of observational measures in dental
populations, Johnson & Baldwin, 1968 and Wright &
Alpern, 1971; in medical populations, see Jay, Ozolins,
Elliot, & Caldwell, 1983 and Katz, Kellerman, & Siegel,
1980). Rating scales are completed by a variety of
sources, including parents, peers, and teachers.
Given the variability in sources, correspondence
among raters becomes a significant concern. A
comprehensive review of correspondence issues in
measurement of both behavioral and emotional childhood
problems found relatively low correspondence among
raters (Achenbach, McConaughy, & Howell, 1987).
Several other studies have documented low correlations
between self-reported depression and parent ratings of
depression (e.g., Kazdin, 1989; Kazdin, French, Unis, &
Esveldt-Dawson, 1983; see Kaslow & Racusin, 1990;
Kazdin, 1987, 1988 for reviews). A few studies have


145


147


these children are scarce. The utility of a new parent
report measure, the Behavioral Upset in Medical
Patients-Revised (BUMP-R), was examined in a pilot
study of 81 mothers, which found increased behavioral
distress upon hospitalization. Research on variables
influencing adjustment has found that gender, race,
socioeconomic status, and family composition are not
significant factors. Greater distress may appear with
younger children, maternal anxiety, and limited parent
contact during hospitalization. Effects of diagnosis,
duration of illness, length of hospitalization, and
previous hospitalizations are unclear.
The current study explored variables influencing
depressive symptoms in hospitalized children aged 4-12.
Parent ratings of child distress were compared to nurse
ratings and to children's responses to a pictorial
measure designed in this study for hospitalized
preschoolers. The day following hospital admission, an
assessment battery containing standard measures of
depression and measures of distress for hospitalized
children was administered to 70 mothers and their
children. Thirty-two nurse ratings were also obtained.
Results indicated that demographic and illness-
related variables were not risk factors for hospital
adjustment difficulties. Based on parent ratings,
children exhibiting behavioral distress at home may be
viii


83
Premorbid psychological functioning, as measured
by the BUMPR-Home rating, was strongly related to
distress upon hospitalization. Therefore, a child's
emotional and behavioral difficulties outside the
hospital is predictive of emotional upset upon
hospitalization. Future research should assess
children prior to hospitalization (e.g., for scheduled
procedures) to determine which children are likely to
experience adjustment difficulties. For children
entering the hospital for unscheduled procedures, a
brief assessment battery shortly after hospital
admission could be investigated to facilitate
identification of children who will likely experience
difficulties in the hospital.
One intriguing direction for pediatric psychology
research moves toward evaluating the influence of the
context of health care delivery on adjustment. For
instance, studies should more closely examine the
impact of a family's access to health care. Parental
experience of the availability of health care for their
children may affect their attitudes towards medical
intervention. Many families may also encounter poor
continuity and follow-up by health care providers.
Availability may also affect a family's health care
orientation, i.e., their attitude towards health care
and medical professionals. Belief systems also likely


90
Lichtenstein, D., Dreger, R. M., & Cattell, R. B.
(1986). Factor structure and standardization of the
Preschool Personality Questionnaire. Journal of
Social Behavior and Personality. 1, 165-181.
Melamed, B. G., & Siegel, L. J. (1975). Reduction of
anxiety in children facing hospitalization and
surgery by use of filmed modeling. Journal of
Consulting and Clinical Psychology. 43. 511-521.
Mullins, L. L., Siegel, L. J., & Hodges, K. (1985).
Cognitive Problem-Solving and life-event correlates
of depressive symptoms in children. Journal of
Abnormal Child Psychology. 13, 305-314.
Myers, J., & Bean, L. (1968). A decade later: A follow
up of social class and mental illness. New York:
Wiley & Sons.
Olson, R. A., Holden, E. W., Friedman, A., Faust, J.,
Kenning, M., & Mason, P. J. (1988). Psychological
consultation in a children's hospital: An evaluation
of services. Journal of Pediatric Psychology. 13,
479-492.
Peterson, L., Mori, L., & Carter, P. (1985). The role
of the family in children's responses to stressful
medical procedures. Journal of Clinical Child
Psychology. 14, 98-104.
Pfeffer, C. R., & Trad, P. V. (1988). Sadness and
suicidal tendencies in preschool children.
Developmental and behavioral pediatrics. 9, 86-88.
Poznanski, E. D., Cook, S. C., & Carroll, B. J. (1979).
A depression rating scale for children. Pediatrics.
M, 442-450.
Prugh, D. G., Staub, E. M., Sands, H. H., Kirschbaum,
R. M., & Lenihan, E. A. (1953). A study of the
emotional reactions of children and families to
hospitalization and illness. American Journal of
Orthopsychiatry. 23. 70-106.
Reynolds, W. M., Anderson, G., & Bartell, N. (1985).
Measuring depression in children: A multimethod
assessment investigation. Journal of Abnormal Child
Psychology. 13, 513-526.


135


15
The effect of age on the adjustment of pediatric
patients contrasts with findings in the general
childhood depression literature. In the general
population, the prevalence of depressive symptomatology
increases through childhood and adolescence. However,
in a study conducted with cancer patients, age was the
strongest predictor of distress during medical
procedures, with younger children exhibiting greater
distress (Jay et al., 1983). Another study of a
pediatric cancer population aged 1-17 also found
younger children expressed more distress during a
painful medical procedure (Katz et al., 1980).
Additional support for the inverse relationship between
age and distress was found when younger hospitalized
children were rated by parents as more distressed
(Saylor et al., 1987). Furthermore, an interaction
between sex and age may occur. In contrast to the
trend in general child depression for adolescent
females to demonstrate more depressive symptoms,
younger females may exhibit more distress in pediatric
settings (e.g., Melamed & Siegel, 1975). Overall,
younger pediatric patients appear more likely to
display behaviors suggestive of depressive symptoms and
emotional distress.
Information on the influence of other demographic
variables (such as race, socioeconomic status, or


