<%BANNER%>

Predicting Treatment Initiation in a Rural, School-Referred Sample

Permanent Link: http://ufdc.ufl.edu/UFE0041769/00001

Material Information

Title: Predicting Treatment Initiation in a Rural, School-Referred Sample
Physical Description: 1 online resource (47 p.)
Language: english
Creator: Matias, Eileen
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: adherence, attendance, child, initiation, psychotherapy, referred, rural, school, treatment
Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Although researchers continue to investigate factors that contribute to treatment attendance and adherence, relatively little is known about what factors influence treatment initiation. This study examines the role of parent-rated symptom severity, and parent-teacher concordance, as predictors of treatment initiation. The current sample consisted of 144 children and adolescents (68% male) referred by their schools for a psychological evaluation in a rural Florida county. One hundred seven families (74%) attended at least one therapy session ( initiators ) while only 37 (36%) families did not attend treatment ( refusers ). The Behavior Assessment System for Children (BASC), Parent and Teacher versions, as well as the Achenbach Child Behavior Checklist and Teacher Report Form, were used to test the relationship between parent-teacher concordance and treatment initiation. The Parent BASC was also used to determine whether higher parent-rated symptom severity predicted greater treatment initiation. Logistic regression analyses were used to test these relationships. Parent-rated symptom severity on the four BASC Composites (Externalizing, Internalizing, Behavior Symptoms Index, and Adaptive Skills) did not significantly predict treatment initiation. Similarly, parent-teacher concordance did not predict treatment initiation in this sample. Implications of these findings are discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Eileen Matias.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Wiens, Brenda A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041769:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041769/00001

Material Information

Title: Predicting Treatment Initiation in a Rural, School-Referred Sample
Physical Description: 1 online resource (47 p.)
Language: english
Creator: Matias, Eileen
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: adherence, attendance, child, initiation, psychotherapy, referred, rural, school, treatment
Clinical and Health Psychology -- Dissertations, Academic -- UF
Genre: Psychology thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Although researchers continue to investigate factors that contribute to treatment attendance and adherence, relatively little is known about what factors influence treatment initiation. This study examines the role of parent-rated symptom severity, and parent-teacher concordance, as predictors of treatment initiation. The current sample consisted of 144 children and adolescents (68% male) referred by their schools for a psychological evaluation in a rural Florida county. One hundred seven families (74%) attended at least one therapy session ( initiators ) while only 37 (36%) families did not attend treatment ( refusers ). The Behavior Assessment System for Children (BASC), Parent and Teacher versions, as well as the Achenbach Child Behavior Checklist and Teacher Report Form, were used to test the relationship between parent-teacher concordance and treatment initiation. The Parent BASC was also used to determine whether higher parent-rated symptom severity predicted greater treatment initiation. Logistic regression analyses were used to test these relationships. Parent-rated symptom severity on the four BASC Composites (Externalizing, Internalizing, Behavior Symptoms Index, and Adaptive Skills) did not significantly predict treatment initiation. Similarly, parent-teacher concordance did not predict treatment initiation in this sample. Implications of these findings are discussed.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Eileen Matias.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Wiens, Brenda A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041769:00001


This item has the following downloads:


Full Text

PAGE 1

1 PREDICTING TREATMENT INITIATION IN A RURAL, SCHOOL REFERRED SAMPLE By EILEEN MATIAS DAVIS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Eileen Matias Davis

PAGE 3

3 To my husband, Rusty, and my wonderful mom and dad for your unconditional love and support

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank Brenda A. Wiens, Ph.D., for her patience, guidance, and support in the development of this project as well as throughout my graduate training. I would also like to thank my lab colleagues, Jennifer E. Rosado, M.A., and Melissa K. Stern, M.S., for their friendship, dependability and willingness to offer assistance or a friendly smile whenever necessary Finally, I would like to extend my deepest love and gratitude to my husband, Rusty, for always encouraging and believing in me; my parents, whose love and support have been a constant presence in my life; and to God, for supplying me with strength, perseverance and a sense of purpose.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 6 ABSTRACT ..................................................................................................................... 7 CHAPTER 1 INTRODUCTION ...................................................................................................... 9 Treatment Initiation ................................................................................................... 9 Study Aims .............................................................................................................. 14 2 METHODS .............................................................................................................. 16 Participants ............................................................................................................. 16 Measures ................................................................................................................ 17 Demographic Data ........................................................................................... 17 Symptom Severity ............................................................................................ 17 Treatment Entry ................................................................................................ 19 Procedures ............................................................................................................. 19 Intake ................................................................................................................ 19 Treatment Initiation ........................................................................................... 20 3 RESULTS ............................................................................................................... 25 Descriptives Parent and Teacher Ratings .............................................................. 25 Parent Teacher Concordance ................................................................................. 25 Treatment Initiation ................................................................................................. 27 Parent perceived Symptom Severity ................................................................ 27 Parent Teacher Concordance .......................................................................... 27 4 DISCUSSION ......................................................................................................... 33 Aim 1 ....................................................................................................................... 33 Aims 2 through 4 ..................................................................................................... 34 Implications ............................................................................................................. 35 Limitations ............................................................................................................... 38 Future Directions .................................................................................................... 41 LIST OF REFERENCES ............................................................................................... 44 BIOGRAPHICAL SKETCH ............................................................................................ 47

PAGE 6

6 LIST OF TABLES Table page 2 1 Demographic characteristics .............................................................................. 22 2 2 Sample items Parent form s ............................................................................. 23 2 3 Sample items Teacher forms ........................................................................... 23 2 4 Psychometric characteristics of BASC and Achenbach scales .......................... 24 3 1 BASC descriptives Initiators vs. refusers ......................................................... 30 3 2 Paired samples T tests Parents vs. teachers .................................................. 30 3 3 Fishers Z test results for BASC parent teacher correlations .............................. 30 3 4 Statistical tests for demographic factors ............................................................. 31 3 5 Logistic regression Parent rated symptom severity ......................................... 31 3 6 Logistic regression Parent teacher difference scores ...................................... 31 3 7 Logistic regression Parent teacher difference scores (absolute value) ........... 31 3 8 BASC vs. Achenbach sca le means .................................................................... 32

PAGE 7

7 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science PREDICTING TREATMENT INITIATION IN A RURAL, SCHOOL REFERRED SAMPLE By Eileen Matias Davis May 2010 Chair: Brenda A. Wiens Major: Psychology Although researchers continue to investigate factors that contribute to treatment attendance and adherence, relatively little is known about what factors influence treatment initiation. This study examines the role of parent rated symptom severity, and parent teacher concordance, as predictors of treatment initiation. The current sample consist ed of 144 children and adolescents (68% male) referred by their schools for a psychological evaluation in a rural Florida county. One hundred seven families (74%) attended at least one therapy ses sion (initiators) while only 37 (36%) families did not attend treatment ( refusers ). The Behavior Assessment System for Children (BASC), Parent and Teacher versions, as well as the Achenbach Child Behavior Checklist and Teacher Report Form were used t o test the relationship between parent teacher concordance and treatment initiation. The Parent BASC was also used to determine whether higher parent rated symptom severity predicted greater treatment initiation. Logistic r egression analyses were used to test these relationships. Parent rated symptom severity on the four BASC Composites (Externalizing, Internalizing, Behavior Symptoms Index, and Adaptive Skills) did not significantly predict treatment initiation.

PAGE 8

8 Similarly, parent teacher concordance di d not predict treatment initiation in this sample. Implications of these findings are discussed.

