PREDICTING TREATMENT AND FOLLOW-UP ATTRITION IN PARENT-CHILD INTERACTION THERAPY (PCIT) By MELANIE ANN FERNANDEZ A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007
Copyright 2007 by Melanie Ann Fernandez
ACKNOWLEDGMENTS I would like to express my gratitude to my mentor, Dr Sheila Eyberg, for her encouragement and support on this project and throughout my graduate school career. I would also like to thank my supervisor y committee members (Drs. Michael Robinson, Shelley Heaton, and Mary Fukuyama) whose thoughts and recommendations made this achievement possible. I am also incredibly grateful for support from my loving family, from Steven Reader, and from the me mbers of the Child Study Laboratory. iii
TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iii LIST OF TABLES ...............................................................................................................v ABSTRACT .......................................................................................................................vi CHAPTER 1 INTRODUCTION........................................................................................................1 2 METHOD...................................................................................................................10 Participants .................................................................................................................10 Measures .....................................................................................................................12 Demographic Measures .......................................................................................12 Diagnostic M easures ............................................................................................13 Cognitive Screening Measures ............................................................................13 Maternal Distress Variable ..................................................................................14 Observational Measures ......................................................................................14 Consumer Satisfaction .........................................................................................15 Procedure ....................................................................................................................16 3 RESULTS...................................................................................................................18 4 DISCUSSION.............................................................................................................25 LIST OF REFERENCES ...................................................................................................30 BIOGRAPHICAL SKETCH .............................................................................................37 iv
LIST OF TABLES Table page 3-1. Pearson Correlation Matrix for Potential Treatment Attrition Predictor Variables .19 3-2. Means and Standard Deviations for Potential Treatment Attrition Predictor Variables ...................................................................................................................20 3-3. Primary Reasons for Dropout Provided by Families for Dropping Out of Parent Child Interaction Therapy (PCIT)............................................................................21 3-4. Means and Standard Deviations for Potential Follow-Up Attrition Predictor Variables in the Assessment-Only Condition ..........................................................22 3-5. Pearson Correlation Matrix for Potential Follow-Up Attrition Predictor Variables in the Assessment-Only Condition ..........................................................22 3-6. Means and Standard Deviations for Potential Follow-Up Attrition Predictor Variables in the Maintenance Treatment Condition ................................................23 3-7. Pearson Correlation Matrix for Potential Follow-Up Attrition Predictor Variables in the Maintenance Treatment Condition ................................................23 v
Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PREDICTING TREATMENT AND FOLLOW-UP ATTRITION IN PARENT-CHILD INTERACTION THERAPY (PCIT) By Melanie Ann Fernandez August 2007 Chair: Sheila M. Eyberg Major: Psychology Our study examined predictors of treatmen t and follow-up attrition in a sample of 99 children between 3 and 6 year s of age with disruptive be havior disorders and their families. Our study added additional parent-support components to the PCIT protocol to alleviate maternal distress and logistical barrie rs identified in earlier studies as possible contributors to attrition. Base d on previous findings, completer and dropout families were compared on maternal Negative Talk and Total Praise, Maternal Distress, and socioeconomic status (SES), and these variables were examined as potential predictors of attrition. Results showed significant group differences in Nega tive Talk, Total Praise, and SES but not Maternal Distress; with lower SES, more Negative Talk, and less Total Praise predicting 71% of treatment dropouts. After PCIT, families were randomly assigned to either an assessment-only or maintenance treatment condition. Prior studi es suggested differences in follow-up completers and dropouts in Maternal Dist ress, Maternal Intellectual Functioning, vi
Maternal Age, and Treatment Satisfaction. Ther efore, these variables were examined as predictors of follow-up attrition. Results showed significant group differences in Maternal Distress for the assessment-onl y condition, and in Ma ternal Intellectual Functioning for the maintenance treatment c ondition. Higher Maternal Distress predicted 63% of dropouts in the assessment-only condition. Lower Mate rnal Intellectual Functioning and younger Maternal Age pred icted 83% of maintenance treatment dropouts. Findings suggested that a ddressing parent stressors beyond the parent-child intervention may help retain distressed mothers in treatment. Results point to the need for pretreatment assessment of barriers to partic ipation in PCIT, particularly related to socioeconomic disadvantage. For follow-up, con tinued contact with therapists may help retain distressed mothers. In conjuncti on with telephone contact, in-person booster sessions may be indicated for mothers of lower Intellectual Functioning. vii
CHAPTER 1 INTRODUCTION Disruptive behavior disorders (DBDs) ar e the most common reason for referral of children to mental health clinics (Loeber, Burke, Lahey, Winters, & Zera, 2000). These conditions include oppositional defiant disord er (ODD) and conduct disorder (CD) and are characterized by clinically si gnificant levels of disregard for authority and violation of social norms and others right s, respectively (American Psychiatric Association [APA], 2000). Disruptive behavior disord ers affect as many at 16% of children and are associated with significant impairments in social, emo tional, and educationa l functioning (Frick & Loney, 1999). Left untreated, disruptive behavior in young children shows a high degree of stability over time (Campbell, 2002; Egeland, Kalkoske, Gottesman, & Erickson, 1990; Farrington, 1995; Hodgins, 1994; La hey et al., 1995) and carrie s a high societal price tag. Children with these disorders account for a larg er percentage of hea lth care costs than children with chronic health c onditions, such as asthma, diab etes, and epilepsy (Guevara, Mandell, Rostain, Zhao, & Hadley, 2003). In addition, DBDs are the most powerful risk factor for subsequent delinque nt behavior, including interp ersonal violence, substance abuse, and property destruction (Fagot, Loeber, & Reid, 1988). In the year 2000, juveniles committed 170,800 violent crimes and over two million property crimes, resulting in an estimated cost of over $3.5 bill ion (Federal Bureau of Investigation [FBI], 1997). Other costs of DBDs are associated with treatment of juvenile delinquency through intervention and educational program s and law enforcement (Frick & Loney, 1
2 1999; Werry, 1997). From many perspectives, res earch supports the need to intervene early in the development of DBDs. Treatment provided at preschool age may be more effective than when children are older (Dishion & Patterson, 1992; Ruma et al ., 1996). Furthermore, effective treatment before school entry has been associated w ith long-term maintenance of child behavior gains (Boggs et al., 2004; Hood & Eyberg, 2003; Reid, Webster-Stratton, & Hammond, 2003). By contrast, treatment attrition is associ ated with the absence of treatment gains at long-term follow-up for young children with DBDs (Boggs et al.). Child treatment attrition is problematic for research also. Attrition interferes with the random composition of research groups, redu ces statistical power, and limits external validity of findings (Kazdin, 1990). Further, attr ition research in child therapy has been characterized as in definitional chaos (A rmbruster & Kazdin, 1994, p. 90). Attrition is broadly defined as premature termination decided unilaterally by a patient against therapist advice (Armbruster & Kazdin), but the differing operational definitions make it difficult to establish attrition rates across stud ies. Some studies define those who agree to treatment but never attend even the first session as dropouts (Weisz, Weiss, & Langmeyer, 1987; McCabe, 2002). Ot her studies define dropout s as those who begin but attend fewer than 25% of the sessions (Kazdi n, 1990), and still others define those who stop treatment before meeting its goals as dropouts (Werba, Eyberg, Boggs, & Algina, in press). In studies of attrition, re searchers have also cautioned against the uniformity assumptionthe notion that predictors of dropout are similar, or uniform, across different variables such as child age, clini cal diagnosis, treatment, or treatment modality