Analyses of Current Study and Pilot Study
Combined 53
Comparison of Samples 53
Analysis of Combined Sample 54
Factor Analysis of BUMPR-Hospital 56
Analyses of Fathers for Combined Sample 61
4 DISCUSSION 69
Background Variables Affecting Distress 69
Evaluation of the BUMP-R 77
Evaluation of the Child-BUMP 80
Implications 81
REFERENCES 8 6
APPENDICES
A BEHAVIORAL UPSET IN MEDICAL PATIENTSREVISED
(BUMP-R) 94
B BACKGROUND INFORMATION SHEET 9 6
C BEHAVIORAL UPSET IN MEDICAL PATIENTS-CHILD
SELF-REPORT VERSION (CHILD-BUMP) 97
D DATA COLLECTION PROCEDURES TRAINING GUIDE 152
BIOGRAPHICAL SKETCH 156
v


ITEM 26
This girl is nice
to be with.
This girl isnt
nice to be with.
ARE YOU:
ARE YOU:
always OR
sometimes
hardly ever
OR never
nice to be
nice to be
nice to be
nice t
with
with
with
with
1
2
3
4
148


ITEM 25
This girl doesnt do what she is told to do. This girl does what she
DO YOU: DO YOU:
never do OR
what you
are told
to do
hardly ever do
what you
are told
to do
sometimes do OR
what you
are told
to do
4
3
2
is told to do.
always do
what you
are told
to do
1
146


80
For the combined samples of the pilot and current
studies, Cronbach's coefficient alpha for the BUMPR-
Hospital rating was .87. Given this strong internal
consistency, the most suitable and parsimonious
solution in the factor analysis of the BUMPR-Hospital
scores was based on the quartimax rotation. Four
factors emerged, with the largest number of items
representing Negativity/Agitation. The second factor
was labelled as Amiability, the third factor as
Dysphoria, and the fourth factor as Noncompliance.
This factor structure appears to adequately capture
those dimensions involved in adjustment to
hospitalization.
Evaluation of Child-BUMP
A self-report pictorial measure, the Child-BUMP,
was designed for the current study to tap emotional
distress in hospitalized children as young as four
years old. Based on a two-step selection process,
children identify pictures reflecting their level of
distress since hospitalization. Most children appeared
to understand the instrument, particularly with
repetition of items. Cronbach's coefficient alpha for
the Child-BUMP was computed as .76, which indicates
moderately strong internal consistency.


123


Ill


53
parent report on the BUMPR-Home form was not
significantly correlated with the Child-BUMP, Nurse-
BUMP, or parental state or trait anxiety. Similarly,
the PIC-D was not significantly related with the Child-
BUMP, Nurse-BUMP, or parental state anxiety. The
Child-BUMP was also not significantly associated with
the CDI, Nurse-BUMP, or parental state or trait
anxiety. The CDI was not significantly correlated with
the Nurse-BUMP or parental state or trait anxiety.
Moreover, contrary to the hypothesis that younger
children would demonstrate more distress, age and grade
were not significantly correlated with the outcome
variables. This finding does not replicate the modest
correlation of age and BUMPR-Hospital scores revealed
in the pilot study. Lastly, SES was not significantly
related to scores on the BUMPR-Hospital, Child-BUMP,
CDI, Nurse-BUMP, or parental state or trait anxiety.
Analyses of Current Study and
Pilot Study Combined
Comparison of Samples
Based on the results obtained from the pilot
study, a difference had been anticipated in the current
study between maternal ratings of distress at
hospitalization and maternal ratings of distress at
home. Moreover, the effects of age and gender were not
observed in the current study compared to the pilot.


89
Kazdin, A. E. (1989). Identifying depression in
children: A comparison of alternative selection
criteria. Journal of Abnormal Child Psychology, 17,
437-454.
Kazdin, A. E., Colbus, D., Rodgers, A. (1986).
Assessment of depression and diagnosis of depressive
disorder among psychiatrically disturbed children.
Journal of Abnormal Child Psychology. 14, 499-515.
Kazdin, A. E., French, N. H., Unis, A. S., &
Esveldt-Dawson, K. (1983). Assessment of childhood
depression: Correspondence of child and parent
ratings. Journal of the American Academy of Child
Psychiatry. 22, 157-164.
Kerr, M. M., Holer, T. S., Versi, M. (1987).
Methodological issues in childhood depression: A
review of the literature. American Journal of
Orthopsychiatry. 57, 193-198.
Kovacs, M. (1983). The Children's Depression Inventory:
A self-rated depression scale for school-aged
youngsters. Unpublished manuscript, University of
Pittsburg School of Medicine.
Kovacs, M. (1985). The Children's Depression Inventory
(CDI). Psvchopharmacoloav Bulletin. 21. 995-998.
Lachar, D., & Gdowski, C. L. (1979). Actuarial
assessment of child and adolescent personality: An
interpretive guide for the Personality Inventory for
Children profile. Los Angeles, CA: Western
Psychological Services.
Lefkowitz, M. M., & Tesiny, E. P. (1980). Assessment of
childhood depression. Journal of Consulting and
Clinical Psychology, 48, 43-50.
Leon, G. R., Kendall, P. C., Garber, J. (1980).
Depression in children: Parent, teacher, and child
perspectives. Abnormal Child Psychology. 8, 221-235.
Levenson, J. L., Hamer, R., Silverman, J. J., Rossiter,
L. F. (1986-87). Psychopathology in medical
inpatients and its relationship to length of
hospital stay: A pilot study. International Journal
of Psychiatry in Medicine, 16, 231-236.