PAGE 9

9 CHAPTER 1 INTRODUCTION Treatment Initiation Every year approximately 1.5 million children receive a psychological evaluation, and one of the most common recommendations offered to these children as a result of the se evaluations is for psychological intervention ( MacNaughton & Rodrigue, 2001). Initial adherence to this recommendation, however, has been shown to be at or below fifty percent in most studies ( Geffken, Keeley, Kellison, Storch, & Rodrigue, 2006; MacNaughton & Rodrigue, 2001). Moreover, r esearch ers have shown that parental adherence to initial recommendations for psychological treatment is poorer than initial adherence to school based recommendations, self help recommendations (i.e. community support group, reading a self help book), and non psychologic al consultation (i.e. speech therapist, pediatrician) ( Bennett, Power, Rostain, & Carr, 1996; MacNaughton & Rodrigue, 2001; Thibodeau, 2007). The low rates of psychological treatment initiation are alarming, for little can be done to meet many of the men tal health needs of children if recommendations offered subsequent to psychological assessment do not result in initiation of psychological intervention. In recent years, greater effort has been made to identify factors that contribute to treatment attendance and completion. Certain family demographic characteristics, such as ethnic minority status, lower socioeconomic status, and lower levels of educational attainment have been shown to be associated with poorer psychotherapy attendance ( Armbrus ter & Fallon, 1994; Wierzbicki & Pekarik, 1993). Research findings have further suggested that certain parent characteristics can also be important factors in the prediction of treatment attendance. Parent psychopathology and parent perceived

PAGE 10

10 stress, for instance, have been found to predict poorer attendance to child therapy ( Armbruster & Kazdin, 1994; Nock & Kazdin, 2001), while higher levels of parent motivation for treatment have predicted higher treatment persistence ( Nock & Photos, 2006). Moreover, parents beliefs about the effectiveness and credibility of treatment have demonstrated a significant relationship with treatment adherence ( Nock, Ferriter, & Holmberg, 2007). Specifically, parents who, during the first clinic visit, rated treatment as mo re reliable and effective for the treatment of their childrens psychological difficulties were more likely to persist in treatment and adhere to treatment recommendations. Additionally, variables associated with child attributes, such as greater clinicia n rated symptom severity and poorer functioning ( Kazdin, Mazurick, & Bass, 1993; Kazdin, Mazurick, & Siegel, 1994) older age ( Piacentini et al. 1995), and lower rates of self reported symptom severity ( BrookmanFrazee, Gabayan, Haine, & Garland, 2008; Kendall & Sugarman, 1997) have been shown to relate to poorer psychotherapy attendance. A popular model often used to help explain some of the possible barriers that interfere with families willingness or ability to participate in child therapy is the barriers to treatment model developed by Kazdin and his colleagues. The four primary barrier domains that were identified are (1) the experience of specific obstacles and life stressors, (2) a negative or weak relationship with the clinician, (3) the perce ption that treatment is irrelevant to the needs of the child and family, and (4) the belief that the demands of treatment are too great ( Kazdin, Holland, & Crowley, 1997; Kazdin, Holland, Crowley, & Breton, 1997). The third barrier stating that families m ay be less inclined to comply with treatment recommendations if they do not perceive them to be

PAGE 11

11 relevant or necessary w as the basis for the primary focus of the current study Although perceived treatment relevance undoubtedly has multiple components, one of these components is likely a parents belief that a child has significant psychological difficulties (Costello, Pescosolido, Angold, & Burns, 1998) Presumably, parents will be less likely to bring their children in for therapy if they do not perceive their child ren s behaviors or psychological difficulties as a significant concern. However, r esearch focused on whether the severity of child psychopathology as perceived by parents predicts treatment attendance is limited and has yielded some mixed find ings. While some researchers have found that severity of child psychopathology, as perceived by parents, is predictive of treatment seeking behaviors ( Griest, Forehand, Wells, & McMahon, 1980; Gustafson, McNamara, & Jensen, 1994; Hricik & Keane, 1988) others have failed to find this relationship ( BrookmanFrazee, et al., 2008; Kendall & Sugarman, 1997; MacNaughton & Rodrigue, 2001). To help explain these conflicting results, Nock and Ferriter (2005) suggested two alternative ways in which parent perce ived c hild symptom severity may influence treatment initiation and attendance. The first possibility they propose d is that greater perceived symptom severity leads to greater perceived need for treatment, thus increasing the likelihood of treatment initiation and attendance. Conversely, greater perceived symptom severity may be associated with more overall barriers to treatment (i.e., more parenting stress, reduced social support, etc.), thus making it less likely that parents will initiate and persist in treatment. It remains unclear whether either of these explanations offered by Nock and Ferriter holds true, or whether other factors influence the relationship between parent perceived symptom severity and adherence behaviors. Nonetheless, it is

PAGE 12

12 important to understand whether parents perceptions of the degree and severity of their child ren s symptoms influences families attendance and adherence in psychological treatment While it is undoubtedly necessary to continue seeking a better understanding of the factors that interfere with or promote families ability and willingness to persist in treatment once treatment has begun, it is arguably equally important to understand the factors that prevent families from entering psychological treatment in the first place once the need for such treatment has been recommended through psychological assessment. This is especially true in light of evidence that approximately onethird of f amilies who do not complete treatment are often able to benefit from only a few sessions of psychotherapy, even if these families terminate prior to achievement of all treatment goals (Kazdin & Wassell, 1998) While many studies have focused on characteristics associated with attendance and adherence to treatment once treatment i s initiated, less is known about factors that predict whether or not families will initiate treatment following recommendations based on psychological assessment. In one of the few studies that attempt ed to identify factors that contribute to treatment attendance at the initiation phase, MacNaughton and Rodrigue (2001) found that the number of reported barriers to treatment, rather than the specific type of barriers reported, significantly predicted treat ment initiation. This finding is important, as it suggests that by eliminating certain treatment barriers, and thus reducing the total number of barriers faced by families, the likelihood of treatment initiation may be increased. More recently, Keeley & Wiens (2008 ) found that family cohesion, as measured with the Family Environment Scale, significantly predicted whether or not families initially accepted

PAGE 13

13 treatment. Additionally, this study showed that families of children whose primary presenting proble ms were internalizing (e.g., depression, anxiety) were less likely to enter treatment than families of children whose symptoms were primarily externalizing (e.g., behavior problems, ADHD) Although these findings provide some insight into the factors that may influence treatment initiat ion, further research is needed to improve the ability to predict which families will adhere to recommendations to init iate psychological intervention. By identifying these, interventions can be developed to implement during the assessment stage in order to increase the likelihood that families will indeed follow through with initiation of treatment. In light of Kazdins research examining the role of barriers in treatment attendance and adherence, it seems likely that parent perceived symptom severity is an important but insufficient component for determining whether or not families enter treatment. Given that the treatment of childhood psychopathology is often a combined effort among parents, teachers, and other professionals, it may be important for parents to receive validation from these other informants regarding their perceptions of a childs symptom severity. In other words, parents may be more likely to initiate psychological treatment for their child if other important individuals in the childs life are also reporting significant difficulties. Studies looking at the relationship between ratings of multiple informants have consistently shown that there is typically moderate to poor concordance in ratings of child symptom severity across raters including parents and teachers, parents and their children, and to a lesser degree two parents of the same child ( Achenbach, McConaughy, & Howell, 1987; Duhig, Renk, Epstein, & Phares, 2000; Smith, 2007).