3 (Armbruster & Kazdin, 1994, p. 100). However, several studies refute the uniformity assumption. For example, in treatments fo r externalizing problems, higher problem severity has predicted attr ition (Kazdin, 1990); whereas fo r internalizing behavior problems, less severe symptoms have been associated with attrition (Kendall & Sugarman, 1997; Flannery-Schroeder, C houdhury, & Kendall, 2005). Findings such as these suggest that attrition ma y be treatment-specific, varyi ng with characteristics of the population for whom the treatment is designed (Armbruster & Kazdin). Attrition rates for parent behavior manage ment training programs range from 40 to 65% (Pekarik & Stephenson, 1988; Wierzbic ki & Pekarik, 1993; Kazdin, 1996), with families continuing to drop out during follow-up (e.g., 22% in Webster-Stratton, 1990; 40% in Cohen, Mannarino, & Knudsen, 2005; 22% in Flannery-Schroeder et al., 2005). Given the methodological and clinical implicat ions of attrition; and given findings that half of children and families drop out of ther apy, and that additional families are lost to follow-up, continued study of attrition is essential. Although many predictors are treatment-speci fic, certain demographic and maternal factors have been repeatedly associated with dropout from treatments for DBDs. Specifically, low socioeconomic status (SES ), single motherhood, ethnic minority status, and high maternal stress have been identifie d across treatments (Dumas & Wahler, 1983; Kazdin, 1990; Kazdin, Mazurick, & Bass, 1993; Armbruster & Fallon, 1994; Kazdin & Mazurick, 1994; Prinz & Miller, 1994; Ca page, Bennett, & McNeil, 2001; McCabe, 2002; Peters, Calam, & Harrington, 2005). Although attrition from child treatments has received considerable research attention, less is known about attrition from follow-up. Long-term follow-up studies of
4 child therapy have focused primarily on either diagnostic ra tes (Cohen, Mannarino, & Knudsen, 2005; Flannery-Schroeder et al., 2005) or statistically and c linically significant change from pretreatment (E yberg et al., 2001), rather than on predictors of attrition. Follow-up studies provide important information on the durability of treatment gains over time, and on the relationship between child and adult psychopathology (Vander Stoep, 1999). Attrition from follow-up, therefore, prev ents acquisition of clinically important information and carries with it methodologica l problems similar to those for treatment (such as reduction of statistical power and li mited generalizability of research findings). The DBD treatment research has thus far been mixed regarding differences between follow-up completers and dropouts. For example, Webster-Stratton (1990) found no differences between follow-up completers and dropouts at preor posttreatment on a variety of variables. In c ontrast, in a 1to 3-year follow-up study of preschoolers, participating mothers were si gnificantly older and had highe r cognitive screening scores than mothers who were lost to follow-up (Hood & Eyberg, 2003). Consumer satisfaction after treatment was also lower for follow-up dropouts than for those who completed treatment (Brestan, Jacobs, Rayfield, & Eyberg, 1999; Luk, Staiger, Mathai, Wong, Birleson, & Adler, 2001). Our study examined attrition from Parent-C hild Interaction Therapy (PCIT), an evidence-based treatment for young childre n with DBDs (Brestan & Eyberg, 1998; Gallagher, 2003; Eyberg, Nelson, & Boggs, in press). Outcome studies show increases in positive parenting behaviors such as praise and reflective listening, and decreases in negative parenting behaviors such as criticism and sarcasm, and the children have shown increases in compliance and decreases in devi ant behaviors (Eisenstadt, Eyberg, McNeil,
5 Newcomb, & Funderburk, 1993; Eyberg et al., 2001). Parents who complete treatment have also reported improvements in parenti ng stress and parenting locus of control and high satisfaction with treatment (Nixon, Sweeney, Erickson, & Touyz, 2003; Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998) alt hough findings suggest that improvements in parenting stress may not be maintained at long-term follow-up (Eyberg et al., 2001). PCIT is not time-limited and continues until parents reach pre-set skill levels and rate their childs behavior within normal limits. Treatment attrition is defined as discontinuing treatment at any point after at tending the first treatment session and before meeting the treatment-completion criteria. Families that complete PCIT are considered treatment successes, and families that drop out before meeting treatment completion criteria are considered treatment failures. In an earlier study (W erba et al., in press), the treatment attrition rate was 33%. In the Werba et al. (in press) study, potent ial pretreatment predictors of attrition were explored with multiple logistic regression. The investigators compared dropouts ( n =17) to completers (n = 33) on a large number of variables, including demographic, maternal, and child characteris tics, logistical barriers (e .g., distance to clinic), and observational variables from mother-child inte ractions (e.g., caregiver prosocial behavior toward child). Variables differing between groups at p < .10 were entered into a regression equation, and results indicated that only Maternal Distress (a variable created by combining scores on the Beck Depression Inventory and the Pare nting Stress Index) and Maternal Negative Talk (combining obser ved criticism of and sarcasm toward the child) predicted dropou t (approximately 56% of dropouts were reliably predicted). A
6 study by Capage et al. (2001) also found that ma ternal parenting stre ss predicted attrition from PCIT. In response to the Werba et al. (in press) findings, changes were made to the PCIT protocol, prior to this study, to reduce attriti on by specifically targe ting maternal distress. Based on findings by Prinz and Miller (2004), therap ists were instructed to discuss briefly with parents their personal stressors unrelat ed to child stressors (e.g., job stress, health problems), in each treatment session. Prinz and Miller had found that addressing parents other life concerns in treatm ent resulted in significantly lower dropout rates relative to families whose therapists focused only on the parent-child intervention. Although additional time was allotted in each PCIT se ssion to address parents personal concerns, this enhancement was not implemented in a controlled or highly structured way. A second change to the protocol made to decrease attrition was based on retrospective parent self-reports ( n = 31) of their reason for dropping out of an earlier study (Boggs et al., 2004). The primary r easons reported for dropping out included logistical problems, such as being unable to obtain transportation or child-care for siblings ( n = 11, or 35%); feeling that treatment was not progressing quickly enough or feeling unable to wait for treatment to begin ( n = 6, or 19%); and dislike of the treatment approach or techniques ( n = 5, or 16%; Boggs et al.). In response, this study protocol included provision of child care for siblings of target children during assessment and treatment sessions and monetary reimbursement for gasoline ($3.00) given at the end of each session. The purpose of this study was to extend th e Werba et al. (in press) findings by examining conceptually based, pretreatment predictors of attrition in PCIT with
7 additional parent support components integrat ed into the protocol. We also sought to examine predictors of attrition from followup. Related to treatment attrition, our first hypothesis was that the proportion of familie s dropping out of the study would be lower than the proportion of dropouts in the Werba et al. study. Moreover, because we attempted to decrease dropout re lated to maternal distress via facilitative listening to personal stressors, we expected that Matern al Distress would no longer differ between dropouts and completers as previously. Howe ver, consistent with Werba et al., we hypothesized that Maternal Negative Talk at pretreatment would be higher among dropouts than treatment completers. We also hypothesized that treatment dr opouts would have less Maternal Total Praise compared to completers. Werba et al. (in press) had examined maternal prosocial behavior as a broad composite category of positive attention at pretreatment and did not find it to distinguish dropouts from complete rs. However, because of the strong emphasis in PCIT on increasing positive feedback, we narrowed the scope of the prosocial behavior parent-child interaction variable. Maternal Total Praise is more directly related to skills taught in treatment and a more salient counterpart to Maternal Negative Talk. SES did not emerge as a significant predicto r of dropout in Werba et al. (in press), but the logistical barriers recalled by the dropout parents in that sample (Boggs et al., 2004) were related to sociodemographic advers ity, and SES has consistently emerged as a predictor of dropout in studies of parent training programs (Dumas & Wahler, 1983; Kazdin, 1990; Kazdin, Mazurick, & Bass, 1993; Armbruster & Fallon, 1994; Prinz & Miller, 1994; Kazdin, Stolar, & Marcia no, 1995; McCabe, 2002; Peters, Calam, & Harrington, 2005). Thus, we hypothesized that SES would distinguish completers and
8 dropouts. We expected the linear combinati on of these variables (Maternal Distress, Maternal Negative Talk, Maternal Total Prai se, and SES) would si gnificantly predict group membership (i.e., treatment completers vs. treatment dropouts), with Maternal Distress contributing least to the equation. Because the treatment study addressed reasons for dropout reported most often in earlier res earch (transportation and child care), we also hypothesized that the proportion of families citi ng logistical problems as their reason for dropout would be lower than in the Boggs et al. (2004) study. Based on their study of father involvemen t in treatment, Bagner and Eyberg (2003) hypothesized that social support from therapis ts during PCIT was related to treatment gains and that loss of so cial support after treatment related to loss of gains. In this study, families that completed PCIT were randomly assigned to either an assessment-only condition or a maintenance treatment condition designed to maintain treatment gains (low maternal stress and few child disruptive beha viors) during a two-y ear follow-up period. Based on the Bagner and Eyberg (2003) hypothesi s as well as earlier fi ndings of relapse in maternal parenting stress after treatment (Eyberg et al., 2001), we expected Maternal Distress to differ in the assessment-only condition between comple ters and dropouts but not to differ between completers and dropout s in the maintenance treatment condition during follow-up. Because a continued rela tionship between therapist and families participating in follow-up has been associated with increased retention (Capaldi & Patterson, 1987), we also expected that th e proportion of families dropping out of maintenance treatment would be lower than the proportion of families dropping out of the assessment-only group.