87
Douglas, J. W. B. (1975). Early hospital admission and
later disturbances of behaviour and learning.
Developmental Medicine and Child Neurology, 17,
456-480.
Eason, L. J. Finch, A. J., Jr., Brasted, W., & Saylor,
C. F. (1985). The assessment of depression and
anxiety in hospitalized pediatric patients. Child
Psychiatry and Human Development. 16, 57-64.
Finch, A. J., Jr., & Saylor, C. F. (1984). An overview
of child depression. In W. J. Burns & J. V. Lavigne
(Eds.), Progress in pediatric psychology (pp.
201-239). Orlando, FL: Grue & Stratton.
Finch, A. J., Jr., Saylor, C. F., Edwards, G. L.
(1985). Children's Depression Inventory: Sex and
grade norms for normal children. Journal of
Consulting and Clinical Psychology. 53., 424-425.
Gauvain-Piquard, A., Rodary, C., Rezvani, A., &
Lemerle, J. (1987). Pain in children aged 2-6 years:
A new observational rating scale elaborated in a
pediatric oncology unitpreliminary report. Pain.
31, 177-188.
Gorsuch, R. L. (1983). Factor analysis (2nd Ed.).
Hillsdale, NJ: Lawrence Erlbaum Assoc.
Harter S., & Pike, R. (1983). Procedural manual to
accompany the Pictorial Scale of Perceived
Competence and Social Acceptance for Young Children.
University of Denver.
Heilgenstein, E., & Jacobsen, P. B. (1988).
Differentiating depression in medically ill children
and adolescents. Journal of the American Academy of
Child & Adolescent Psychiatry. 27, 716-719.
Jay, S. M., Ozolins, M., Elliott, C. H., & Caldwell, S.
(1983). Assessment of children's distress during
painful medical procedures. Health Psychology. 2,
133-147.
Jensen, R. A. (1955). The hospitalized child: Round
Table, 1954. American Journal of Orthopsychiatry.
25, 293-318.
Jessner, L., Blom, G. E., & Waldfogel, S. (1952).
Emotional implications of tonsillectomy and
adenoidectomy of children. Psychoanalytic Study of
the Child. 7, 126-169.


10
corresponding to early hospitalization, including
conduct problems, delinquency, academic difficulty, and
unstable work history. Another study of pediatric
surgery patients found that, without intervention,
children displayed more behavioral difficulties nearly
one month after hospitalization (Melamed & Siegel,
1975). This early research demonstrated that children
often adjust poorly to hospitalization or medical
procedures.
Prevalence
The study of poor adjustment to hospitalization is
susceptible to similar difficulties in estimating
prevalence as those difficulties in general childhood
depression. Different age groups, definitions, and
assessment techniques are utilized. Although one study
found nearly 90% of pediatric surgery patients
displayed behavioral difficulties following
hospitalization (Prugh et al., 1953), most estimates of
depressive symptoms in pediatric populations are more
conservative. In children with physical handicaps or
chronic illness, poor adjustment is reported as ranging
from 13-26% based on parent report, approximately two
times the rate found in healthy children (Wallander,
Varni, Babani, Banis, & Wilcox, 1988). In a review of
studies using diagnostic criteria (derived from DSM-III
criteria), estimates of depressive symptoms in both


ACKNOWLEDGMENTS
First and foremost, I must thank two individuals
who helped transform an idea into reality: Steve Boggs,
for his guidance, incredible patience, unwavering
support, and long hours; and Sheila Eyberg, as a female
role model of unparalleled professionalism as well as
for her steadfast enthusiasm and faith. Also, I thank
Jim Rodrigue, Mike Geisser, and Faye Gary-Harris, whose
excitement about this project energized me and whose
suggestions, comments, and questions facilitated the
direction and critical evaluation of the project. I
thank all of these committee members for helping me
pull this together.
Secondly, I thank my mother not only for her
invaluable assistance in creating the Child-BUMP
pictures but also for instilling in me a deep
compassion for human suffering that led me to the field
of clinical psychology. I also thank Becky, Brigette,
Elena, Jennifer, Melodye, and Randi, my close friends
and fellow graduate school survivors, for the emotional
support that sustained me and for redefining the depths
of friendship. I also appreciate the help of other
friends, faculty, and staff at the Dept, of Clinical
ii


6
Only a few studies have demonstrated sex
differences in depressive symptomatology for
prepubertal children. For instance, minor differences
were found in which girls scored higher than boys on a
different self-report measure (Children's Depression
Scale; Reynolds et al., 1985). A study of 8- to 13-
year-olds also found a trend for girls to report more
depressive symptoms than boys on the CDI (Seligman,
Peterson, Kaslow, Tanenbaum, Alloy, & Abramson, 1984).
These conflicting findings may reflect variability in
assessment or in populations, or they may indicate that
observers respond differentially based on the child's
sex (Saylor, Finch, Spirito, & Bennett, 1984). Despite
these mixed results, the bulk of the literature
continues to support the 1:1 sex ratio for depression
in prepubertal children.
As mentioned above, an interaction between sex and
age appears, leading to an increased incidence of
depression in adolescent girls (Kazdin, 1987, 1988). A
significant interaction effect between gender and age
was found in a study of over 1000 schoolchildren, with
adolescent girls reporting the most unhappiness in a
structured interview (Webb & VanDevere, 1985).
Although depressive disorder is relatively uncommon in
prepubertal children, the prevalence of depression
rises with increasing age (Angold, 1988; Kaslow &


115


This girl doesnt talk about things that bug her.
DO YOU:
ITEM 14
never talk
about
things that
bug you
1
OR hardly ever
talk about
things that
bug you
2
This girl talks about a lot of things that bug her.
DO YOU:
sometimes talk OR always talk
about things about things
that bug you that bug you
3
4
124