PAGE 14

14 Although numerous studies have shown a tendency for poor inter rater agreem ent little is known about t he role that poor concordance among informants may play in determining treatment initiation or adherence. R esearchers have argued that poor agreement between parents and children may play an important role in the poor outcomes that are commonly seen for clinic based child therapy. If parents and their children do not agree on the nature and severity of the problem for which therapy is being sought or on the overall need for therapy, then it would seem likely that they would struggle to work together to achieve established treatment goals. In fact, BrookmanFrazee and colleagues (2008) found that stronger parent child agreement regarding treatment goals predicted better treatment attendance. To date, however, no research has been done to explore the role of parent teacher concordance in determining treatment initiation or attendance. In child therapy, teachers often play an important role, not only in helping to meet the therapeutic goals by implementing strategies in the classroom, but also by helping to identify behavioral and psychological difficulties or by providing corroboration of difficulties in a class room setting. Child therapists must often work closely with these teachers to provide a comprehensive treatment plan that addresses the childs needs in multiple settings. If teachers and parents agree on the degree of child symptom severity, and the need for treatment is thus corroborated across settings, parents may be more inclined to initiate child treatment. Study Aims The first aim of the present study was to provide descriptive information about the degree of parent teacher concordance in a sample of students from a rural county in North central Florida. The present sample is unique in that children were referred for

PAGE 15

15 initial assessment by their teachers or school counselors. Parents, after choosing to accept this initial evaluation, then decided whether or not to follow through with recommendations for treatment. Given that parents were not the initiators of psychological evaluation in this population, as would normally be the case, it was expected that the correlation between parent and teacher ratings would be significantly weaker than the ratings reported in the manual for the Behavior Assessment System for Childre n, the primary measure used to assess perceived s ymptom severity in this study. The second aim of this study was to determine whether parent rated symptom severity predicted treatment initiation. With several important barriers to treatment (i.e. cost, a ccessibility) greatly reduced or eliminated through the U niversity of F lorida F amily S upport S ervices (UF FSS) the relationship between perceived symptom severity and treatment initiation was expected to be significant. The third aim of this study was t o explore whether the relationship between parent and teacher ratings predicted treatment initiation. The expectation was that greater concordance of elevated scores among teachers and parents would predict treatment initiation. Similarly, it was expecte d that children whose parent and teacher both rated their symptom severity in the elevated range would be more likely to initiate treatment. Overall, this study aimed to elucidate how parent perceived symptom severity and agreement between parents and teachers regarding severity of symptoms impact the likelihood that families will follow through with the recommendation to enter therapy for their child.

PAGE 16

16 CHAPTER 2 METHODS Participants Participants were 144 children (46 girls and 98 boys) and their par ents and teachers from a primarily rural county in N orth central Florida. These children were all referred by their schools for a psychological intake assessment and subsequently referred for treatment with the University of FloridaFamily Support Service s (UF FSS) based on the results of this evaluation. Only those families who attended th e initial intake assessment were included in the study. Al together, 495 students were referred for an initial intake assessment. Of these, 359 (72.5%) attended the intake, while the remaining 136 were unreachable by the P roject CAT Ch staff, failed to reschedule missed appointments, or refused to participate in the initial intake. Of the 359 who attended an initial intake, 144 (40.1%) were then referred for treatment with UF FSS. Data were unavailable for families who did not attend the initial intake assessment and for families who were referred for treatment with other providers precluding their inclu sion in the analyses for the current study Demographics The children who were included in this study were elementary, middle, and high school students ranging in age from 6 to 19 years ( M = 11.15, SD = 3.23). Eighty seven percent were Caucasian, 8.3% w ere African American, and the remaining 5% were Asian American, Hispanic American, Native American, or Bi racial Family income data were available for 100 of the 144 families. Of these, 42% reported a yearly household income of less than $20,000, and 75% reported an annual income of less than $40,000. See Table 2 1 for a more detailed break down of family demographic characteristics.

PAGE 17

17 Measures Demographic Data During the intake session, parents completed a general information form to assess demographic factors including child age, gender, ethnicity, family income, and parental marital status Symptom Severity At the time of the initial referral, teachers were given a behavioral measure to complete prior to the scheduled intake assessment date. In addition, parents completed several standardized measures to assess different areas of emotional and behavioral functioning. Two different standardized measures were used to assess child symptom severity as perceived by parents and teachers. Due to a mid project change in measures families who participated in the study during the first half of the data collection period were administered the Achenbach scales, including the Child Behavior Checklist (CBCL) for parents and the Teacher Report Form (TRF) for teachers. In contrast, those who par ticipated during the second portion of the data collection phase completed the Behavior Assessment System for Children, Parent Rating Scale (BASC PRS) and Teacher Rating Scale (BASC TRS). The current study focused prim arily on the BASC with the exception of the final aim, for which the Achenbach scales were also included The BASC PRS and TRS each have three versions one for each of three different age groups. T he forms include a preschool form for ages 2 5, a child form for ages 611, and an adolescent form for ages 1221. Given the very small number of children under the age of 6 who attended an intake session, these children, and hence the 2 5 age range form, were not included in the present study. The BASC scal es have been shown to have excellent psychometric characteristics ( Reynolds &

PAGE 18

18 Kamphaus, 1998) from r = .88 to r = .95 for the parent version and r = .87 to r = .97 for the teacher version in the present sample. The BASC Externalizing Problems, Internalizing Problems, Adaptive Skills, and Behavioral Symptoms Index composites were used to compare parent and teacher ratings, as these are the four composites that are common among the parent and teacher forms. Although the BASC scales w ere the primary measures used in this study, power was increased for the final analysis by combining the BASC parent and teacher scales and the A chenbach scales, allowing for inclusion of a greater number of st udents in this analysis. For the process of combining measures for the final analysis, only the Internalizing Symptoms and Externalizing Symptoms scales for each measure were used, given that these scales are highly comparable across measures. The report ed correlation between the CBCL Internalizing Symptoms Composite Score and the BASC PRS Internalizing Symptoms Composite Score is r = .65. The correlation between the TRF Internalizing Symptoms Composite Score and the BASC TRS Internalizing Symptoms Compo site Score is r = .73. The CBCL and BASC PRS Externalizing Symptoms Composite Scores are reported to correlate at r = .79. Finally, the Externalizing Symptoms Composite Scores for the TRF and BASC TRS are correlated at r = .88 (Reynolds & Kamphaus, 1998) Tables 2 2 and 2 3 offer examples of the overlap between the BASC scales and the Achenbach scales. These scales were combined by coding the T scores from the Internalizing Symptoms and Externalizing Symptoms Composites of the BASC PRS and TRS as well as the CBCL and TRF into dichotomous variables, indicating normal scores or elevated scores (including

PAGE 19

19 borderline/at risk and clinical severity). Following the example of MacNaughton and Rodrigue (2001) a cutoff T score of 60 was used to make the distinction between normal and elevated ratings. Ratings > 60 were classified in the elevated range and coded as 1, while ratings of See Table 24 for psychometric properties of the BASC PRS/TRS and the CBCL/TRF as published in their respective manuals. Treatment Entry After families were offered intervention services through UF FSS, they then had the option of entering treatment or refusing treatment. Families responses to this initial tre atment recomm endation were classified as either accepted treatment if they attended one or more therapy sessions following the intake assessment (coded as 1), or refused treatment if they attended zero therapy sessions (coded as 0). Families who accepted treatment were referred to as treatment initiators and those who did not enter treatment were referred to as treatment refusers. Procedures Intake The participants in this study we re families who were offered services through Project Columbia County Acting Together for Children (CATCh) and the University of Florida Family Support Services (UF FSS), a school linked mental health intervention program in Columbia County, Florida. All procedures were reviewed and approved by the Institutional Review Board01. Fa milies of children who were identified as at risk (i.e., having behavioral, emotional, and/or academic problems) were referred by their school guidance counselors for psychological assessment. Following this initial referral, families came in for a formal evaluation for which graduate student, intern, or post -