9 In an earlier PCIT follow-up study (Hood & Eyberg, 2003), participating mothers were distinguishable from mothers lost to follow-up in age and intellectual functioning. Therefore, in this study we hypothesized that the assessment-only and maintenance treatment follow-up completers and dropouts wo uld differ significantly in Maternal Age and Intellectual Functioni ng, with dropouts being younger with lower intellectual functioning. Because follow-up dropouts have reported lower treatment satisfaction in studies of other treatments (Luk et al., 2001) we hypothesized that consumer satisfaction following PCIT would distinguish followup dropouts and completers in both the assessment-only and maintenance treatment conditions, with dropouts in both conditions reporting lower treatment satisfaction. Given our hypothesis that Ma ternal Distress would dist inguish the assessment-only but not maintenance treatment condition comp leters versus dropouts, we sought to identify a linear combination of variable s predicting follow-up group membership (i.e., follow-up completers versus dropouts) for both the assessment-only and the maintenance treatment conditions. We hypothesized that Maternal Age, Maternal Intellectual Functioning, Maternal Distress, and Treatment Satisfaction would predict follow-up group membership; however, we expected that Maternal Distress would contribute least to the equation for the maintenance treatment group.
CHAPTER 2 METHOD Participants Participants were 99 mother-c hild dyads of 3to 6-year -old children with disruptive behavior disorders enrolled in a treatment study examining the efficacy of maintenance treatment following completion of PCIT. Child ren (69 boys, 30 girls) were referred for treatment by pediatricians, child psychiatrist s, child neurologists, teachers, and day care providers. The childrens mean age was 4 years, 4 months ( SD = 1 year, 1 month), and racial/ethnic composition was 76% Caucasian, 11% Biracial, 8% Af rican American, 4% Hispanic, and 1% Asian. For mothers, the racial/ethnic breakdown was 84% Caucasian, 7% African American, 5% Bi racial, and 4% Hispanic. With all five socioeconomic categories represented, families had a mean SES of 38.44 ( SD = 13.98), placing them, on average, in the lower middle SES range (e.g., skilled craftsmen, clerical, sales workers) according to the Hollingshead (1975) Index. Ma le caregivers in 59 families participated in treatment, but their data we re not included in this study. Ma ternal caregivers included 92% biological or adoptive mothers, 3% st epmothers, 4% grandmothers, and 1 foster mother. Mothers mean age was 33.78 years ( SD = 9.49 years), and most mothers (58%) were married; 6% were separated, 17% were divorced, 17% were single, and 1 was widowed. For inclusion in this study, children had to meet Jens en et al. (1996) criteria for oppositional defiant disorder (ODD), which re quires both categorical and dimensional indices of the disorder. The children met diagnostic criteria for ODD on the Diagnostic Interview Schedule for Children-IV-Parent (DIS C-IV; Shaffer, Fisher, Lucas, Dulcan, & 10
11 Schwab-Stone, 2000) and obtaine d a clinically elevated ( T > 61) score on the Aggressive Behavior subscale of the Child Behavi or Checklist (CBCL/2-3; Achenbach, 1992; CBCL/4-18; Achenbach, 1991). Partic ipants also had to obtain a standard score of at least 75 on a cognitive screening measure (Peabody Picture Vocabulary Test-III, Dunn & Dunn, 1997, for children; Wonderlic Personnel Te st, Dodrill, 1981, for parents). Children were excluded if parents described severe sensory or mental impairment (e.g., blindness, autism) during the clinical interview. Ch ildren taking psychotropic medications for behavior problems (29%) had to maintain a consistent medication regimen and dosage schedule for at least one month before enrol ling in the study, and mothers were asked not to alter their childs medicati on or dosage during treatment. Parents of children not taking psychotropic medication were asked not to begin medication for their child during treatment. In addition to ODD, children in th is study met DISC-IV cr iteria for attention deficit hyperactivity disorder (74%), conduct disorder (46%), separation anxiety disorder (26%), and major depressive disorder (4%). Of the 99 families that participated in PCIT, 63 completed treatment. Following treatment, thirty-two families were randomly assigned to an assessment-only follow-up condition and thirty-one to a maintenance tr eatment condition. In the assessment-only follow-up condition, the childrens (22 boys, 10 gi rls) mean age was 4 years, 6 months ( SD = 1 year), and racial/ethnic compositi on was 81% Caucasian, 13% Biracial, 3% African American, and 3% Asian. In additi on to meeting ODD criteri a at pretreatment, children in the assessment-only condition al so met criteria for attention deficit hyperactivity disorder (65%), conduct disorder (46%), separation anxi ety disorder (21%), and major depressive disorder (6%). For mothers, the racial/ethnic breakdown was 94%
12 Caucasian and 6% Biracial. Fam ilies had a mean SES of 41.16 ( SD = 14.66), placing them, on average, in the upper middle SE S range (e.g., medium business, minor professional, technical) according to the Holl ingshead (1975) Index. Maternal caregivers included 97% biological or a doptive mothers and 3% stepmothers. Mothers mean age was 34 years ( SD = 9 years), and most mothers (61 %) were married; 23% were divorced, 10% were single, and 7% separated. For the maintenance treatment follow-up condition, the childrens (20 boys, 11 girls) mean age was 4 years, 4 months ( SD = 1 year), and racial/ethnic composition was 71% Caucasian, 13% Biracial, 10% African Am erican, and 7% Hispanic. In addition to meeting ODD criteria at pret reatment, children in the maintenance treatment condition also met criteria for atten tion deficit hyperactivity diso rder (78%), conduct disorder (43%), separation anxiety disorder (33%), and major depressive disorder (3%). For mothers, the racial/ethnic breakdown was 77% Caucasian, 10% African American, 7% Hispanic, and 7% Biracial. Families had a mean SES of 36.20 ( SD = 9.41), placing them, on average, in the lower middle SES range (e .g., skilled craftsmen, clerical, sales worker) according to the Hollingshead (1975) Index. Ma ternal caregivers included 90% biological or adoptive mothers, 7% grandmothers, and 3% stepmothers. Mothers mean age was 36 years ( SD = 9 years, 6 months), and most mo thers (57%) were married; 20% were divorced, 17% were single, 3% separated, and 3% widowed. Measures Demographic Measures Mothers completed a demographic questionn aire including descriptive information on age, sex, race/ethnicity, occupation, educat ion level, and marital status of family members. The Hollingshead Four Factor Inde x of Social Status (Hollingshead, 1975), an