ITEM 9
This girl has to be told to follow
hospital rules.
DO YOU:
always OR
have to be
told to
follow
hospital
rules
sometimes
have to be
told to
follow
hospital
rules
4
3
This girl doesnt have to be told to follow
hospital rules.
DO YOU:
hardly ever have OR never have to
to be told to be told to
follow hospital follow hospital
rules rules
114


ITEM 7
This girl has to be told many times what to do.
DO YOU:
always OR
have to be
told many
times what
to do
sometimes have
to be told
many times
what to do
4
3
This girl doesnt have to be told many times
what to do.
DO YOU:
hardly ever OR
have to be
told many times
what to do
never have
to be told
many times
what to do


5
throughout adolescence and approaching adulthood
(Kazdin, 1987, 1988). Consequently, the sex ratio of
depression for females to males nears 2:1 by adulthood
(Boyd & Weissmann, 1981).
Nevertheless, for prepubertal children, several
studies utilizing various assessment instruments have
corroborated the 1:1 sex ratio. Among these studies,
scores on a clinician-administered interview rating
scale revealed no sex differences (Shanahan,
Zolkowski-Wynne, Coury, Collins, & O'Shea, 1987). In
addition, normative data on over 1400 schoolchildren
grades 2-8 for one of the most widely-used self-report
depression scales, the Children's Depression Inventory
(CDI; Kovacs, 1983, 1985), found negligible gender
differences, and thus separate norms were not
recommended (Finch, Saylor, & Edwards, 1985). This
finding was substantiated when no gender differences on
the CDI were found in a smaller sample of 166 children
grades 3-6 (Reynolds, Anderson, & Bartell, 1985). A
study of parent-reported childhood depression also
reflects this equal prevalence for children grades 3-6
(Leon et al., 1980; Reynolds et al., 1985).
Furthermore, no statistically significant sex
differences were found on a measure of peer-nominated
depression (Lefkowitz & Tesiny, 1980).


36
Trait Anxiety items for their general feeling since
their child was hospitalized.
The Children's Depression Inventory (CPI) (Kovacs.
1983, 1985; Finch et al., 1985). The CDI is a 27-item
self-report depression scale appropriate for children
ages 8 and older. Each item consists of three
statements representing graded levels of severity of a
depressive symptom and the child selects one of the
three statements. A sample item is:
I feel like crying everyday.
I feel crying many days.
I feel like crying once in a while.
The choices are valued from 0 to 2, with high scores
indicating depression. The internal consistency of the
CDI (coefficient alphas) was reported as ranging from
.70 to .94 (Kovacs, 1985; Saylor et al., 1984).
Test-retest reliability ranged from correlations of .38
to .87 (Saylor et al., 1984).
The Behavioral Upset in Medical Patients-Child
Self-Report Version (Child-BUMP). The Child-BUMP is a
27-item pictorial scale designed for use in this study
(See Appendix C). All questions from the BUMP-R for
the hospital setting (with one exception) were
rephrased in language understandable for children ages
4-12. The questionnaire was rephrased by three
independent sources and the simplest items selected.


LIST OF TABLES
TABLE Page
1Means and Standard Deviations of the Outcome
Measures 42
2 Demographic Differences for the BUMPR-Hospital.. 44
3 Demographic Differences for the PIC-Depression.. 45
4 Demographic Differences for the Child-BUMP 46
5 Demographic Differences for the CDI 47
6 Demographic Differences for the Nurse-BUMP 48
7 Spearman Correlations Between Illness-Related
Variables and Outcome Measures 50
8 Correlations Among Outcome Measures and
Demographic Variables 52
9 Item-Total Correlations for the BUMPR-Hospital.. 57
10 Eigenvalues for the BUMPR-Hospital Factor
Analysis 59
11 Factor Structure for the BUMPR-Hospital 62
12 Item Loadings for Factor OneNegativity/
Agitation 63
13 Item Loadings for Factor TwoAmiability 64
14 Item Loadings for Factor ThreeDysphoria 65
15Item Loadings for Factor FourNoncompliance.... 66
vi


91
Ryan, N. D., Puig-Antich, J., Ambrosini, P.,
Rabinovich, H., Robinson, D., Nelson, B., Iyengar,
S., & Twomey, J. (1987). The clinical picture of
major depression in children and adolescents.
Archives of General Psychiatry. 44, 854-861.
Saylor, C. F., Finch, A. J., Jr., & McIntosh, J. A.
(1988). Self-reported depression in psychiatric,
pediatric, and normal populations. Child Psychiatry
& Human Development. 18., 250-254.
Saylor, C. F., Finch, A. J., Jr., Spirito, A., &
Bennett, B. (1984). The Children's Depression
Inventory: A systematic evaluation of psychometric
properties. Journal of Consulting and Clinical
Psychology. 52, 955-967.
Saylor, C. F., Pallmeyer, T. P., Finch, A. J., Jr.,
Eason, L., Trieber, F., & Folger, C. (1987).
Predictors of psychological distress in hospitalized
pediatric patients. Journal of the American Academy
of Child and Adolescent Psychiatry. 26, 232-236.
Seligman, M. E. P., Peterson, C., Kaslow, N. J.,
Tanenbaum, R. L., Alloy, L. B., & Abramson, L. Y.
(1984). Attributional style and depressive symptoms
among children. Journal of Abnormal Psychology. 93.,
235-238.
Shanahan, K. M., Zolkowski-Wynne, J., Coury D. L.,
Collins, E. W., O'Shea, J. S. (1987). The Children's
Depression Rating Scale for normal and depressed
outpatients. Clinical Pediatrics. 26. 245-247.
Shaw, E. G., & Routh, D. K. (1982). Effect of mother
presence on children's reaction to aversive
procedures. Journal of Pediatric Psychology. 7,
33-42.
Spielberger, C. D. (1983) Manual for the State-Trait
Anxiety Inventory (Form Y, Self-Evaluation
Questionnaire). Palo Alto, CA: Consulting
Psychologists Press.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E.
(1970). Manual for the State-Trait Anxiety Inventory
(Self-Evaluation Questionnaire). Palo Alto, CA:
Consulting Psychologists Press.


