This article was downloaded by: [Van Pelt and Opie Library] On: 20 October 2014, At: 20:15 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Child & Adolescent Mental Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/rcmh20
Risk factors for premature termination of treatment at a child and family mental health clinic a
a
Laura Taylor , Debra Kaminer & Anneli Hardy
b
a
Department of Psychology , University of Cape Town , Rondebosch, 7700, South Africa b
Statistical Consulting Service, Department of Statistical Sciences , University of Cape Town Published online: 19 Dec 2011.
To cite this article: Laura Taylor , Debra Kaminer & Anneli Hardy (2011) Risk factors for premature termination of treatment at a child and family mental health clinic, Journal of Child & Adolescent Mental Health, 23:2, 155-164, DOI: 10.2989/17280583.2011.634553 To link to this article: http://dx.doi.org/10.2989/17280583.2011.634553
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
& Conditions of access and use can be found at http://www.tandfonline.com/page/ terms-and-conditions
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164 Printed in South Africa — All rights reserved
Copyright © NISC Pty Ltd
JOURNAL OF CHILD AND ADOLESCENT MENTAL HEALTH ISSN 1728–0583 EISSN 1728–0591 DOI: 10.2989/17280583.2011.634553
Research Paper Risk factors for premature termination of treatment at a child and family mental health clinic
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
Laura Taylor1, Debra Kaminer1* and Anneli Hardy2 Department of Psychology, University of Cape Town, Rondebosch 7700, South Africa Statistical Consulting Service, Department of Statistical Sciences, University of Cape Town *Corresponding author, email: [email protected]
1
2
Objective: Premature termination of treatment amongst children and families attending mental health services is a significant problem for both outcomes research and clinical practice in South Africa and elsewhere. This study investigated factors that are associated with premature termination of treatment at a public service child and family clinic in Cape Town. Method: A retrospective archival analysis of clinic files from 2002–2009 was conducted. Administrative, child and family factors, and type of treatment were explored as risk factors for premature termination. Results: A single-parent household and the presence of a child or oppositional defiant disorder were risk factors for premature termination of treatment, while the presence of a maternal psychiatric diagnosis was associated with a lower likelihood of terminating treatment prematurely. Conducting a scholastic assessment with the child was associated with a lower risk of premature termination, while there was a trend towards a higher risk of premature termination when individual child therapy was the recommended treatment. Conclusions: This study offers recommendations for how these findings could assist South African clinicians to enhance client retention in child and family mental health services, and suggestions for future research.
Introduction Although attrition of children and families from mental health services has received little attention relative to attrition among adult clients, existing studies report high rates of drop out for children and their families across different stages of treatment (Staudt 2003). For example, McKay, McAdam and Gonzales (1996) report that 39% of the clients attending a child and family clinic failed to attend the first scheduled appointment; dropout rates of 45% (Armbruster and Fallon 1994) and 49% (Johnson, Mellor and Brann 2008) have been found amongst all clients attending treatment at child and family clinics over a one-year period; and Kazdin and Mazurick (1994) report that nearly 48% of child and family cases terminated treatment prematurely after psychotherapy had commenced. Dropping out of treatment presents a significant obstacle to treatment outcome research, as the loss of cases during the course of treatment, and resulting changes in group composition, threaten all aspects of validity through the introduction of potential sampling bias, statistical power reductions and limits to the generalisability of research results (Kazdin, Holland and Cowley 1997). In terms of clinical practice, those who terminate treatment prematurely are less likely to benefit from treatment than those who do not (Reitzel et al. 2006). Finally, premature termination of treatment affects the efficient delivery of clinical services, as frequent cancellations or no-shows increase service delivery costs and reduce the opportunity for waiting list clients to receive treatment (Kazdin et al. 1997). In Journal of Child & Adolescent Mental Health is co-published by NISC (Pty) Ltd and Routledge, Taylor & Francis Group
156
Taylor, Kaminer and Hardy
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
light of these barriers to effective treatment practice, service delivery and research, research must identify factors that predict premature termination of treatment in child and family mental health clinics (Johnson et al. 2008). Factors associated with premature termination of treatment in child and family mental health clinics Some research on predictors of premature termination of treatment amongst children and families attending mental health services has focused on samples that have completed the intake or assessment stage and commenced psychotherapy (Kazdin, Mazurick and Bass 1993, Kazdin and Mazurick 1994); other studies have explored dropout across different stages of clinical contact. Although research findings are somewhat inconsistent, possibly due to variations in sampling, methodology and definitions of “dropout” or “premature termination” (Staudt 2003), administrative, child and family factors have all been identified as predictors of premature termination of child and family mental