13 index of socioeconomic status that is co mprised of education, occupation, sex, and marital status, was calculated based on information obtained from the demographic questionnaires. Diagnostic Measures Three measures were used to screen for ODD. The Diagnostic Interview Schedule for Children, Fourth Edition (DISC-IV; Shaffer et al., 1997) was administered to all mothers to assess ODD symptom frequency and duration as well as functional impairment in the home and school se ttings. In its earlier versions, the DISC was shown to have moderate to substantial te st-retest reliability and internal consistency (Fisher et al., 1993; Jensen et al., 1995), a nd the test-retest reliab ility of the DISC-IV compares favorably with the earlie r versions (Shaffer et al., 2000). Child Behavior Checklist for 4 to 18 year olds (CBCL/4-18; Achenbach, 1991) consists of 118 behavior-problem items rated by the parent on a 3-point scale from (0) not true to (2) very true or often true to assess the frequency of a variety of specific behaviors in children during th e past 6 months. One-week te st-retest reliability for the CBCL/4-18 Aggressive subscale has b een reported at .91 (Achenbach). Child Behavior Checklist for 2 to 3 year olds (CBCL/2-3; Achenbach, 1992) is similar in format to the CBCL/4-18 and contains 99 items rated by the parent for frequency in the past 2 months on a 3-point scale. Test-retest reliability of the CBCL/2-3 Aggressive subscale has been reported at .85 over a 3-week period (Koot, Van Den Oord, Verhulst, & Boomsman, 1997). Cognitive Screening Measures Cognitive screening of parents was completed using the Wonderlic Personnel Test (WPT; Dodrill, 1981), a 12-minute timed test developed to screen adult intellectual
14 ability that correlates with WAIS Full Scale IQ scores at .93 (Dodrill). The Peabody Picture Vocabulary Test-Third Edition (PPVT-III; Dunn & Dunn, 1997) was used to screen children in this study. The PPVT-III is a well-standardized measure of receptive language, which correlates highly (.90) with th e Wechsler Intelligence Scale for Children III Full Scale IQ. Maternal Distress Variable Consistent with earlier research (Werba et al., in press), we found a strong correlation (.58) between total scores on the Beck Depression Inventory Second Edition (BDI-II; Beck, Steer, & Brown, 1996) and the Parenting Stress Index-Short Form (PSI-SF; Abidin, 1995). Therefore, sc ores on these two measures were standardized and combined as in Werba et al. into a composite Maternal Distress variable. The PSI-SF is a 36-it em parent self-report scale that measures stress in the parent-child relationship due to parent distress, difficult ch ild behavior, and dysfunctional parent-child interaction. The BDI-II is a 21-item multiple-cho ice self-report measure of depressive symptomatology. Studi es have reported 1-week te st-retest reliability at .93 (Beck, Steer, & Brown). Internal consistency for the Maternal Distress composite score created for this study was .72. Observational Measures Two composite variables from the Dyadic ParentChild Interaction Coding System Second Edition (DPICS-II; Eyberg, Bessmer, Newcomb, Edwards, & Robinson, 1994) were created for this study. The DPICS is a behavioral coding system that measures the quality of parentchild social interactions during three 5-min standardized play situations (c hild-led, parent-led, and cleanup) that vary in the degree of parental control required. Four categories of maternal verbalizations coded during
15 these parentchild interaction observations were analyzed fo r this study: Criticism, Smart Talk (sarcasm), Labeled Praise, and Unlabele d Praise. Criticism and Smart Talk were combined into one composite Negative Talk category, and Labeled and Unlabeled Praise were combined into one composite Total Praise category. Kappas for Criticism, Smart Talk, Labeled Praise, and Unlabeled Pr aise were .51, .33, .64, and .87, respectively. Because the Kappa value for Smart Talk was unacceptably low (Fleiss, 1981), we examined the Kappa confusion matrix, whic h indicated that whenever one of two observers coded a Smart Talk and the othe r coded another category, the other category was Criticism in 100% of occurrences. Sim ilarly, when one of two observers coded Criticism and the other another category, th at other category was Smart Talk 100% of occurrences. Thus, by combining these two categ ories into the single composite Negative Talk category, the data set for Negative Ta lk contained no inter-coder disagreements. Consumer Satisfaction After treatment, mothers completed the Therapy Attitude Inventory (TAI), a parent-report instrument speci fically designed to evaluate consumer satisfaction for parent training, parent-child treatments, and family therapy (Eyberg, 1993). The TAI includes 10 items addressing the impact of th erapy on parenting skills and child behavior. Parents rate items on a scale from 1 (dissa tisfaction with treatment or worsening of problems) to 5 (maximum satisfaction with treatment or im provement of problems). The item ratings are summed to yield a total sc ore between 10 and 50. Earlier studies have demonstrated 4-month test-retest stability (.85; Brestan, Jacobs, Rayfield, & Eyberg, 1999) and internal consiste ncy (.88; Eisenstadt, E yberg, McNeil, Newcomb, & Funderburk, 1993), and TAI scores from differ ent parent training methods have provided
16 evidence of discriminative validity (Eyberg & Matarazzo, 1980). Cronbachs alpha for the TAI in this study was .78 at posttreatment. Procedure Families were screened for inclusion during the first of two pre-treatmenttreatment assessment visits. These visits included a clinical interview and child diagnostic interview with the parents, the cognitive sc reening measures, and several questionnaires including a demographic questionna ire, the BDI-II, and the PSI-S F. At each visit, parentchild dyads were also videotaped in three stru ctured play situations later coded using the DPICS system. Families were then seen for PCIT by two graduate student co-therapists during weekly one-hour sessions, which were videotaped for later integrity checking. All sessions were conducted according to proce dures outlined in the treatment manual (Eyberg and the Child Study Laboratory, 1999). Therapists were asked to record the reason(s) for dropout for any family th at discontinued treatment prematurely. Undergraduate research assistants us ed session checklists to code treatment integrity for a randomly selected 50% of the session tapes for each family. Treatment integrity, calculated as percent agreemen t with the session checklist, was 90%. A randomly selected 50% of the coded session tapes were recoded by a second undergraduate research a ssistant to assess the re liability of the treatm ent integrity coding. Intercoder percentage agreement reliability was 91%. After successful completion of PCIT, fam ilies were randomly assigned to either a maintenance treatment group or follow-up assessment-only group. Families receiving maintenance treatment ( n = 31) were called once per mont h by their PCIT therapist to assess maintenance and provide support or advice as indicated. Families in the
17 assessment-only group ( n = 32) were invited to call thei r PCIT therapist for support or advice if needed, but otherwise did not have fu rther contact with their therapist. For both the maintenance and assessment-only groups, families were contacted by phone every three months by an assessor masked to thei r follow-up assignment (i.e., maintenance treatment plus assessment versus assessment-onl y) to complete measures relating to child behavior and maternal functioning. At the oneand two-year follow-up points, families were seen again for full two-visit assessments that included the behavioral observation measures as well as questionnaire measures. Observational Data. At the pre-treatment, post-treatment, and oneand two-year follow up assessments, parents and children were observed in three 5-min structured play situations (Child-Led Play, Parent-Led Pl ay, and Clean Up) on two occasions, one week apart. For this study, only data from the Parent-Led Play (PLP, 5 minutes) and Clean Up (CU; 5 minutes) situations (10 minutes from each occasion) at pretreatment were used, averaged across the two situa tions and the two observation occasions. For the 12 cases in which one of the two observations of the sa me situation (i.e., PLP or CU) was missing, data from the one observation that was not mi ssing was substituted for the average of the two situations.