PAGE 20

20 doctoral clinicians obtained parental consent as well as child assent. Standardized measures and a clinical interview were used to assess emotional and behavioral functioning from the perspective of the child, the parents, and the childs teacher(s). As part of the initial referral process childrens teachers were contacted and asked to complete a behavioral measure (BASC TRS or TRF) to assess the childs emotional and behavioral symptoms from the t eachers perspectives. These measures were usually completed and returned prior to the scheduled evaluation date. W hen the families came in for the intake assessment, parents were asked to complete several measures (including the BASC PRS or CBCL) to ass ess the childs functioning from the parents perspective. Following this initial intake assessment children were referred for psychological treatment and/or other services based on their needs One hundred and forty four children were referred back to UF FSS and were given the opportunity to enter treatment with UF FSS therapists. Data on treatment entry and treatment course were not available for children who were referred for treatment with other community agencies (e.g., Meridian Behavioral Healthcare, White Foundation, Sexual Assault Center) due to health privacy laws. Treatment Initiation Once the intake assessment was complete, families were offered the opportunity to enter their children in treatment. For children who were referred to services with UF FSS, t hese services were offered to families at no cost and available at the Columbia County School Board building in Lake City, F lorida or at the childs elementary, middle, or high school, depending on the particular needs and preference s of eac h family. Given that many of the families served through Project CATCh live in traditionally

PAGE 21

21 underserved rural communities and that many of these families have limited financial means, the accessible location and free nature of these services largely reduces major obstacles that often serve as barriers to treatment participation. Of the 144 families offered services through UF FSS, 107 (74%) families initiated treatment and 37 (26%) families refused treatment at this time. Reasons for refusing treatment c ould not always be determined, as some families could not be reached to schedule treatment. However, the inability to reach families in order to schedule an initial session or families failure to attend an initial scheduled session are some of the most c ommon reasons for which families do not enter treatment with UF FSS.

PAGE 22

22 Table 21 Demographic characteristics Note: Differences between treatment initiators and treatment refusers were not significant Total Sample (%) N = 144 Treatment Initiators (%) n = 107 Treatment Refusers (%) n = 37 Gender Male Female 68.1 31.9 68.0 32 .0 67.0 33 .0 Ethnicity Caucasian African American Hispanic Asian American Native American Bi Racial 86.8 8.3 1.4 0 .7 0 .7 2.1 87.9 7 .5 1.9 0 .9 0 .9 0 .9 83.8 10.8 0.0 0 .0 0 .0 5 .4 Income < 6,000 6,00010,000 11,00020,000 21,00030,000 31,00040,000 41,00050,000 51,00060,000 5.9 11.8 26.5 23.5 8.8 14.7 5.9 2.9 3.3 15.4 24.2 18.7 16.5 15.4 3.3 3.3 5.9 11.8 26.5 23.5 8.8 14.7 5.9 2.9 Marital Status Married Divorced/Separat ed Living Together Single 48.1 35.7 8.5 7.8 44.2 36.8 10.5 8.4 58.8 32.4 2.9 5.9

PAGE 23

23 Table 22. Sample i tems Parent forms BASC P RS CBCL Externalizing Is overly active Cant sit still, restless, or hyperactive Argues when denied own way Argues a lot Is cruel to animals Cruel to animals Plays with fire Sets fires Threatens to hurt others Threatens people Runs away from home Runs away from home Internalizing Worries Worries Cries easily Cries a lot Is shy with other children Too shy or timid Is sad Unhappy, sad, or depressed Says, I want to kill myself. Talks about killing self Complains of dizziness Feels dizzy or lightheaded Table 2 3 Sample i tems Teacher forms BASC TRS TRF Externalizing Steals at school Steals Disrupts the schoolwork of other children Disturbs other pupils Threatens to hurt others Threatens people Breaks other childrens things Destroys property belonging to others Has friends who are in trouble Hangs around with others who are in trouble Throws tantrums Temper tantrums or hot temper Internalizing Refuses to talk Refuses to talk Says, Im afraid I will make a mistake. Is afraid of making mistakes Is sad Unhappy, sad, or depressed Complains of pain Aches or pains ( not stomach or headaches) Says, Im not very good at this. Self conscious or easily embarrassed Is nervous Nervous, high strung, or tense

PAGE 24

24 Table 2 4. Psychometric characteristics of BASC and Achenbach scales BASC PRS CBCL BASC TRS TRF # of items 138 118 148 118 Item response range 0 to 3 0 to 2 0 to 3 0 to 2 .69 .89 .78 .94 .70 .94 .72 .95 Internalizing Scales 84 .91 90 82 .90 90 Externalizing Scales 92 .94 94 94 .95 95 Notes: BASC PRS: Behavior Assessment System for Children Parent Rating Scale; CBCL: Child Behavior Checklist; BASC TRS: Behavior Assessment System for Children Teacher Rating Scale; TRF: Teacher Report Form

PAGE 25

25 CHAPTER 3 RESULTS DescriptivesParent and Teacher Ratings Parent and teacher BASC forms were scored using the general normative sample as the reference group. Overall, parents rated their childrens symptoms as more severe than did teachers on all four composite scales, with an average difference among informant s of 6.19 for treatment initiators and 7.64 for treatment refusers. The mean parent ratings for Externalizing Problems, Internalizing Problems, and the Behavioral Symptoms Index were all at or above the clinical cut off of 60 (indicating at risk or greate r severity) for both groups. In contrast, mean teacher ratings on these composites were all at or below this cut off. BASC PRS and BASC TRS means and standard deviations for each group are presented in Table 3 1 To test whether these differences between parents and teachers were significant, pairedsamples t tests were conducted. Results indicate d that parents rated their childrens symptoms as significantly worse than did teachers on the Externalizing Problems scale [Treatment initiators: t (47) = 4. 47, p < .001; Treatment refusers: t ( 20) = 3.30, p < .01] and the Behavioral Symptoms Index [Treatment initiators: t (49) = 4.66, p < .001; Treatment refusers: t (21) = 3.62, p < 01] for both groups (See Table 32 ). Differences between informants were not significant for Internalizing Problems [Treatment initiators: t (49) = 1.65, p = .11 Treatment refusers: t (21) = 1.73, p = .10] or Adaptive Skills [Treatment initiators: t (51) = .28, p = .78 ; Treatment refusers: t (21) = .48, p = .64]. Parent Teacher Concordance Fishers Z tests were used to test the first hypothesis that parent teacher concordance would be lower in the current sample as compared to the normative