76
Jay et al., 1983). The current findings suggest that
previous medical history may be unrelated to the
child's reaction to hospitalization. The current study
included children with a wide variety of diagnoses and
medical histories whereas the Jay et al. (1983) study
involved a sample of children with cancer. Perhaps the
quality of previous medical experience (i.e., negative
or positive perceptions) is more influential in a
child's adjustment to hospitalization than frequency of
prior hospitalizations.
Contrary to the hypothesis, duration of illness
and length of hospital stay were not positively
correlated with the child's emotional distress.
Although previous research has not studied these
variables closely, children with chronic illness and
longer hospitalizations were expected to obtain higher
scores on measures of depression and behavioral upset.
However, both the current study and pilot study did not
find these relationships. One interesting negative
association between maternal state anxiety and duration
of illness emerged. The concept of habituation may
account for this relationship, with mothers reporting
less anxiety as they become accustomed to the chronic
nature of their child's illness. With regard to length
of hospitalization, most children were hospitalized for
short periods and greater variability in length of


125


and Health Psychology, whose daily contributions to my
well-being and training made accomplishing this project
meaningful. I thank the group of research assistants
who helped with data collection and turned what
initially seemed to be an insurmountable daily task
into an organized demonstration of teamwork. Finally,
I thank the parents and nurses who took the energy from
their pressing concerns in order to help the sons and
daughters of tomorrow.
iii


Table 7: Spearman Correlations Between Illness-Related Variables and Outcome Measures
Length of
hospitalization
r (D)
Number of Prior
hospitalizations
r (n)
Duration of
illness
r (n)
Number of Hours
spent w/ parent
r (n)
BUMPR-Hospital
.14 (70)
.04 (70)
.02 (64)
. 16
(70)
BUMPR-Home
.13 (70)
.21 (70)
.14 (64)
. 05
(70)
PIC-Depression
.04 (70)
.11 (70)
.00 (64)
. 05
(70)
Child-BUMP
-.14 (70)
-.04 (70)
-.17 (64)
.29
(70)*
CDI
.00 (33)
.09 (33)
-.08 (31)
. 01
(33)
Nurse-BUMP
-.15 (32)
-.15 (32)
-.29 (30)
. 04
(32)
STAI-State
.20 (70)
-.15 (70)
-.30 (64)*
-.26
(70)a
STAI-Trait
.03 (65)
.09 (65)
-.17 (59)
-.13
(65)
* E < -01
a Because the significance level was reduced to .01, the marginal relationship was found at
only p < .05.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EPPA501SU_AXTJVN INGEST_TIME 2015-04-14T18:34:45Z PACKAGE AA00029761_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


92
Stone, W. L., & Lamenek, K. L. (1990). Developmental
issues in children's self-reports. In A. M. Lagreca
(Ed.), Through the eves of a child (pp. 18-56).
Boston: Allyn & Bacon.
Taylor, J. A. (1953). A personality scale of manifest
anxiety. Journal of Abnormal and Social Psychology.
48, 285-290.
Vardaro, J. A. (1978). Preadmission anxiety and
mother-child relationships. Journal of the
Association for the Care of Children in Hospitals.
7, 8-15.
Walker, L. S., & Greene, J. W. (1989). Children with
recurrent abdominal pain and their parents: More
somatic complaints, anxiety, and depression than
other patient families? Journal of Pediatric
Psychology. 14. 231-243.
Wallander, J. L., Varni, J. W., Babani, L., Banis, H.
T., & Wilcox, K. T. (1988). Children with chronic
physical disorders: Maternal reports of their
psychological adjustment. Journal of Pediatric
Psychology. 13, 197-212.
Webb, T. E., & VanDevere, C. A. (1985). Sex differences
in the expression of depression: A developmental
interaction effect. Sex Roles. 12, 91-95.
Wirt, R. D., Lachar, D., Klinedinst, J. K., & Seat, P.
D. (1984) Multidimensional description of child
personality: A manual for the Personality Inventory
for Children. Los Angeles, CA: Western Psychological
Services.
Wisniewski, J. J., Naglieri, J. A., & Mulick, J. A.
(1988). Psychometric properties of a Children's
Psychosomatic Symptom Checklist. Journal of
Behavioral Medicine. 11, 497-507.
Worchel, F. F., Nolan, B. F., Willson, V. L., Purser,
J. S., Copeland, D. R., & Pfefferbaum, B. (1988).
Assessment of depression in children with cancer.
Journal of Pediatric Psychology. 13, 101-112.
Wright, G. Z., & Alpern, G. D. (1971). Variables
influencing children's cooperative behavior at the
first dental visit. Journal of Dentistry for
Children. 38, 124-128.


16
family composition) on adjustment in pediatric
populations is even more limited than that found in
research on general childhood depression. Of the few
studies available, appropriate behavior in a dental
setting was associated with socioeconomic status, with
upper SES children exhibiting less negative behavior
(Wright & Alpern, 1971). Influence of race and family
composition are not known and thus may correspond to
data obtained on general childhood depression.
Consequently, future research should explore the
influence of these demographic variables on the
emotional adjustment of pediatric patients and whether
these influences correspond to those found in general
childhood depression.
Several other variables related to pediatric
populations may also affect adjustment to
hospitalization. These variables include maternal
anxiety, parental presence in the hospital, diagnosis,
prior medical experience, onset of illness, and length
of hospitalization. With regard to maternal anxiety,
mothers' self-report of anxiety was positively related
to their children's negative behavior during dental
visits (Johnson & Baldwin, 1968; Wright & Alpern,
1971). Moreover, mother's state anxiety was positively
correlated with a physiological measure of anxiety in
pediatric patients (Vardaro, 1978). Maternal anxiety