health treatment. Administrative factors Some studies have found that children and families who are waitlisted for an extended period before their first appointment tend to be more at risk for pre-intake attrition (Sherman et al. 2008). Length of time on the waiting list does not appear to be predictive of treatment dropout amongst adult clients once treatment has started (Reitzel et al. 2006), however, this relationship has not been explored in child and family client populations. Another administrative factor that may be associated with premature termination of child and family treatment is referral source. Reis and Brown (1999) suggest that clients who are involuntarily referred for treatment, such as through the court or schools, are more likely candidates for premature termination than those who seek treatment voluntarily. Children and families may therefore be more likely to remain in treatment if parents have sought the treatment themselves, as opposed to being involuntarily referred by institutions. However, a study by Armbruster and Fallon (1994) found that coercive referral sources such as schools, courts, and social services agencies were not associated with premature termination of treatment. Family factors Family factors appear to consistently account for the greatest amount of variance in premature termination of child and family mental health treatment. Family socio-economic status, family structure and parental mental health history have frequently been found to be predictors of premature termination. Children from socio-economically disadvantaged families, and whose parents have a low level of education or are unemployed, have been found to be more likely premature termination candidates than children from high socio-economic status families (Armbruster and Fallon 1994, Kazdin et al. 1997). It is possible that for families with fewer resources, receiving therapeutic treatment might be accorded less importance than meeting more basic survival needs. For example, families from groups with lower socio-economic status may not have the financial resources for mental health services when they can ill afford other necessities, may be less able to negotiate time off work to attend sessions, and may struggle to pay for transport to and from a clinic. The children of parents who have little social and emotional support from intimate partners or other family members may be more likely premature termination candidates. For example, children who have young mothers, and children who come from single parented or non-biologically headed homes, are at greater risk for premature termination (Kazdin and Mazurick 1994, Kazdin et al. 1997), as are children whose families have low levels of cohesion (Armbruster and Fallon 1994). Both maternal and paternal anti-social behaviours have been found to be predictors of premature termination of treatment at child and family clinics (Armbruster and Fallon 1994, Kazdin and Mazurick 1994, Kazdin et al. 1997). However, the influence of other forms of parental psychopathology on premature termination of child and family therapy is unclear.
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164
157
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
The risk created by family variables appears to be cumulative, with premature termination of individual child therapy more likely to occur among families headed by a single parent who is young and of a lower socio-economic status (Campbell, Baker and Bratton 2000). Diagnostic factors The child’s diagnosis has also been reported to be a predictor of premature termination. Kazdin et al. (1997) found that children presenting with conduct problems are more likely premature termination candidates. Johnson et al. (2008) found that attrition was more common amongst children with diagnoses of conduct disorders or attention deficit/hyperactivity disorder (ADHD) and less common among those with diagnoses of anxiety disorders or with no diagnosis. The Johnson et al. (2008) study was one of the first to concentrate solely on child diagnosis as a predictor of premature termination of child treatment, and the authors note that further research is needed on how and why certain diagnoses are associated with a greater prevalence of premature termination. Type of treatment Child and family mental health clinics in South Africa and elsewhere tend to offer a range of possible treatment options, including scholastic assessments, parent counselling, family therapy, individual child therapy or some combination of these. However, the relationship between type of treatment and premature termination has seldom been explored in child and family mental health settings. It is possible that treatments involving parents or families may have a higher risk of premature termination than individual child therapy, due either to logistical barriers such as taking time off work, transport costs and lack of child care arrangements or to psychological resistance on the part of family members. However, it is also possible that treatments that focus only on the child and do not provide parallel therapeutic holding for family members may be more at risk of premature termination. Aims of this study While international research has identified multiple administrative, family and child diagnostic predictors of premature termination of treatment in child and family mental health service settings, there is a lack of South African research addressing this topic. Since the South African context may be characterised by specific and unique challenges with regard to patient retention in mental health settings, local research identifying factors that are associated with premature termination of treatment at child and family mental health clinics is warranted. Such research will enable clinicians to identify families at high risk for terminating treatment