PAGE 26

26 sample as reported in the BASC manual ( Reynolds and Kamphaus, 1998; see Vianna, 1980 and Brandner, 1933 for detailed descriptions of the Fishers Z test). The Fishers Z test takes into account differences in sample sizes, allowing for comparison between the current study sample and the sample used to develop BASC norms. Becaus e the goal of this analysis was to compare the inter rater correlations in the current sample as a whole to the normative sample, treatment initiat o rs and refusers were combined for this analysis. However, separate analyses were conducted depending on the form that was used for youths of different ages (child form versus adolescent form), as the BASC manual reports these correlations separately Sample size and correlations for the study sample and BASC normative sample, as well as the Fishers Z and pva lues for each test, are reported in Table 33 On the child form, parent teacher correlations ranged from .01 to .39. Fishers Z test results show that correlations from the study sample were not significantly different from the correlations reported in the BASC manual for Externalizing Problems (Fishers Z = .08, p = .42), Internalizing Problems (Fishers Z = .11, p = .91), or the Behavioral Symptoms Index (Fishers Z = 1.69, p = .09). Conversely, the correlations on the Adaptive Skills Index were signi ficantly different across samples (Fishers Z = 2.77, p < .01) with a stronger parent teacher correlation in the normative sample (.48) than in the study sample (.01) On the adolescent form, significant differences across samples were found for the Internalizing Problems composite scale only (Fishers Z = 2.21, p < .05). Responses of parents and teachers from the study sample (.63) correlated more strongly than responses of parents and teachers from the normative sample (.32).

PAGE 27

27 Treatment Initiation Pa rent perceived Symptom Severity To test the hypothesis that higher parent rated symptom severity would predict higher rates of treatment initiation, a logistic regression analysis was conducted with treatment initiation as the dependent variable and the s cores for each of the four BASC composite scales as the predictors. Prior to conducting this analysis, an i ndependent samples t test and several Chi square tests were conducted to test whether groups differed on a number of demographic variables, including age, gender, ethnicity family income, and marital status Results indicated that treatment initiators and treatment refusers did not differ significantly on these variables, and thus these demographic factors were not controlled for in subsequent anal yses. Table 34 presents the results of these analyses Contrary to expectations, the overall regression model was not significant, indicating that parent ratings of child symptom severity at the time of the intake assessment did not predict whether or not families entered treatment [ 2(4) = 2.18, p = .70, Nagelkerke Pseudo R2 = .03]. Odds ratios for the individual composite scales were all near one, indicating that higher parent ratings were not associated with hi gher odds of entering treatment. See Table 35 for individual bweights, p values, and odds ratios for each composite scale in this analysis. Parent Teacher C oncordance Next, the hypothesis that parent and teacher concordance on ratings of ch ild symptom severity would predict treatment initiation was tested using a second logistic regression analysis. Prior to running the analysis, difference scores were obtained by subtracting parent BASC scores from teacher BASC scores, creating a new variable of difference scores ranging from 50 to 27 for Externalizing Problems, 56 to 26 for

PAGE 28

28 Internalizing Problems, 49 to 18 for the Behavioral Symptoms Index, and 55 to 29 for Adaptive Skills. Negative values indicated higher parent ratings, whereas posi tive difference scores were indicative of higher teacher ratings. Values closer to zero were indicative of less discrepancy among informants. These difference scores were then entered as simultaneous predictors into a logistic regression analysis, with t reatment initiation status as the outcome variable. The results of this analysis were not significant, indicating that the overall difference between parent and teacher ratings was not predictive of treatment initiation [ 2(4) = .55, p > .05, Nagelkerke P seudo R2 = .01]. Table 36 displays the individual bweights, p values, and odds ratios for this analysis. A separate logistic regression was conducted using the absolute value of the difference score between parent and teacher ratings. For this new var iable, a value of zero, indicating complete agreement, was the minimum obtainable score and higher scores indicated greater overall discrepancy between informants. Similar results were obtained in this analysis, with absolute value difference scores faili ng to predict treatment initiation [ 2(4) = 2.09, p > .05, Nagelkerke Pseudo R2 = .04]. Table 3 7 display s the individual bweights, p values, and odds ratios for th is analys i s. The next analysis was designed to test whether children whose parents and teachers both rated them in the at risk or clinical range (T scores > 60) were more likely to initiate treatment than children for whom this was not the case. Using the cutoff of 60, four groups were created and entered into two separate C hi square analyses, one for Externalizing Problems and another for Internalizing Problems. For these analyses, children whose parents and teachers completed the Achenbach scales were included along with the children for whom BASC scales were completed. Given that data were

PAGE 29

29 combined across measures, the choice was made to use only the Externalizing Symptoms and Internalizing Symptoms scales, as these are the scales with greatest overlap across the t wo different measures. Table 3 8 juxtaposes the descriptive information for corresponding forms of the BASC and Achenbach scales. Overall, means for corresponding forms across the different measures (BASC and Achenbach) were all within four T score point s of one another. The maximum score was consistently higher for the BASC scales than the Achenbach scales for both the Externalizing Problems and Internalizing Problems scales. This, however, may be largely due to the fact that the Achenbach scales have a more limited range of T scores (max = 100) than do the BASC scales (max = 120). The four groups created for this analysis were: (1) children whose parents and teachers both rated them in the at risk or clinical range; (2) children for whom neither parent nor teacher ratings were in the clinical or at risk range; (3) children whose teachers but not parents rated them as at risk or clinical; and (4) children whose parents but not teachers rated them in the at risk or clinical range. Results showed that the likelihood of entering treatment did not differ based on the agreement category to which the child belonged. This was true both for the Externalizing Problems analysis ( 2 = 4.55; p = .21) and the Internalizing Problems analysis ( 2 = 2.97; p = .40 ). Contrary to expectations, children were just as likely to enter treatment when both parent and teacher agreed that the child exhibited elevated symptoms as when neither parent nor teacher reported these elevated scores.

PAGE 30

30 Table 31. BASC descriptive s Initiators vs. r efusers Table 32 Paired samples T tests Parents vs. t eachers Initiators Refusers t d f p t df p BASC Externalizing 4.47 47 <.001 3.30 20 <.01 BASC Internalizing 1.65 49 .11 1.73 21 .10 BASC Behavioral Sx Index 4.66 49 <.001 3.62 21 <.01 BASC Adaptive Skills 0.28 51 .78 0.48 21 .64 Table 33 Fishers Z test results for BASC parent teacher correlations Study Sample Normative sample Pearson r ( n ) Pearson r ( n ) Fishers Z p Child form Externalizing .39 (34) .51 (470) 0 .08 .42 Internalizing .21 (34) .23 (470) 0 .11 .91 Behavioral Sx Index .17 (34) .45 (470) 1.69 .09 Adaptive Skills .01 (34) .48 (470) 2.77** <.01 Adolescent form Externalizing Internalizing Behavioral Sx Index Adaptive Skills .60 (31) .63 (34) .55 (34) .20 (36) .51 (470) .32 (470) .40 (470) .48 (470) 0 .67 2.21 1.05* 1.78 .50 <.05 .29 .08 p < .05 ** p < .01 Initiators n = 48 Refusers n = 20 Mean (SD) Mean (SD) BASC PRS Externalizing 71.31 (2.44) 68.70 (3.95) BASC TRS Externalizing 60.27 (1.91) 57.35 (3.14) BASC PRS Internalizing 60.56 (1.92) 61.75 (3.25) BASC TRS Internalizing 56.81 (2.20) 54.95 (2.69) BASC PRS Behavioral Sx Index 70.19 (1.96) 69.60 (3.24) BASC TRS Behavioral Sx Index 60.29 (1.74) 58.10 (2.58) BASC PRS Adaptive Skills 37.60 (1.77) 39.50 (1.63) BASC TRS Adaptive Skills 37.52 (1.02) 38.60 (1.44)