11
inpatient and outpatient pediatric samples ranged from
7-40%, with the wide variability possibly due to
different age ranges and different medical populations
(Finch & Saylor, 1984). In a study of 7- to 12-year-
old hospitalized children, 38% exhibited dysphoric mood
(Kashani, Barbero, & Bolander, 1981). An examination
of a pediatric psychology service in a children's
hospital found 19% of consultations were referred for
depression or suicide attempts and 12% of the
consultations were referred for adjustment to chronic
illness (Olson, Holden, Friedman, Faust, Kenning, &
Mason, 1988). Thus, pediatric populations may be more
likely to display depressive symptoms than the general
child population.
Assessment Issues
As mentioned above, the assessment of pediatric
patients' adjustment involves symptoms from overlapping
constructs and may include symptoms of major
depression, dysphoric mood, behavioral concomitants of
emotional upset, and symptoms related to anxiety.
Assessment in pediatric populations shares difficulties
similar to the assessment of depression in the general
population. Problems with correspondence among raters
appear in the pediatric psychology literature as well
as in the general childhood depression literature.
Rating scales are typically administered to parents,


ITEM 4
This girl usually isnt grouchy.
ARE YOU:
never OR hardly ever
grouchy grouchy
1
9
This girl is usually grouchy.
ARE YOU:
sometimes OR
grouchy
always
grouchy


149


84
influence the perceived acceptability of medical
intervention. In addition, social support systems may
contribute to health care orientation as well as
provide assistance to parent and child during health
care crises. These variables may affect the parent and
child's perspective and adjustment when encountering a
given hospitalization.
With regard to assessment of pediatric
populations, overall, the BUMP-R appears to be a useful
instrument in the assessment of emotional distress in
hospitalized children, with concurrent validity, high
internal consistency, and appropriateness for various
raters. Further investigation of the BUMP-R should
evaluate the consistency of distress ratings over time
as well as comparing scores of hospitalized children
with outpatient pediatric samples in order to ascertain
the specific effect of hospitalization. The current
multimethod investigation utilized self-report, parent-
report, and nurse-report. Correspondence between
raters should be further explored, focusing on
interrater reliabilities between mothers, fathers,
children, and health personnel (e.g., nurses, doctors).
In addition, the Child-BUMP warrants further
investigation regarding its utility for pre-literate
hospitalized children. Items should be further
examined to ensure the suitability for these young


88
Johnson, R., & Baldwin, D. C. (1968). Relationship of
maternal anxiety to the behavior of young children
undergoing dental extraction. Journal of Dental
Research. 47, 801-805.
Kandel, D. B., & Davies, M. (1982). Epidemiology of
depressive mood in adolescents: An empirical study.
Archives of General Psychiatry. 39., 1205-1212.
Kashani, J. H., Barbero, G. J., & Bolander, F. D.
(1981). Depression in hospitalized pediatric
patients. Journal of the American Academy of Child
Psychiatry, 20. 123-134.
Kashani, J. H., Holcomb, W. R., & Orvaschel, H. (1986).
Depression and depressive symptoms in preschool
children from the general population. American
Journal of Psychiatry. 143. 1138-1143.
Kashani, J. H., Husain, A., Shekim, W. O., Hodges, K.
K., Cytryn, L., & McKnew, D. H. (1981). Current
perspectives on childhood depression: An overview.
American Journal of Psychiatry, 138. 143-153.
Kaslow, N. J., & Racusin, G. R. (1990). Childhood
depression: Current status and future directions. In
A. S. Bellack, M. Hersen, & A. E. Kazdin (Eds.),
International handbook of behavior modification and
therapy (2nd ed., pp. 649-667). New York: Plenum
Press.
Kaslow, N. J., Rehm, L. P., Pollack, S. L., & Siegel,
A. W. (1988). Attributional style and self-control
behavior in depressed and nondepressed children and
their parents. Journal of Abnormal Child Psychology,
16, 163-175.
Katz, E. R., Kellerman, J., & Siegel, S. (1980).
Behavioral distress in children with cancer
undergoing medical procedures: Developmental
considerations. Journal of Consulting and Clinical
Psychology, 48. 356-365.
Kazdin, A. E. (1987). Assessment of childhood
depression: Current issues and strategies.
Behavioral Assessment. 9, 291-319.
Kazdin, A. E. (1988). Childhood depression. In E. J.
Mash & L. G. Terdal (Eds.), Behavioral assessment of
childhood disorders (2nd ed., pp. 157-195). New
York: Guilford Press.


30
Based on diagnoses reported by mothers, 77% of the
children were divided into six diagnostic groups based
on Nelson's Textbook of Pediatrics (Behrman et al.,
1987). Of those diagnoses categorized, 19% of the
children were hospitalized for cardiovascular or
respiratory difficulties (e.g., cystic fibrosis,
cardiac myopathy, coronary artery disease). Twenty-two
percent of the children were diagnosed with immunity,
allergy, or related diseases (e.g., HIV, asthma,
juvenile rheumatoid arthritis) and 20% percent of the
children were experiencing difficulties involving the
digestive system (e.g., cleft lip and palate,
alpha-antitrypsin deficiency, gastroenteritis). An
additional 13% of the children were diagnosed with
urinary difficulties (e.g., nephrotic syndrome, urinary
tract infection, bladder infection). Thirteen percent
of the patients had been hospitalized for diseases
affecting the nervous system (e.g., spina bifida,
seizure disorders). For another 13% of the children,
diagnosis had not been determined (e.g., fever of
unknown origin). The sixteen children not classified
into diagnostic groups had received a variety of
diagnoses, preventing assignment to a group of adeguate
number for analyses. These unclassified diagnoses
included viral meningitis, diabetes, snakebite, and
sickle cell anemia.