prematurely, and to put measures in place to enhance retention. This study therefore examined administrative, family, child and type of treatment factors that may be associated with premature termination of treatment at a child and family clinic in South Africa. Method Study design and setting A retrospective archival analysis of clinic files was conducted at the University of Cape Town (UCT) Child Guidance Clinic (CGC). The CGC is a postgraduate teaching centre for training clinical psychology masters students. It is also an outpatient treatment clinic for children and families who are unable to afford private clinical services. The therapeutic process at the CGC begins when a parent calls to make an appointment, identifying in the process which child they may be concerned about. When an appointment is available, the whole family (or as many members of the household as possible) is seen by a student psychologist for one or two intake sessions, and the child is also interviewed alone where appropriate. Thereafter, the student psychologist determines what further information needs to be gathered, for instance further history-taking meetings with one or both parents, collateral from the child’s teacher, and scholastic or other forms of assessment with the child. After the assessment process is completed, feedback is provided to the family and a treatment plan is recommended. This treatment
158
Taylor, Kaminer and Hardy
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
plan may entail parent counselling with one or both parents, family therapy, individual child psychotherapy, referral of the child or other family members to other practitioners or a combination of these. There is no limit to the number of sessions that may be offered, and treatment contracts are negotiated with families on a case-by-case basis. Sessions are usually offered once a week but this may vary depending on the family’s circumstances. Sample The sampling frame included all cases over the 8-year period from 2002–2009 where the identified client was a child (that is, younger than 18 years of age). The clinic also sees a small number of adult patients, however, these were not included in the study. Those case files in which several of the variables under study were not recorded (n = 61) were excluded from the sample while case files which contained only one or two missing variables were retained. The final sample comprised 332 case files. Data collection The definition of premature termination used in this study was based on that used by Johnson et al. (2008:520), namely “the unilateral decision of the client to terminate treatment, either explicitly, through stating their desire to discontinue and not making any more appointments, or implicitly, by failing to attend sessions even when followed up. Completion of treatment was defined as mutually agreed discontinuation of treatment”. To determine whether or not each case met the definition for premature termination, the termination/discharge summary and clinical notes on the last contact with the client were examined. The independent variables selected for inclusion were based on current premature termination literature as well as the anecdotal clinical experience of the CGC staff, and were divided into four categories: administrative factors, family factors, child factors and type of treatment. Administrative factors The administrative variables of “days on waitlist” and whether the child was “school referred” were included, because they have been reported as possible predictors of premature termination in the literature (Reitzel et al. 2006, Sherman et al. 2008). The continuous variable of “days on waitlist” was calculated by looking at the date of the first phone call on the referral card and the date of the first intake session on the file contact sheet. Information about whether or not the child was referred by his or her school was obtained from the referral card. Family factors Socio-economic status. Although families attending the CGC do report their monthly income on their application form, in order to establish which fee they should pay on a sliding scale, the reliability of this report may be limited since there is some incentive for families to under-report their income. Consequently, in line with previous studies (Armbruster and Fallon 2004, Sherman et al. 2008), “maternal and paternal education level” (whether the parent completed high school) and “current employment status” were selected as indicators of family socio-economic status. This information was found in the clinical history completed by the clinician. Family structure Four independent variables were used to reflect family structure: the presence or absence of a “single parent home”, the presence or absence of a “multiple fathered home” (that is, the children living in the home have the same mother but different fathers), the “number of siblings” of the child client, and the “total number of people living in the household”. This information was also found in the clinical history. Family history “Maternal psychiatric diagnosis” (whether the child’s mother has a past or current psychiatric diagnosis) and “maternal trauma history” (whether the child’s mother has experienced a traumatic
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164
159
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