CHAPTER 3 RESULTS Analysis of Observational Data. Frequency counts for both categories of verbalizations (i.e., Maternal Negative Tal k, Maternal Total Praise) were calculated and averaged across the two pretreatment asse ssments. The average frequencies of each verbalization category were then averaged ac ross observation situations (i.e., PLP, CU). For cases in which one of the two observations of the same situation was missing (i.e., PLP or CU), due to assessor error in record ing the observation or technical problems with the recording equipment, data from the observation that was not missing was substituted for an average of the two situations. Because the data had positively skewed distributions, log transformations were performed to achieve more normal distributions. In cases where no Total Praise or Negative Talk occurred (i.e., 0 values), a nominal value of .01 was added prior to transformation, as 0 values cannot undergo log transformation. Log transformations resulted in kurtosis and skewness values within acceptable limits (i.e., < 2 ), although Total Praise remained kurtotic (kurtosis = 3.24). Treatment attrition was 36% in this study. To test the significance of the difference between the proportion of families that dropped out of the Werba et al. study and the proportion that dropped out of this study, we calculated a z score by dividing the observed difference between the proportions by the estimate of the standard error of the difference. The proportion of families dropping out of this study (36 of 99) did not differ significantly from the proportion in Werba et al. (17 of 52), z = -.026, p = .49. 18
19 One-way ANOVAs were used to examine pretreatment differences in Maternal Distress, Total Praise, and SES between PC IT dropouts and completers, and the Welch t test was used to examine Negative Talk. Initia l evaluation of the data suggested that SES and Maternal Distress were univariate normal, with skewness and kurtosis estimates less than 1 Log transformations of Total Praise and Negative Talk resulted in roughly univariate normal distributions (i.e., < 2 ). Intercorrelations among potential predictor variables for treatment attri tion are shown in Table 3-1. Table 3-1. Pearson Correlation Matrix for Potential Treatment Attrition Predictor Variables Potential predictor variables 1. 2. 3. 4. 1. Maternal Distress 1 2. SES -.14 1 3. Negative Talk -.07 -.19 1 4. Total Praise -.17 .19 -.22* 1 p < .05. As shown in Table 3-2, treatment comple ters and dropouts diffe red significantly in SES, F (1, 98) = 8.03, p = .006, and Total Praise, F (1, 97) = 4.17, p = .044. SES and Total Praise for dropouts were significantly lower th an for completers. Maternal Distress did not differ between treatment completers and dropouts, F (1 95) = .23, p = .633. Of the 38 mothers who entered treatment with clinically elevated depression (BDI-II score > 14) and parenting stress (PSI-SF score > 90), 61% (23 of 38) completed treatment. Of these 23 completers, 65% (15 of 23) no longer reported clinically significant levels of distress on both measures at posttreatment. Levenes Te st of Equality of Variance was significant for Negative Talk ( p = .025), which can increase the pr obability of Type I error with groups of different sizes (Refinet ti, 1996). For this reason, Welchs t test was used to keep Type I error close to nominal signi ficance level (Zimmerman, 2004). This test showed a significant difference in Negative Talk between dropouts and completers, t (1,
20 93.39) = 7.08, p = .009, with dropouts having significantly more Negative Talk than completers (Table 3-2). Table 3-2. Means and Standard Deviations for Potential Treatment Attrition Predictor Variables Treatment Completers Treatment Dropouts Potential predictor variables M SD M SD F or t p SES 41.40 14.08 33.36 12.62 8.03 .006 Total Praise .23 .69 -.07 .73 4.17 .044 Maternal Distress .03 .83 -.06 .99 .23 .633 Negative Talk .61 .42 .79 .28 7.08 .009 Discriminant function analysis (DFA) was conducted to identify a linear combination of variables predictive of group membership (treatment completers versus dropouts). Total Praise, Negative Talk, SES, and Maternal Distress were included as independent variables. The discrimina nt function was significant, Wilks = .85, 2 (4, N = 95) = 14.59, p = .006. A jackknife procedure was used to obtain a cross-validation estimate to account for potential sample bias and provide a more conservative estimate of classification. Using this procedure, 62% of families were classified correctly. For dropouts, 24 of 34 (71%) families were classifi ed correctly and for completers, 35 of 61 (57%) families were classified correctly. Loadings for the four variables on group membership (i.e., the function structure matrix) resulted in correlati ons of .73 for SES, -.63 for Negative Talk, .49 for Total Praise, and .12 for Maternal Distress. Examina tion of standardized canonical discriminant function coefficients suggested that when controlling for the remaining predictors, SES was the best predictor of group membership (.67), followed by Negative Talk (-.48), Total Praise (.35), and Maternal Distress (.27). Afifi a nd Clark (1996) recommended a
21 cutoff value of .3 for interpreting predictor variables. Therefore, treatment dropout was predicted by lower SES, more Maternal Negative Talk, and less Ma ternal Total Praise. Primary reasons for dropout were obtaine d for 31 of 36 (86%) families and are listed in Table 3-3. To test the significance of the difference between the proportions of families that indicated logistical problems as their primary reason for dropout in this study compared to Boggs et al. (2004), we calculated a z score by dividing the observed difference between the proportions by the estimate of the standard error of the difference. The proportion of families with logistical problems as their primary reason for dropout in this study (4 of 31) was significantly lower th an in the Boggs et al. sample (11 of 23), z = -2.86, p = .002. Table 3-3. Primary Reasons for Dropout Provided by Families for Dropping Out of PCIT Primary Reasons Number of families ( N = 31) Percent of families Family disagreed with treatment approach 8 26% Family too busy to participate in treatment 4 13% Life/Family stressors interfered 4 13% Logistical problems (e.g., transportation) 4 13% Childs behavior improved 3 10% Childs/parents health problems interfered 3 10% Other treatment wanted/needed 2 6% Childs school behavior pr oblems did not improve 1 3% Family lost custody of child 1 3% Family moved 1 3% For the assessment-only condition, oneway ANOVAs were used to examine differences between completers and dropouts in Maternal Ag e, Maternal Intellectual Functioning, Maternal Distress, and Treatment Satisfaction. Initial ev aluation of the data suggested that Maternal Dist ress, Maternal Intellectual Functioning, and Maternal Age were roughly univariate normal, with sk ewness and kurtosis estimates less than 1 Treatment Satisfaction was negatively skewed (i.e., > 1 ); therefore, square root
22 transformation was used to normalize the di stribution. As shown in Table 3-4, results suggested that follow-up completers and dropouts differed significantly in Maternal Distress, F (1, 29) = 6.47, p = .017, but were indistinguishable in Maternal Age, F (1, 30) = .01, p = .923, Maternal Intelle ctual Functioning, F (1, 30) = .26, p = .616, and Treatment Satisfaction, F (1, 28) = .42, p = .523. Table 3-4. Means and Standard Deviations for Potential Follow-Up Attrition Predictor Variables in the Assessment-Only Condition Follow-up Completers Follow-up Dropouts Potential predictor variables M SD M SD F p Maternal Distress -.47 .75 .27 .83 6.47 .017 Maternal Age 34.20 7.41 33.88 10.77 .01 .923 Maternal Intellectual Functioning 108.33 12.23 106.31 9.92 .26 .616 Treatment Satisfaction 6.86 .20 6.81 .24 .42 .523 DFA was conducted to identify a linear co mbination of variables predictive of group membership (i.e., follow-up completers versus dropouts) for the assessment-only condition. Intercorrelations among potentia l follow-up predictor variables for the assessment-only condition are presented in Table 3-5. Maternal Distress, Maternal Intellectual Functioning, Mate rnal Age, and Treatment Satisfaction were included as independent variables. The discriminant function was not significant, Wilks = .76, 2 (4, N = 26) = 6.10, p = .192. Table 3-5. Pearson Correlation Matrix for Potential Follow-Up Attrition Predictor Variables in the Assessment-Only Condition Potential predictor variables 1. 2. 3. 4. 1. Maternal Distress 1 2. Maternal Intellectual Functioning -.03 1 3. Maternal Age .10 .11 1 4. Treatment Satisfaction -.01 -.12 -.20 1 For the maintenance treatment condition, one-way ANOVAs were used to examine differences between completers and dropout s in Maternal Intellectual Functioning,
23 Maternal Distress, and Treatment Satisfaction. Levenes Test of Equality of Variance was significant for Maternal Age ( p = .025). Therefore, Welchs t test was used to examine group differences for this variable to keep T ype I error close to nominal significance level (Zimmerman, 2004). As shown in Table 36, results suggested that dropouts had significantly lower Intellectual Functioning than completers, F (1, 30) = 12.40, p = .001, but the groups did not differ in Maternal Distress, F (1, 30) = 1.98, p = .17, Treatment Satisfaction, F (1, 29) = .03, p = .858, or Maternal Age, t (1, 25.806) = 2.49, p = .127. Table 3-6. Means and Standard Deviations for Potential Follow-Up Attrition Predictor Variables in the Maintenance Treatment Condition Follow-up Completers Follow-up Dropouts Potential predictor variables M SD M SD F or t p Maternal Distress -.02 .80 .39 .77 1.98 .17 Maternal Age 38.11 11.11 33.33 5.25 2.49 .127 Maternal Intellectual Functioning 113.05 9.61 101.17 8.35 12.40 .001 Treatment Satisfaction 6.84 .18 6.83 .24 .03 .858 DFA was conducted to identify a linear co mbination of variables predictive of group membership (i.e., follow-up completers versus dropouts) for the maintenance treatment condition. Intercorre lations among potential predic tor variables for follow-up attrition are presented in Table 3-7. The di scriminant function was significant, Wilks = .53, 2 (4, N = 29) = 15.89, p = .003. A jackknife procedure wa s used to obtain a crossvalidation estimate to account fo r potential sample bias and provide a more conservative estimate of classification. Using this proce dure, 79% of all families were classified correctly. For maintenance treatment condition dropouts, 10 of 12 (83%) families were classified correctly and for completers, 13 of 17 (76%) families were classified correctly. Table 3-7. Pearson Correlation Matrix for Potential Follow-Up Attrition Predictor Variables in the Maintenance Treatment Condition Potential predictor variables 1. 2. 3. 4. 1. Maternal Distress 1
24 2. Maternal Intellectual Functioning -.26 1 3. Maternal Age .05 .18 1 4. Treatment Satisfaction .10 -.17 .04 1 Loadings for the four variables on group membership (i.e., the function structure matrix) resulted in correlati ons of .85 for Maternal Intellectual Functioning, .36 for Maternal Age, -.28 for Maternal Dist ress, and .03 for Treatment Satisfaction. Examination of standardized canonical discrimi nant function coeffici ents suggested that when controlling for the remain ing predictors, Maternal In tellectual Functioning was the best predictor of group membership (.90), followed by Maternal Age (.45), Treatment Satisfaction (.27), and Maternal Distress (.24). Following the Afifi and Clark (1996) recommendation of .3 as a cutoff for interpre ting predictor variable s, results suggested that mothers lower intellectual functi oning and younger age predicted dropout from maintenance treatment. Follow-up attrition was 53% for the a ssessment-only condition and 39% for the maintenance treatment condition in this study. To test the significance of the difference between proportions of families that dropped out of the assessment-only and maintenance treatment conditions, we calculated a z score by dividing the observed difference between the proportions by the estimate of the standard error of the difference. The proportion of families that dropped out of the maintenance treatment group (12 of 31) did not differ significantly from the proportion that droppe d out of the assessment-only group (17 of 32), z = .13, p = .55.