This girl doesnt sleep well.
DO YOU:
never
sleep
well
OR hardly ever
sleep
well
ITEM 19
This girl sleeps well.
DO YOU:
usually OR always
sleep sleep
well well
2 l
134


45
Table 3: Demographic Differences for the PIC-Depression
N Mean (SD) t
Males
29
62.1
(13.9)
.59
Females
41
60.0
(15.9)
Acre Group
4-8
32
63.2
(18.2)
1.20
8-12
38
58.9
(11.6)
Race
White
46
60.8
(16.2)
. 04
African-American
19
60.6
(13.2)
Family Composition
2-Parent
44
58.3
(12.8)
1.70
1-Parent
23
64.9
(18.7)
Note: All t values are nonsignificant.


42
Table 1: Means and Standard Deviations of the Outcome
Measures
N
Mean
(SD)
BUMPR-Hospital
70
27.2
(13.3)
BUMPR-Home
70
24.7
(11.2)
PIC-D
70
60.9
(15.0)
STAI-State
70
55.3
(12.2)
STAI-Trait
65
52.2
(10.4)
Child-BUMP
70
51.5
(8.4)
CDI
33
6.9
(5.1)
Nurse-BUMP
32
28.6
(18.8)
for the BUMP-R scores or for the new Child-BUMP. The
sample mean on the PIC-D (M = 60.9, SD = 15.0) was one
standard deviation above the normative sample mean,
suggesting that mothers reported more depressive
symptomatology in this population than in the general
population. However, sample means on parent report of
anxiety on both the STAI State scale (M = 55.3, SD =
12.2) and the STAI Trait scale (M = 52.2, SD = 10.4)
were within normal limits. In addition, the current
sample mean on the CDI (M = 6.9, SD =5.1) is
comparable to the mean reported for newly diagnosed
diabetics (Kovacs, 1983) and is below the recommended
cut-off for diagnosis of depression.


70
Contrary to the hypothesis, younger children did
not obtain higher distress scores in the current study.
Many previous studies have found an inverse
relationship between age and depressive symptoms in
pediatric populations (e.g., Jay et al., 1983; Katz et
al., 1980; Saylor et al., 1987). Moreover, age was
found to be negatively correlated with BUMPR-Hospital
scores in the pilot study. Thus, the current study did
not confirm the negative correlation between age and
distress found in the pilot study. Again, a comparison
revealed that children in the pilot study were
significantly younger than those in the current sample.
With both samples combined, a low but significant
negative correlation was found between age and mothers'
BUMPR-Hospital ratings. The magnitude of this age
correlation is comparable to that found in the study
that introduced the BUMP-R (Saylor et al., 1987).
Thus, younger children may show more distress upon
hospitalization, although the relationship does not
appear strong.
As anticipated, no significant differences were
found for race on any of the outcome measures. The
influence of ethnicity on distress in pediatric
populations has not been studied. However, the absence
of racial differences in distress found in the current
study is consistent with research in depression for the


Table 8: Correlations Among Outcome Measures and Demographic Variables
BUMPR-Hosp
r (n)
BUMPR-Home
E (n)
PIC-D
r (n)
Child-BUMP
E (n)
CDI Nurse-BUMP
r (n) r (n)
STAI-State
r (n)
STAI-Trait
E (n)
BUMPR-Home
.53
(70)
PIC-D
.45
(70)
.49 (70)
Child-BUMP
.36
(70)*
.10 (70)
.14 (70)
CDI
.43
(33)
.52 (33)
.54 (33)
.30 (33)
Nurse-BUMP
.39
(32)*
.02 (32)
.04 (32)
.32 (32)
-.17 (14)
STAI-State
.15
(70)
.07 (70)
.22 (70)
-.05 (70)
-.03 (33)
-.16 (32)
STAI-Trait
.14
(65)
.21 (65)
.40 (65)
-.02 (65)
.26 (31)
-.19 (29)
.64 (65)
AGE
-.15
(70)
.04 (70)
-.17 (70)
.04 (70)
.36 (33)* -
-.23 (32)
-
.03 (70)
-.08 (70)
GRADE
-.20
(70)
-.04 (70)
-.22 (70)
-.01 (70)
.30 (33)
-.25 (32)
-
.02 (70)
-.07 (65)
SES
-.02
(70)
.31 (70)*
.36 (70)
.16 (70)
.27 (33)
.29 (32)
.00 (70)
.23 (65)
* n < .01
" a < .001
' Because the significance level was reduced to .01, this marginal relationship was found at only p < .05.


43
The new Child-BUMP measure appeared to function as
designed, with the children readily grasping the two-
step response process. Four-year-old children
occasionally had difficulty with the measure if the
parent had suggested that the child had comprehension
problems. Repeating items was often helpful for
children, particularly for those with more limited
attention spans. Cronbach's coefficient alpha for the
Child-BUMP was .76, suggesting fairly strong internal
consistency. For the Nurse-BUMP, internal consistency
was .93, indicating that the individual items are very
strongly intercorrelated.
Analyses of Background Data
An examination of the demographic variables sex,
age group, race, and family composition indicated no
significant differences for the outcome measures (See
Tables 2-6). No differences had been anticipated based
on sex, race, or family composition. However, a
marginal relationship was found between the CDI scores
and family composition (See Table 5), with children of
single parent homes reporting more depressive symptoms
than children in two-parent homes (t = 1.99, p = .055).
Contrary to the expectation that younger children (ages
4-8) would exhibit more distress than older children
(ages 8-12), no differences between age groups were
found on any of the outcome measures. Analyses also