event) were selected as indicators of maternal psychological history. This information was again found in the clinical history. Paternal psychiatric and trauma histories could not be included due to the high number of single parent households in the sample (n = 146; 44.5%) and the consequent amount of missing data about fathers in the case files. Child factors The child’s “gender” and their “age at the time of referral” were examined as independent variables. Since the child’s diagnosis has emerged as an important predictor of premature termination in previous studies (Kazdin et al. 1997, Johnson et al. 2008), the presence or absence of a diagnosis of: 1) ADHD/attention deficit disorder (ADD); 2) depression; 3) conduct disorder/oppositional defiant disorder; and 4) anxiety disorder were also included as independent variables. These diagnoses were selected on the basis of the premature termination literature and through an examination of common presenting problems in the case files at the CGC. Diagnostic information was taken from the DSM-IV-TR (APA 2000) multi-axial diagnosis provided by the clinician. Type of treatment Cases were coded for whether or not a scholastic assessment with the child had been conducted. For cases where some form of psychotherapy was recommended, treatment was categorised as: 1) individual child therapy; 2) parent or family counselling; or 3) a combination of 1) and 2). Decisions about treatment categorisation were based on the management plan and clinical session notes in each case file. Scholastic assessment and psychotherapy were not mutually exclusive, since intervention at the CGC may include both assessment and psychotherapy. Procedure The files were examined manually by two researchers, and were coded according to a coding schedule. Before the coding started, both researchers coded several files together to ensure consistency. During the coding process there was ongoing communication between the two researchers to ensure continued consistency of coding. Data analysis The statistical package PASW Statistics (SPSS) version 18.0, with alpha set at 0.05, was used to analyse the data (PASW 2010). Descriptive data were first examined, making use of frequencies and percentages to describe the categorical variables, and means, standard deviations and ranges to describe the continuous variables. Following this, Fisher’s exact test and Pearson’s chi-square test of independence were performed to examine associations between the categorical independent variables and premature termination. The phi and Cramer’s V measures of the strength of association and the odds ratio were also calculated to evaluate the size of effect. Independent sample two-tailed t-tests were used to establish whether or not the mean scores of the continuous independent variables differed significantly across the premature termination and non-premature termination groups. Ethical considerations Before the first appointment at the CGC, parents or legal guardians are asked to sign a consent form giving UCT permission to use all case information, except names or other identifying data, for research. This form had been signed in all 332 case files retained in the sample. Because of the archival nature of the data, this study presented no obvious risks or benefits to participants. Results Sample characteristics Tables 1 and 2 summarise the characteristics of families presenting for treatment at the CGC over the period 2002–2009. Frequencies and percentages for the categorical independent variables are
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
160
Taylor, Kaminer and Hardy
presented in Table 1, and means, standard deviations and ranges for the continuous independent variables are reported in Table 2. With regard to referral patterns, a minority of the sample (n = 95; 28.7%) was referred by the child’s school, but a scholastic assessment was conducted for almost two thirds of the sample (n = 182; 60%) on the basis of the presenting problems described by the family, the school or both. More boys were seen at the clinic than girls (n = 201; 60.5% vs. n = 131; 39.5%) in this period, and the average age at referral was 9.5 years. The average time spent on the waiting list was just under four months, however, the standard deviation indicates a large amount of variability in the waitlist variable. With regard to socio-economic indicators, only about half the mothers (n = 173; 57.1%) and fathers (n = 124; 46.4%) had completed high school, but the majority of both mothers (n = 236; 74.9%) and fathers (n = 227; 83.5%) were currently employed. A substantial portion (n = 146; 44.5%) of families attending the clinic lived in single-parent households and almost 20% (n = 65) lived in multiple-fathered households. The average household size was about four family members, and the average number of siblings was approximately one, suggesting that overcrowding and large family size are not common features of the households of families presenting to this clinic. With regard to maternal psychological history, a minority of mothers (n = 51; 16%) had a past or current psychiatric diagnosis while a greater proportion (n = 92; 29%) had experienced a traumatic event. Only a minority of children received a diagnosis of ADHD/ADD (n = 49; 14.8%), depression
Table 1: Characteristics of families presenting for treatment: frequencies and percentages for categorical variables Variable School referred Scholastic assessment Mother completed high school Father completed high school Mother employed Father employed Single-parent household Multiple-fathered household Presence of past or current maternal psychiatric diagnosis Presence of maternal trauma exposure Child’s gender Male Female ADHD/ADD diagnosis present Depression diagnosis present Conduct or oppositional defiant disorder present Anxiety disorder present Type of psychotherapy recommended Parent or family counseling Individual child therapy Combination
Frequency (Valid total) 95 (331) 182 (303) 173 (303) 124 (267) 236 (315) 227 (272) 146 (328) 65 (329) 51 (316) 92 (315)
Percentage (%) 28.7 60.1 57.1 46.4 74.9 83.5 44.5 19.8 16.1 29.2
201 131 49 (331) 34 (331) 28 (331) 34 (331)
60.5 39.5 14.8 10.3 8.5 10.3
80 (209) 34 (209) 95 (209)
38.3 16.3 45.5
Table 2: Characteristics of families presenting for treatment: means, standard deviations and ranges for continuous variables Variable Days on waitlist Number of people living at home Number of siblings Child’s age at referral