PAGE 31

31 Table 34 Statistical t ests for d emographic f actors t 2 df p Age .39 142 .70 Gender 0 .01 1 .94 Ethnicity 4.44 5 .49 Income Marital Status 2.42 3.15 7 3 .93 .37 Table 35 Logistic r egression Parent rated symptom severity B df p Odds Ratio BASC Externalizing .008 1 .78 1.01 BASC Internalizing .004 1 .89 1.00 BASC Behavioral Symptoms Index .034 1 .45 .97 BASC Adaptive Skills .007 1 .78 .99 Notes: Parent BASC scores as predictors of treatment initiation. Table 36 Logistic regression Parent teacher difference scores B df p Odds Ratio BASC Externalizing .01 1 .86 1.00 BASC Internalizing .01 1 .62 1.01 BASC Behavioral Symptoms Index .00 1 .98 1.00 BASC Adaptive Skills .00 1 .95 1.00 Notes: Parent teacher difference scores as predictors of treatment initiation Table 37 Logistic regression Parent teacher difference scores (absolute value) B df p Odds Ratio BASC Externalizing .04 1 .18 1.04 BASC Internalizing .02 1 .62 1.02 BASC Behavioral Symptoms Index .04 1 .27 .96 BASC Adaptive Skills .00 1 .94 1.00 Notes: Parent teacher difference scores (absolute value) as predictors of treatment initiation

PAGE 32

32 Table 38 BASC vs. Achenbach scale m eans n Mean ( SD ) Min Max BASC PRS Externalizing 97 67.31 ( 17.50 ) 36 118 CBCL Externalizing 69 64.51 ( 10.82 ) 32 82 BASC TRS Externalizing 71 59.49 ( 13.31 ) 40 90 TRF Externalizing 74 60.76 ( 9.17 ) 41 79 BASC PRS Internalizing 96 59.81 ( 13.53 ) 38 98 CBCL Internalizing 69 63.12 ( 10.84 ) 33 88 BASC TRS Internalizing 75 56.67 ( 14.15 ) 39 102 TRF Internalizing 74 58.09 ( 10.45 ) 36 77

PAGE 33

33 CHAPTER 4 DISCUSSION This study was designed to explore the role of parent perceived symptom severity and parent teacher concordance i n treatment initiation in a school referred sample of children and adolescents from a rural county in Northcentral Florida. Of the 144 families who were offered services through UF FSS, 107 entered treatment. This 74% initiation rate is substantially higher than the 50% or lower initiation rates often reported in the t reatment adherence literature ( Geffken et al. 2001) Demographic variables were not significant predictors of treatment initiation. Although previous research has shown some of these demographic factors, including annual income and minority status, to sig nificantly predict treatment initiation ( Armbruster & Fallon, 1994; Wierzbicki & Pekarik, 1993), this relationship was not found in the present study. However, this is likely due to the restricted range of income in the study sample (most families earned les s than $ 50,000 per year) and the small number of minority participants (86.8% were Caucasian) in the present sample. This likely limited the ability to detect income and ethnicity effects in the logistic regression analyses. Ai m 1 The first aim of this study was to compare the degree to which parent and teacher concordance on ratings of child symptom severity in the present sample differed from the correlations obtained from the normative sample as reported in the BASC manual ( R eynolds & Kamphaus, 1998). Given that these children and adolescents were referred for a psychological evaluation by their teachers rather than their parents, as is typically the case, the expectation was that the correlations would be significantly weake r than those of the normative sample. The near zero positive correlation in the

PAGE 34

34 study sample for the Adaptive Skills composite (child form) was significantly weaker than the reported correlation in the BASC manual. Conversely, on the adolescent form, the re was a significant difference between the study sample and normative sample in parent teacher correlations for Internalizing Problems. Parents and teachers in the study sample were significantly more in agreement regarding adolescents internalizing dif ficulties than were parents and teachers in the normative sample. Differences between the study sample and normative sample did not reach significance for the Externalizing Problems, Internalizing Problems, or Behavioral Symptoms Index composites on the c hild form, or the Externalizing Problems, Behavioral Symptoms Index, and Adaptive Skills composites on the adolescent form. Overall, the majority of the correlations in this analysis were comparable to those reported in the BASC manual. The few instances where there appeared to be significant differences among the two samples are likely attributable to the relatively small number of cases for the study sample as compared to the BASC normative group. Aims 2 through 4 The second goal of this study was to determine whether parent perceptions of symptom severity pr edicted treatment initiation. In line with the findings of MacNaughton & Rodrigue (2001), r esults showed that the degree of child symptom severity, as perceived by parents, did not significantly predict whether or not families followed through with the recommendation to enter psychological treatment. The third and fourth aims of the study extended this second hypothesis by looking at whether corroboration from teachers regarding childrens symptom severity predicted treatment initiation. The assumption underlying this hypothesis was that parents would be more likely to initiate treatment for their children if they and the childrens teachers were both

PAGE 35

35 reporting higher levels of symptom severity, especially given that teachers are often important participants in the treatment of childhood psychopathology. Contrary to expectations, higher levels of parent and teacher concordance were not associated with greater likelihood of initiating treatment. A lso, c hildren whose parents and teachers both rated their symptoms in the at risk or clinical range were not more likely to enter treatment than children whose parents and teachers rated them in the nonclinical range. Implications The relatively hig h rates of treatment initiation in the present study may be explained by several factors. One possible reason for this high rate of treatment initiation is the reduction of several obstacles that have been found in previous studies to serve as barriers to treatment. Specifically, families in this study were offered treatment free of charge and services were available either in a nearby school board building or in the childs school, reducing the barriers of cost and transportation difficulties. Minimizat ion of these factors, which often preclude families from entering treatment, likely made treatment more accessible for these families, many of whom are low income and have limited resources. Although no formal exploration of this was conducted, the high r ate of treatment initiation in this study, a study in which the barriers of cost and accessibility we re greatly reduced, provides some indirect support for earlier findings showing that the number of barriers reported by families predicted treatment initia tion regardless of the specific type of barriers they were ( MacNaughton & Rodrigue, 2001). The relatively high rates of treatment initiation in the present sample highlights the importance of offering affordable and accessible mental health services to ch ildren from low SES families and underserved communities, for whom barriers to treatment

PAGE 36

36 often interfere with access to mental health services. By minimizing these barriers it is likely that a greater proportion of families may be able to attend treatmen t initially. Another possible reason for the high rates of treatment initiation in this study is that all families in this sample did at the very least, attend an intake session, thus contact had already been initiated with mental health care providers when families were faced with the decision of whether or not to initiate treatment. Many families may have already considered this intake session the start of treatment and attendance at a first therapy session may have been a natural progression of this. While, from a research perspective, treatment was not considered to have been initiated until the first official therapy session was attended, these families may have instead considered the intake session to be the actual start of treatment. Also, having already undergone an extensive interview and assessment process during the intake, families w ho were offered treatment may have felt they had much already invested in the UF FSS service thus increasing their likelihood of entering treatment with this provider Moreover, these families already showed some degree of interest and willingness to obtain psychological services just by attending the initial intake assessment Because data were unavailable regarding the percentage of families who agreed to the schools initial referral, it is unknown whether the rates of initial referral acceptance as opposed to treatment acceptance more closely resemble the less than 50% rates typically found for initiation of psychotherapy treatment. It is possible that t he families who attended the intake session are a biased sample of families more likely to exhibit interest in psychological treatment. Another important factor to consider is that families were referred for the intake assessment by the childs school, perhaps leading families to feel more