CHAPTER 4 DISCUSSION One important finding of this study was that maternal distress did not emerge as a predictor of dropout. Although maternal distress was implicated in dropout in two earlier PCIT studies examining attrition (Capage et al., 2001; Werba et al., in press), treatment completers and dropouts were indistinguishable in maternal distress in this study, and maternal distress did not contri bute to the linear combinati on of variables predictive of dropout. Approximately 60% of mothers who repor ted clinical elevati ons in depressive symptoms and parenting stress at pretre atment completed PCIT, and 65% of these completers no longer reported both clinical elevations following PCIT. Our findings suggest that incorporating a dditional parent support com ponents to the protocol (e.g., increased time spent speaking with parents a bout stressors apart from the parent-child intervention, monetary reim bursement) may have contributed to the retention of distressed mothers. Consistent with our hypothesis, a smaller proportion of families indicated logistical problems as their primary reason for dropout than in the Werba et al. (in press) sample. The provision of child care and mone tary reimbursement for treatment sessions may have alleviated these barriers to treatment completion for some families. However, because SES emerged as the strongest predictor of attrition, additional logistical supports provided seemed insufficient to overcome the greater adversity faced by families in lower SES groups. Because SES is a proxy variable, co ntinued research is needed to elucidate the specific barriers associ ated with socioeconomic disadvantage (e.g., lack of 25
26 transportation to sessions). Additionally, pe rceived barriers to tr eatment participation, which have been shown to predict attrition (K azdin et al., 1997), will need to be assessed before families begin PCIT. In contrast to retrospective assessment of barriers, which may include only those families who complete treatment and posttreatment assessments, prospective assessment will enable analysis of the relationship between perceived barriers to treatment and dropout and will include dropout families. As hypothesized, less total praise and more negative talk by mothers at pretreatment both predicted dropout. These two va riables are characteristic of a parenting style that is antithetical to the core pr inciples and skills ta ught in PCIT and many parenting programs. Consistent with this inte rpretation, it is notable that 26% of families that dropped out of treatment reported dis like or disagreement with the treatment approach. These findings suggest that pretreatment assessment of treatment acceptability may be important in treatment planning. In addition, treatment acceptability has been associated with a stronger therapeutic alli ance and fewer perceived barriers (Kazdin, Marciano, & Whitley, 2005), highlighting the im portance of the therapeutic alliance in treatment retention. Modifications made to the protocol aime d at reducing treatment attrition in this study were unsuccessful. The dropout rate was comparable to Werba et al. (in press). Although we expected lower attrition due to additional support components, the 36% attrition rate was consistent with dropout rates demonstrated in other PCIT studies focusing on children with DBDs (e.g., 35% in Eisenstadt et al., 1993; 33% in Werba et al., in press).
27 Inclusion of both the assessment-only a nd maintenance treatment conditions in follow-up analyses allowed for more controlle d examination of the role of maternal distress in attrition. Our hypothe sis about the role of mate rnal distress in predicting dropout from the assessment-only condition wa s supported. Maternal distress re-emerged as a predictor of dropout at follow-up in th e subset of families that did not maintain regular contact with their therapists. Our findings are c onsistent with Bagner and Eybergs (2003) hypothesis of the importance of therapist support in maintenance of treatment gains as well as findings that pare nting stress relapses after treatment ends (Eyberg et al., 2001). Monthly telephone c ontact may provide families with needed support and represent a successful reten tion strategy for distressed families. For families in the maintenance treatment condition, which involved monthly contact via telephone with their therapists, dropouts had lower intellectual functioning than completers. Intellectual functioning represented the strongest predictor of attrition and, in combination with maternal age, predicted dropout from follow-up in this treatment condition. Contrary to our hypotheses, treatment satisfaction was unrelated to dropout, and attrition rates we re not lower for maintenance treatment than assessmentonly follow-up families. Although maintenance treatment seemed to alleviate maternal distress, the format of maintenance sessions may ha ve been less helpful for mothers with lower intellectual functioning. Maintenance sessi ons involved problem solv ing via telephone, when mothers may have been more distracted by their environment than when seen in-person in the clinic setting. As a result, these te lephone maintenance sessions might have been less effective and could have contributed to dropout. Future studies should consider use
28 of supportive telephone calls in conjunction with booster sessions in the clinic setting, with the goals of alleviating distress and reviewing treatm ent strategies in a more concrete manner. Another option recomme nded by Eyberg, Edwards, Boggs, and Foote (1998) would be to conduct booster sessions either in families homes or locations close to their homes. It might also be important for therapists and assessors to emphasize frequently the importance of continued participation in maintenance treatment for treatment gains to persist over time. This concept may become less intuitive for mothers of lower intellectual functioni ng as sessions move from bei ng conducted in-person to via telephone. One limitation of this study is that changes made to the PCIT protocol in response to earlier findings (i.e., parent and logistical support com ponents) were not manipulated experimentally in the study design. That is, fa milies were not randomly assigned either to an additional support condition or no addi tional support condition during standard treatment. Instead, all families in this study received additional facilitative listening to concerns beyond the parent-child interacti on as well as monetary reimbursement and child care. As such, it is not possible to conclu de definitively that re tention of distressed mothers in this study was related to the addi tional therapist support or that logistical barriers were less often cited as the reason for attrition due to the provision of child care and monetary reimbursement. A second limitation is the absence of father data included in the analyses. Because more mothers than fathers tend to participat e in parent training programs, including PCIT, and maternal participation tends to be more consistent than for fathers (Budd & OBrien, 1982; Webster-Stratton, 1985; Bagner & Eyberg), we sought to focus the scope
29 of this study solely on maternal variables related to attriti on. However, research suggests that father involvement in PCIT may be re lated to improved long-term outcome (Bagner & Eyberg, 2003). Attrition from PCIT has been related to absence of gains at long-term follow-up (Boggs et al., 2004). Future research should consider pate rnal variables for their unique role in predicting attrition. This study extends the existing literature through identificat ion of significant predictors of follow-up attrition. However, s lightly more than half of families in the assessment-only condition dropped out of follow-up. Although long-term outcomes for treatment dropouts have been studied (Boggs et al., 2004), increased retention of families during follow-up would allow for a greater understanding of long-term functioning following PCIT. Despite these limitations, the four conceptually based variables examined in relation to treatment attrition contributed to an improved formula for predicting attrition from PCIT relative to Werba et al. (in pre ss), which correctly classified 62% of dropout families. Furthermore, the results extended the literature by elucidating predictors of attrition in follow-up, such that 63% of dropouts were correctly classified in the assessment-only condition and 83% in the maintenance treatment condition. Because attrition from both treatment and follow-up presents an ongoing challenge to researchers and clinicians, our findings point to a need for conti nued study of attrition and development of empirically based retention st rategies in PCIT. Our results suggest that examination of predictors of attrition informs small changes in treatment that may enhance treatment outcomes for more families completing treatment.