55
comparable to the children aged 8-12 years on the
BUMPR-Hospital (t(149) = .91, p > .05) and on the
BUMPR-Home (t(149) = 1.19, p > .05). Age evidenced
only a small relationship (r = -.17, p < .05) with
ratings of behavioral distress in the hospital.
There were no differences in the BUMPR-Hospital
scores (F(117) = 1.55, p > .05) or in the BUMPR-Home
scores (F(117) = 1.45, p > .05) for the six diagnostic
groups. The combined sample demonstrated no
differences for chronicity of illness on the ratings of
hospital distress (t(128) = .78, p > .05) or the
ratings of home distress (t(128) = .29, p > .05). No
significant relationship was found between the number
of prior hospitalizations and the BUMPR-Hospital scores
(r = .06, p > .05) or the BUMPR-Home scores (r = .06,
p > .05). Similarly, duration of illness was unrelated
to the BUMPR-Hospital ratings (r = .06, p > .05) or the
BUMPR-Home ratings (r = .05, p > .05).
With this combined sample, maternal ratings of
distress following hospitalization (M = 27.0, SD =
13.1) were significantly higher (t(151) = 3.15, p <
.01) than ratings of distress at home (M = 23.8, SD =
11.2). Consistent with the results of both the pilot
study and the current study, ratings of distress in the
hospital were significantly correlated with ratings of
distress in the home (r = .50, p < .0001).


74
anxiety emerged. Therefore, mothers reported greater
anxiety the more time they spent with their child.
This may reflect parent fatigue or it may suggest that
concern for their child's illness may influence both
amount of contact and anxiety ratings.
Overall, this pattern of results regarding parent
contact does not support the Peterson et al. (1985)
position promoting parental contact in order to avoid
difficulties following separation upon hospitalization.
One salient limitation in drawing conclusions about
parent contact from the current study stems from
subject participation being contingent upon parental
presence. Obviously, those patients who spent little
time with their hospitalized child were particularly
likely to be unavailable for study participation.
Thus, distress in children who could not participate in
this study may indeed be greater with limited parent
contact. In addition, parents themselves reported the
number of hours spent with their child and they may
have exaggerated the amount of time. Indeed, there was
minimal statistical variability in reported parent
contact, and, given the skewed distribution and the
possible influence of outliers, results regarding
amount of time the parent spent with the child should
be interpreted cautiously.


1986-1987). Both variables, onset of illness and
length of hospitalization, warrant further
investigation.
20
Pilot Study
Introduction and Rationale
Given the many unanswered issues regarding
adjustment, a pilot study was conducted to examine the
applicability of a new measure in studying variables
influencing distress in a hospitalized pediatric
population aged 4-12. Few studies have investigated
depressive symptomatology in pediatric populations of
preschool and pre-literate children. Moreover,
measures specifically designed to assess distress in
medical settings are limited.
The pilot study focused on the application of a
new parent rating scale of behaviors associated with
depression and anxiety (Saylor et al., 1987). This
rating scale involves the evaluation of the frequency
of specific behaviors, not a parental interpretation of
emotional distress. Moreover, this brief questionnaire
does not require direct observation by a trained
clinician. Thus, preliminary findings regarding the
influence of background variables on adjustment to
hospitalization were gathered.


26
composition differences were found. There were also no
significant differences in the BUMPR-Hospital ratings
due to diagnosis based on the six categories of illness
(F(63) = 1.12, p > .05). Only marginal differences for
chronicity of illness (t(66) = 1.73, p = .09) were
found on the BUMPR-Hospital, with children diagnosed
with acute illnesses (M = 32.1, SD = 14.4) scoring
higher than those diagnosed with chronic illnesses (M =
26.3, SD = 11.6).
Age was significantly correlated with hospital
distress (r = -.24, p < .05), with younger children
obtaining higher distress ratings than older children.
Behavioral distress in the hospital or at home was not
significantly related to SES, duration of illness, or
number of previous hospitalizations.
Ratings of hospital distress were significantly
associated with home distress (r = .48, p < .0001). In
addition, maternal ratings of their child's behavioral
upset in the hospital (M = 26.8, SD = 13.0) were
significantly higher (t(79) = 2.67, p < .01) than
ratings of distress behaviors at home (M = 23.1, SD =
11.3), suggesting that mothers observed increased
behavioral distress following hospitalization.


This girl tricks people to get what she wants,
wants.
DO YOU:
always OR sometimes
trick trick
people people
4
3
ITEM 17
This girl doesnt trick people to get what she
DO YOU:
hardly ever
trick
people
2
OR never
trick
people
1
130


ITEM 11
This girl usually looks kind of sad.
DO YOU:
always OR sometimes
look sad look sad
4
3
This girl doesnt usually look sad.
DO YOU:
hardly ever OR
look sad
never
look sad


ITEM 18
This girl doesnt want a lot of things.
This girl wants a lot of
things.
DO YOU:
DO YOU:
never OR
hardly ever
sometimes OR
never
want a lot
want a lot
want a lot
want a lot
of things
of things
of things
of things
1
2
3 4
132


27
Purpose of Study and Hypotheses
The purpose of the current investigation was two
fold. First, variables affecting behavioral upset in a
hospitalized population aged 4-12 were further
explored. Second, issues regarding assessment
modalities were studied. Specifically, differences
among raters were investigated as well as a comparison
of measures designed to assess distress in medical
populations with measures more commonly used in the
assessment of depression. As part of the assessment
battery, a newly designed self-report scale appropriate
for pre-literate children in hospital settings was
included. The usefulness of this new measure for
hospitalized children was of particular interest.
Assessment of depression was obtained following
hospital admission via self-reported behavioral
distress and depression, nurse report of the child's
behavioral distress, and parental report of their
child's depression and behavioral distress as well as
the parent's own anxiety in the hospital setting. No
significant differences in the outcome measures due to
sex, race, socioeconomic status, or family composition
(i.e., single v. two-parent homes) were anticipated.
The following hypotheses were tested.
(1) Age and grade differences were expected, with
younger children exhibiting greater emotional distress.