PAGE 37

37 compelled to follow through with the intake and ultimately with treatment given the school district s involvement in the process An additional important characteristic of the present study that should be considered is that only treatment initiation was assessed. Although more than 70% of families initiated treatment, the present study did not look at whether this high rate of initiation corresponded with high rates of persistence in treatment or treatment completion It is possible that many families initiated treatment because the referrals were made by the childrens schools and families felt they had to comply with the schools request. However, they may not have been truly committed to treatment participation a nd thus may have terminated treatment early. Nonetheless, in light of previous findings showing that some families can benefit from just a few therapy sessions, the high rate of initiation in the present study is important even if many of these families m ay have terminated treatment prematurely. The lack of a significant relationship between parent rated symptom severity as well as parent teacher concordance and treatment initiation in this study indicates that there are other factors contributing to parents decisions of whether or not to initiate psychological treatment for their children. One likely possibility is that other barriers to treatment that were not captured in this study are more important predictors of treatment initiation in this sam ple. For families whose childrens symptoms were rated more severely, the perceived relevance of treatment may not have corresponded with these high rates of perceived symptom severity Other forms of treatment, such as prescription medication, may have seemed more pertinent for these families. Another possibility is that factors associated with the initial intake session may have informed

PAGE 38

38 families decisions. A positive experience during the intake session, for instance, may have motivated families to enter treatment whereas a more negative intake experience, such as families disapproving of the assessor s interpretation of their childrens difficulties, may have led families to opt out of treatment with UF FSS. Overall, based on the findings of this study, it appears that other factors may be greater predictors of treatment initiation in this population. Limitations Several important limitations of the current study warrant consideration when interpreting these results. First, although there were 144 families in the overall study sample, parent or teacher data were missing for several children, substantially reducing the sample size available for several of the analyses. Of particular relevance, expected cell sizes for the C hi squar e analyses used to test the final hypothesis were under the minimum of five for several of the cells. Although this suggests the need to interpret the C hi square results with some caution, the markedly high pvalues for these analyses suggest that these n ull findings would hold true even with a larger sample size. For the final hypothesis, sample size was increased by combining children whose parents and teachers filled out BASC forms with children whose parents and teachers instead completed the Achenbach scales. While these separate measures have many similarities in terms of specific item content and scale composition, they are distinct measures that differ somewhat in length, wording, and rating scale format (the BASC forms have a four point rating scale while the Achenbach forms have a threepoint rating scale). Also, the BASC allows for T scores up to 120 while T scores on the Achenbach scales are capped at 100. These differences may affect the reliability of results obtained by combining the meas ures. However, by dichotomizing these

PAGE 39

39 measures using a cutoff value rather than using the actual scores on these measures as continuous variable s, combining these measures is more easily justified and reliable as a method of analysis. Nevertheless the use of a dichotomy does not allow the full range of variability on these measures to be utilized in the analyses thus limiting its ability to detect smaller differences A third possible limitation is that only broadband measures were used to gauge parents and teachers perceptions of symptom severity (BASC and Achenbach scales). Although these measures capture a broad range of potential problems, they may not be focused enough to capture gradations in symptom severity for specific behavioral and p sychological problems. Nonetheless, exploratory analyses using measures more specifically targeted toward externalizing problems (the Eyberg Child Behavior Inventory and the Sutter Eyberg Student Behavior Inventory) yielded very similar findings, suggesti ng that the findings of the current study were not merely an artifact of the measures that were used. While this study attempt ed to predict treatment initiation based on parent and teacher ratings of symptom severity, families were not directly inquired regarding their reasons for not initiating treatment. Some of these families may have elected to seek treatment elsewhere, and thus would not be accurately classified as treatment refusers. Alternatively, some families may have moved to a different locat ion, thus precluding them from being able to attend treatment with UF FSS. Although less likely, there is also the possibility that the problems for which treatment was recommended may have resolved on their own for some of these children prior to treatment initiation as some families waited several months to enter treatment A number of unknown factors may

PAGE 40

40 have contributed to families decisions not to initiate treatment with UF FSS and these factors, because they were not captured in this study, may be masking the relationship between parent rated symptom severity or parent teacher concordance and treatment initiation. Moreover, although the services offered through UF FSS do minimize two important barriers to treatment (cost and accessibility of locat ion), families we re not systematically asked about other barriers that may actually be important factors contributing to treatment refusal. The present study looked at the relationship between parent perceived symptom severity as well as parent teacher concordance and treatment initiation for the sample as a whole. It is possible that a relationship may have emerged between these factors if these analyses had been conducted separately for different age groups or for males and females. However, due to t he relatively small sample size, these analyses could not be reliably run separately, as the even smaller sample sizes would not have permitted the use of logistic regression analyses, which requires a certain number of subjects per predictor. When interpr eting the results of the current study, it is also important to consider the unique characteristics of the sample as well as of the services that were offered to families. All families in this sample were from a rural county in the South and were referred for services by the childs school While perceived symptom severity and parent teacher concordance did not predict treatment initiation in this sample, these findings may not be generalizable to other samples of families of different socioeconomic stat us or from different regions While unmeasured barriers to treatment or other factors may have been more relevant for this low income, rural community,

PAGE 41

41 parent perceived symptom severity and parent teacher concordance may be more relevant predictors in com munities and populations with higher incomes who perceive fewer barriers to successful treatment attendance. Although income was not found to significantly differ in the current sample between initiators and refusers, the overall income status of most fam ilies was fairly low, perhaps masking a true effect of income status. Future Directions Overall, the null findings of the present study indicate that much is yet to be discovered regarding factors that contribute to treatment initiation. Although the present study did not find a relationship between parent rated symptom severity, or parent teacher concordance, and treatment initiation, these findings can help to inform future research. These findings showed that even in the absence of major financial a nd locationrelated barriers to treatment, a quarter of families failed to follow through with the recommendation to initiate treatment. For these families, other factors were likely involved in their decision not to enter treatment with UF FSS. Future r esearch in this underserved, low income community of children and adolescents should aim to explore what specific factors interfere with families willingness or ability to initiate psychological intervention services following psychological evaluation. B y asking families to identify the factors contributing to their decision to enter or refuse treatment, interventions can be implemented at the intake phase that are designed to improve the likelihood that families will initiate treatment. Also, services c ould be made available to help families further minimize the number of perceived barriers that may interfere with treatment initiation. One possible barrier to treatment that can be specifically asked about in future studies is parenting stress. Although some families who participated in Project

PAGE 42

42 CATCh did complete a parenting stress measure, the number of families who did so was relatively small precluding investigation into the role of parenting stress on treatment initiation in the present sample. However, given that parenting stress may be an important barrier to treatment, this factor warrants investigation in the future. The adult literature on treatment adherence behaviors has employed the transtheoretical stages of change model to explore the relationship between patients readiness for change and treatment attendance. Derisley and Reynolds (2000), for instance, found that patients who scored low on Contemplation were more likely to terminate treatment prematurely. This same model can be applied to research on child psychotherapy attendance. When exploring the role of stages of change in child psychotherapy attendance, it is reasonable to assume that both parent and child readiness for change would be important cont ributors of attendance. The role of readiness for change in adherence behaviors may differ for parents and children at different stages of treatment (including initiation, attendance, and termination). There may also be differential importance of parent and child stages of change for treatment attendance, and this relationship may be different for youths of different ages. Exploring both parent and child stages of change at different points of the psychotherapy process may provide important insight into adherence behaviors in child psychotherapy. In the present study, parent teacher concordance measured the degree to which parents and teachers agreed on the severity of childrens symptoms. However, parent teacher concordance on severity ratings may not t ranslate into concordance on the need for psychological treatment. Our understanding of factors that contribute to treatment initiation will likely be improved if concordance in future studies is instead

PAGE 43

43 operationalized as the degree to which parents and teachers agree, not only on the severity of psychological symptoms but on the extent to which they believe the child needs psychological intervention. By using a measure that taps into actual beliefs about the need for and relevance of treatment, it may be possible to more accurately determine the role that parent teacher concordance plays in treatment initiation.