3 LIST OF REFERENCES Abidin, R. (1995). Parenting Stress Index manual (3rd ed.). Odessa, FL: Psychological Assessment Resources. Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 Profile. Burlington, VT: University of Vermont Department of Psychiatry. Achenbach, T. M. (1992). Manual for the Child Behavior Checklist/2-3 and 1992 Profile Burlington, VT: University of Verm ont Department of Psychiatry. Afifi, A. A., & Clark, V. (1996). Computer-aided multivariate analysis. London: Chapman & Hall. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders. (4th ed., text revision). Wash ington, DC: American Psychiatric Association. Armbruster, P., & Fallon, T. (1994). Clinical, sociodemographic, and systems risk factors for attrition in a children's mental health clinic. American Journal of Orthopsychiatry, 64, 577-585. Armbruster, P., & Kazdin, A. E. (1994) Attrition in ch ild psychotherapy. Advances in Clinical Child Psychology, 16, 81-108. Bagner, D. M., & Eyberg, S. M. (2003). Fath er involvement in pa rent training: When does it matter? Journal of Clinical Child and Adolescent Psychology, 32, 599-605. Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory-II San Antonio, TX: Psychological Corporation. Boggs, S. R., Eyberg, S. M., Edwards, D. L., Rayfield, A., Jacobs, J., Bagner, D., & Hood, K. K. (2004). Outcomes of parent-chi ld interaction therapy: A comparison of treatment completers and study drop outs one to three years later. Child & Family Behavior Therapy, 26(4) 1-22. Brestan, E. V., & Eyberg, S. M. (1998). E ffective psychosocial tr eatments of conductdisordered children and adolescents: 29 years, 82 studies, and 5,272 kids. Journal of Clinical Child Psychology, 27, 180-189. 30
31 Brestan, E., Jacobs, J., Rayfield, A., & Eyberg, S. M. (1999). A consumer satisfaction measure for parent-child treatments and its relationship to measures of child behavior change. Behavior Therapy, 30, 17-30. Budd, K. S., & OBrien, T. P. (1982). Father i nvolvement in behavior al parent training: An area in need of research. Behavior Therapist, 5, 85-89. Campbell, S. B. (2002). Follow-up and outcome studies. In Behavior problems in preschool children: Clinical and developmental issues (2 nd Ed., pp. 241-275). New York: Guildford Press. Capage, L. C., Bennett, G. M., & McNeil, C. B. (2001). A comparison between African American and Caucasian children referred for treatment of disruptive behavior disorders. Child and Family Behavior Therapy, 23(1), 1-14. Capaldi, D., & Patterson, R. (1987). An a pproach to the problem of recruitment and retention rates for longitudinal research. Behavioral Assessment, 9 169. Cohen, J. A., Mannarino, A. P., & Knudsen, K. (2005). Treating sexually abused children: 1 year follow-up of a randomized controlled trial. Child Abuse & Neglect,29 135-145. Dishion, T. J., & Patterson, G. R. (1992). Age effects in parent training outcome. Behavior Therapy, 23, 719-729. Dodrill, C. B. (1981). An economical method for the evaluation of general intelligence in adults. Journal of Consulting and ClinicalPsychology, 49, 668. Dumas, J. E., & Wahler, R. G. (1983). Predictors of treatment outcome in parent training: Mother insularity and socioeconomic di sadvantage. Behavioral Assessment, 5, 301-313. Dunn L. M., & Dunn, L. M. (1997). Examiners manual for the PPVTIII: Peabody Picture Vocabulary Test (3rd ed.). Circle Pines, MN: American Guidance Services. Egeland, B., Kalkoske, M., Gottesman, N., & Erickson, M. F. (1990). Preschool behavior problems: Stability and factors accounting for change. Journal of Child Psychology and Psychiatry, 31, 891-909. Eisenstadt, T. H., Eyberg, S. M., McNeil, C. B., Newcomb, K., & Funderburk, B. (1993). Parent-child interaction therapy with behavior problem children: Relative effectiveness of two stages and overall treatment outcome. Journal of Clinical Child Psychology, 22, 42-51. Eyberg, S. M. (1993). Consumer satisfaction measures for assessing parent training programs. In L. VandeCreek, S. Knapp, & T.L. Jackson (Eds .), Innovations in Clinical Practice: A Sourcebook (Vol. 12). Sarasota. Professional Resource Press.
32 Eyberg, S. M., Bessmer, J., Newcomb, K., Edwards, D.,& Robinson, E. A. (1994). Manual for the Dyadic ParentChild Interaction Coding SystemII. Social and Behavioral Sciences Documents (No. 2897). Retrieved April 17, 2005, from www.pcit.org Eyberg, S.M., and the Child Study Lab (1999). Pa rent child interacti on therapy: Integrity checklists and session materials. Retrieved April 17, 2005, from http://www.childstudylab.com Eyberg, S.M., Edwards, D., Boggs, S., & F oote, R. (1998). Maintaining the treatment effects of parent training: The role of booster sessions and other maintenance strategies. Clinical Psychology: Science and Practice, 5, 544-554. Eyberg, S. M., Funderburk, B. W., Hembree-Ki gin, T. L., McNeil, C. B., Querido, J. G., & Hood, K. K. (2001). Parent-child intera ction therapy with behavior problem children: One and two year maintenance of treatment effects in the family. Child & Family Behavior Therapy, 23(4), 1-20. Eyberg, S. M., & Matarazzo, R. G. (1980). Tr aining parents as therapists: A comparison between individual parent-c hild interaction training and parent group diactic training. Journal of Clinical Psychology, 42 594-606. Eyberg, S.M., Nelson, M.M., & Boggs, S.R. (in press). Evidence-based treatments for child and adolescent disrup tive behavior disorders. Journal of Clinical Child and Adolescent Psychology. Fagot, B. I., Loeber, R., & Reid, J. B. (1988). Developmental determinants of male-tofemale aggression. In G. W. Russell (Ed.), Violence in intimate relationships (pp. 91-105). Costa Mesa, CA: PMA Publishing Corp. Farrington, D. P. (1995). The stability a nd prediction of aggressive behavior. Gruppendynamik, 26, 23-40. Federal Bureau of Investigation (FBI), Crime in US: Uniform Crime Reports (J1.14). 1997, U.S. Government Printing Office: Washington, DC. Fisher, P. W., Shaffer, D., Piacentini, J., La pkin, J., Kafantaris, V., Leonard, H., et al. (1993). Sensitivity of the Diagnostic Inte rview Schedule for Children, 2nd edition (DISC-2.1) for specific diagnoses of children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 32 666-673. Flannery-Schroeder, E., Choudhury, M. S., & Kendall, P. C. (2005). Group and Individual Cognitive-Behavioral Treatments for Youth With Anxiety Disorders: 1Year Follow-Up. Cognitive Therapy & Research, 29, 253-259. Fleiss, J.L. (1981). The Measurement of Interrater Agreement. In Statistical Methods for Rates and Proportions. (2 nd Ed., pp. 212-304). New York: John Wiley & Sons.