PAGE 44

44 LIST OF 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. Psychological Bulletin, 101, 213232. Armbruster, P., & Fallon, T. (1994). Clinical, sociodemographic, and systems risk factors for attrition in a childrens mental health clinic. American Journal of Orthopsychiatry, 64, 577 585. Armbruster, P., & Kazdin, A.E. (1994). Attrition in child psychotherapy. Advances in Clinical Child Psychology, 16, 18 108. Bennett, D.S., Power, T.J., Rostain, A.L., & Carr, D.E. (1996). Parent acceptability and feasibility of ADHD interventions: Assessment, correlates, and predictive validity. Journal of Pediatric Psychology, 21, 643 657. Brandner F.A. (1933). A test of the significance of the difference of the correlation coefficients in normal bivariate samples. Biometrika, 25(1/2), 102109. BrookmanFrazee, L., Gabayan, E.N., Haine, R.A., & Garland, A.F. (2008). Predicting frequency of treatment visits in community based youth psychotherapy. Psychological Services, 5 (2), 126 138. Costello, E.J., Pescosolido, B.A., Angold, A., & Burns, B.J. (1998). A family network based model of access to child mental health services. Research in Community Mental Health, 9, 165190. Derisley J., & Reynolds S. (2000) The transtheoretical stages of change as a predictor of premature termination, attendance and alliance in psychotherapy. British Journal of Clinical Psychology, 39, 371382. Duhig A.M., Renk, K., Epstein, M.K., & Phares, V. (2000). Interpersonal agreement on internalizing, externalizing, and total behavior problems: A metaanalysis. Clinical Psychology: Science and Practice, 7, 435453. Geffken, G.R., Keeley, M.L., Kellison, I., Rodrigue, J.R., & Storch, E.A. (2006). Parental adherence to child psychologists recommendations from psychological testing. Professional Psychology: Research and Practice, 37(5), 499 505. Griest, D.L., Forehand, R., Wells, K.C., & McMahon, R.J. (1980) An examination of differences between nonclinic and behavior problem clinic referred children and their mothers. Journal of Abnormal Psychology, 89, 497 500.

PAGE 45

45 Gustafson, K.E., McNamara, J.R., & Jensen, J. (1994). Parents informed consent deci sions regarding psychotherapy for their children: Consideration of therapeutic risks and benefits. Professional Psychology: Research and Practice, 25, 1622. Hricik, D.A., & Phillips Keane, S. (1988). Referred and nonreferred childrens predictions of their mothers behavior evaluations. Journal of Clinical Child Psychology, 17, 8 13. Kazdin, A.E., Holland, L., & Crowley, M. (1997). Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting and Cli nical Psychology, 65, 453 463. Kazdin, A.E., Holland, L., Crowley, M., & Breton, S. (1997). Barriers to treatment participation scale: Evaluation and validation in the context of outpatient therapy. Journal of Child Psychology and Psychiatry, 38, 10511 062. Kazdin, A.E., Mazurick, J.L., & Bass, D. (1993). Risk for attrition in treatment of antisocial children and families. Journal of Clinical Child Psychology, 22, 2 16. Kazdin, A.E., Mazurick, J.L., & Siegel, D. (1994). Treatment outcome among children with externalizing disorders who terminate prematurely versus those who complete psychotherapy. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 549 557. Kazdin, A.E., & Wassell, G. (1999). Barriers to treatment participation and therapeutic change among children referred for conduct disorder. Journal of Clinical Child Psychology, 28, 160172. Keeley, M.L., & Wiens, B.A. (2008). Family influences on treatment refusal in school linked mental health services. Journal of Child and Family Studies, 17, 109126. Kendall, P.C., Sugarman, A. (1997). Attrition in the treatment of childhood anxiety disorders. Journal of Consulting and Clinical Psychology, 65(5), 883 888. MacNaughton K.L., & Rodrigue, J.R. (2001). Predicting adherence to recommendations by parents of clinic referred children. Journal of Consulting and Clinical Psychology, 69(2), 262 270. Nock, M., & Ferriter C. (2005). Parent management of attendance and adherence in child and adolescent therapy: A conceptual and empirical review. Clinical Child and Family Psychology Review, 8(2), 149166. Nock, M., Ferriter, C., & Holmberg, E. (2007). Parent beliefs about treatment credibility and effectiveness: Assessment and relation to subsequent treatment participation. Journal of Child and Family Studies, 16, 27 38.

PAGE 46

46 Nock, M.K., & Kazdin, A.E. (2001). Parent expectancies for child therapy: Assessment and relation to participation in treatment. Journal of Child and Family Studies, 10, 155 180. Nock, M.K., & Photos, V. (2006). Parent motivation to participate in treatment: Assessment and prediction of subsequent participation. Journal of Child and Family S tudies, 15(3), 345358. Piacentini, J., Rotheram Borus, M.J., Graae, F., Gillis, J.R., Trautman, P., Cantwell, C., Garcia Leeds, C., & Shaffer, D. (1995). Demographic predictors of treatment attendance among adolescent suicide attempters. Journal of Consulting and Clinical Psychology, 63 (3), 469 473. Reynolds, C.R., & Kamphaus, R.W. (1998). Behavior Assessment System for Children Manual. Minnesota: American Guidance Service, Inc. Smith, S.R. (2007). Making sense of multiple informants in child and adolescent psychotherapy. Journal of psychoeducational assessment, 25(2), 139 149. Thibodeau, A.S. (2007). Parental compliance to clinical recommendations in an ADHD clinic. Dissertation Abstracts International: Section B. Sciences and Engineering, 68(2 B), 1322. Vianna, M.A.G. (1980). Statistical methods for summarizing independent correlational results. Journal of Educational Statistics, 5(10), 83 104. Wierzbicki, M., & Pekarik, G. (1993). A meta analysis of psychotherapy dropout. Pr ofessional Psychology: Research and Practice, 24(2), 190 195.

PAGE 47

47 BIOGRAPHICAL SKETCH Eileen Matias Davis is originally from Hialeah, Florida where she graduated as valedictorian in 2002 from Hialeah High School. In 2006 she earned a Bachelor o f Arts degree in psychology from Harvard University in Cambridge, Massachusetts. During her time as an undergraduate, Eileen worked at Harvards Laboratory for Developmental Studies from 20032006 and Clinical Research Laboratory in 2006. Upon graduating, Eileen took a position as a first grade teacher at Meadowlane Elementary School, a Title I primary school in South Florida. In 2008, s he entered the Clinical and Health Psychology doctoral program at the University of Florida with a concentration in cl inical child psychology. As a graduate student at UF, she has worked under the mentorship of Brenda A. Wiens, Ph.D ., in the National Rural Behavioral Health Center, providing clinical services to children and adolescents from the Columbia County School Bo ard system.