33 Frick, P. J., & Loney, B. R. (1999). Outc omes of children and adolescents with oppositional defiant disorder and conduct disorder. In H. C. Quay & A. E. Hogan (Eds.), Handbook of disruptive behavior disorders (pp. 507-524). Dordrecht, Netherlands: Kluwer Academic Publishers. Gallagher, N. (2003). Effects of parent-ch ild interaction thera py on young children with disruptive behavior disorders. Bridges Practice-Based Research Syntheses, 1, 1-17. Guevara, J. P., Mandell, D. S., Rostain, A. L., Zhao, H., & Hadley, T. R. (2003). National estimates of health services expenditures for children with behavioral disorders: an analysis of the medical expenditure panel survey. Pediatrics, 112, e440-e446. Hodgins, S. (1994). Status at age 30 of children with conduct problems. Studies on crime and crime prevention, 3, 41-62. Hollingshead, A. B. (1975). Four-Factor Index of Social Status. Unpublished manuscript, Yale University, New Haven, CT. Hood, K. K., & Eyberg, S. M. (2003). Outcomes of Parent-Child In teraction Therapy: Mothers reports of maintenance three to six years after treatment. Journal of Clinical Child and Adolescent Psychology, 32, 419-429. Jensen, P., Roper, M., Fisher, P., Piacentini, J., Canino, G., Richte rs, J., et al. (1995). Test-retest reliability of the Diagnosti c Interview Schedule for Children (DISC 2.1). Parent, child, and combined algorithms. Archives of General Psychiatry, 52 61-71. Jensen, P.S., Watanabe, H.K., Richters, J.E., Roper, M., Hibbs, E.D., Salzberg, A.D., & Lui, S. (1996). Scales, diagnoses, and child psychopathology II: Comparing the CBCL and the DISC against external validators. Journal of Abnormal Child Psychology, 24, 151-168. Kazdin A. E. (1996). Dropping out of child psychotherapy: Issues for research and implications for practice. Clinical Child Psychology and Psychiatry, 1, 133-156. Kazdin, A. E. (1990). Premature termination from treat ment among children referred for antisocial behavior. Journal of Child Psychology & Psychiatry & Allied Disciplines, 31 415-425. Kazdin, A. E., Holland, L., & Crowley, M. (1997). Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting & Clinical Psychology, 65 453-463. Kazdin, A. E., Marciano, P. L., & Whitley, M. K. (2005). The therapeutic alliance in cognitive-behavioral treatment of children referred for oppositional, aggressive, and antisocial behavior. Journal of Consulting and Clinical Psychology, 73, 726-730.
34 Kazdin, A. E., & Mazurick, J. L. (19 94). Dropping out of child psychotherapy: Distinguishing early and late dropout s over the course of treatment. Journal of Consulting & Clinical Psychology, 62, 1069-1074. Kazdin, A. E., Mazurick, J. L., & Bass, D. (1993). Risk for attrit ion in treatment of antisocial children and families. Journal of Clinical Child Psychology, 22, 2-16. Kazdin, A. E., Stolar, M. J., & Marciano, P. L. (1995). Risk factors for dropping out of treatment among White and Black families. Journal of Family Psychology, 9, 402417. Kendall, P. C., & Sugarman, A. (1997). Attr ition in the treatmen t of childhood anxiety disorders. Journal of Consulting and Clinical Psychology, 65, 883-888. Koot, H. M., Van Den Oord, E. J., Verhulst, F. C., & Boomsma, D. I. (1997). Behavioral and emotional problems in young preschool ers: Cross-cultura l testing of the validity of the Child Behavior Checklist/2-3. Journal of Abnormal Child Psychology, 25, 183-196. Lahey, B. B., Loeber, R., Hart, E. L., & Fric k, P. J. (1995). Four-year longitudinal study of conduct disorder in boys: Patterns and predictors of persistence. Journal of Abnormal Psychology, 104, 83-93. Loeber, R., Burke, J. D., Lahey, B. B., Winters, A., & Zera, M. (2000). Oppositional defiant and conduct disorder: A review of the past 10 years, part I. Journal of the American Academy of Child & Adolescent Psychiatry, 39 1468-1484. Luk, E. S., Staiger, P. K., Mathai, J., Wong, L., Birleson, P., & Adler, R. (2001). Children with persistent conduct problem s who dropout of treatment. European Child and Adolescent Psychiatry, 10, 28-36. McCabe, K. M. (2002). Factors that predict premature termination among MexicanAmerican children in outpatient psychotherapy. Journal of Child and Family Studies, 11, 347-359. Nixon, R. D., Sweeney, L., Erickson, D. B., & Touyz, S. W. (2003). Parent-child interaction therapy: A comparison of st andard and abbreviated treatments for oppositional defiant preschoolers. Journal of Consulting an d Clinical Psychology, 71, 251-260. Pekarik, G., & Stephenson, L. A. (1988). Adult and child differences in therapy dropout research. Journal of Clinical Child Psychology, 17, 316-321. Peters, S., Calam, R., & Harrington, R. (2005) Maternal attributions and expressed emotion as predictors of attendance at parent management training. Journal of Child Psychology and Psychiatry, 46, 436-48.
35 Prinz, R. J., & Miller, G. E. (1994). Fa mily-based treatment for childhood antisocial behavior: Experimental influe nces on dropout and engagement. Journal of Consulting and Clinical Psychology, 62, 645-650. Reid, M. J., Webster-Stratton, C., & Ha mmond, M. (2003). Follow-up of children who received the Incredible Years interv ention for Oppositional Defiant Disorder: Maintenance and prediction of 2-year outcome. Behavior Therapy, 34, 471-491. Refinetti, R. (1996). Demonstrating the conseq uences of violations of assumptions in between-subjects analysis of variance. Teaching of Psychology, 23, 51-54. Ruma, P. R., & Burke, R. V. (1996). Group pare nt training: Is it eff ective for children of all ages? Behavior Therapy, 27, 159-169. Schuhmann, E. M., Foote, R. C., Eyberg, S. M., Boggs, S. R., & Al gina, J. A. (1998). Efficacy of parent-child in teraction therapy: Interim report of a randomized trial with short-term maintenance. Journal of Clinical Child Psychology, 27, 34-45. Shaffer, D., Fisher, P., Lucas, C.P., Dulca n, M.K., & Schwab-Stone, M.E. (2000). NIMH Diagnostic Interview Schedule for Ch ildren Version IV (NIMH DISC-IV): Description, differences from previous versions, and reliability of some common diagnoses. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 28-38. Vander Stoep, A. (1999). Maintaining high subject retention in follow-up studies of children with mental illness. Journal of Child and Family Studies, 8, 305-318. Webster-Stratton, C. (1985). The effects of father involvement in parent training for conduct problem children. Journal of Child Psyc hology and Psychiatry, 26, 801810. Webster-Stratton, C. (1990). Long-term follow-up of families with young conduct problem children: From preschool to grade school. Journal of Clinical Child Psychology, 19, 144-149. Weisz, J. R., Weiss, B., & Langmeyer, D. B. (1987). Giving up on child psychotherapy: Who drops out? Journal of Consulting & Clinical Psychology, 55, 916-918. Werba, B. E., Eyberg, S. E., Boggs, S. R., & Algina, J. A. (in press). Predicting outcome in ParentChild Interaction Therapy: Success and attrition. Behavior Modification Werry, J. S. (1997). Severe conduct disorder: Some key issues. Canadian Journal of Psychiatry, 42, 577-583. Wierzbicki M., & Pekarik G. (1993). A meta-analysi s of psychotherapy dropout. Professional Psychology: Research & Practice 24, 190-195.
36 Zimmerman, D. W. (2004). A note on preliminary tests of equality of variances. British Journal of Mathematical and Statistical Psychology, 57, 173-181.
BIOGRAPHICAL SKETCH Melanie Fernandez was born in Tampa, Flor ida, to Rafael and Rebeca Fernandez. She has one sister, Laura, 3 y ears her senior. Melanie earned her Bachelor of Arts degree (with honors) in psychology, at Brown Univer sity in 1998. She entered graduate school in 2002 in the Department of Clinical and Heal th Psychology at the Un iversity of Florida and will complete her predoctoral internship training at Duke University Medical Center. 37