N 288 329 332 332
Mean 119.09 4.43 1.14 9.58
SD 110.37 1.72 1.05 3.29
Minimum 2 2 0 3
Maximum 720 12 7 17
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164
161
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
(n = 34; 10.3%), conduct or oppositional defiant disorder (n = 28; 8.5%) or an anxiety disorder (n = 34; 10.3%). Where psychotherapy was recommended, the largest proportion of cases received a recommendation of a combination of individual child therapy and parent or family counselling (n = 95; 45.5%). In 80 (38.3%) cases only parent or family counselling was recommended, and in 34 (16.3%) cases only individual child therapy was recommended. Rates of premature termination The overall rate of premature termination in the sample was 27.1% (n = 90), which represents almost a third of all cases included in this study. A small number of cases (n = 23 (305); 7.5%) unilaterally withdrew from treatment immediately after the intake interview. In those cases where a scholastic assessment with the child was recommended (n = 182 (303); 54.8%), 17% (n = 31) terminated prematurely (that is, the assessment process was not completed, or the family did not return for feedback and recommendations after the assessment was completed). Of those cases for which some form of psychotherapy was recommended (n = 209 (305); 63%), 26.79% (n = 56) terminated the therapy prematurely. Premature termination rates were highest in cases where individual child therapy was the recommended treatment (n = 14; 41.2%), lower in cases where parent or family counselling was recommended (n = 22; 27.5%), and lowest in cases where a combination of individual child therapy and parent or family counselling was recommended (n = 20; 21.1%). Association between independent variables and premature termination Fisher’s exact test was performed to examine the associations between premature termination and each of the categorical independent variables. The p values yielded by Fisher’s exact test are summarised in Table 3. Results indicate that living in a single-parent household, and a child diagnosis of conduct or oppositional defiant disorder, were associated with an increased likelihood of premature termination of therapy. However, the presence of a maternal psychiatric diagnosis was associated with a decreased likelihood of terminating therapy prematurely. A larger proportion of children from single-parent homes terminated therapy prematurely than children from homes with both parents (33.6% versus 22%; Fisher’s Exact test, p = 0.024), although the effect size (ϕ = 0.129) was small. Based on the odds ratio, within this sample a child from a single-parent home was 1.8 times more likely to prematurely terminate therapy than children from homes with both parents. Table 3: Association between categorical independent variables and premature termination Variable School referred Scholastic assessment conducted Mother completed high school Father completed high school Mother employed Father employed Single-parent household Multiple-fathered household Presence of past or current maternal psychiatric diagnosis Presence of maternal trauma exposure Child’s gender ADHD/ADD diagnosis present Depression diagnosis present Conduct or oppositional defiant disorder present Anxiety disorder present * p value from Fisher’s Exact test
p* 0.49
Journal of Child & Adolescent Mental Health Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/rcmh20
Risk factors for premature termination of treatment at a child and family mental health clinic a
a
Laura Taylor , Debra Kaminer & Anneli Hardy
b
a
Department of Psychology , University of Cape Town , Rondebosch, 7700, South Africa b
Statistical Consulting Service, Department of Statistical Sciences , University of Cape Town Published online: 19 Dec 2011.
To cite this article: Laura Taylor , Debra Kaminer & Anneli Hardy (2011) Risk factors for premature termination of treatment at a child and family mental health clinic, Journal of Child & Adolescent Mental Health, 23:2, 155-164, DOI: 10.2989/17280583.2011.634553 To link to this article: http://dx.doi.org/10.2989/17280583.2011.634553
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
& Conditions of access and use can be found at http://www.tandfonline.com/page/ terms-and-conditions
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164 Printed in South Africa — All rights reserved
Copyright © NISC Pty Ltd
JOURNAL OF CHILD AND ADOLESCENT MENTAL HEALTH ISSN 1728–0583 EISSN 1728–0591 DOI: 10.2989/17280583.2011.634553
Research Paper Risk factors for premature termination of treatment at a child and family mental health clinic
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
Laura Taylor1, Debra Kaminer1* and Anneli Hardy2 Department of Psychology, University of Cape Town, Rondebosch 7700, South Africa Statistical Consulting Service, Department of Statistical Sciences, University of Cape Town *Corresponding author, email: [email protected]
1
2
Objective: Premature termination of treatment amongst children and families attending mental health services is a significant problem for both outcomes research and clinical practice in South Africa and elsewhere. This study investigated factors that are associated with premature termination of treatment at a public service child and family clinic in Cape Town. Method: A retrospective archival analysis of clinic files from 2002–2009 was conducted. Administrative, child and family factors, and type of treatment were explored as risk factors for premature termination. Results: A single-parent household and the presence of a child or oppositional defiant disorder were risk factors for premature termination of treatment, while the presence of a maternal psychiatric diagnosis was associated with a lower likelihood of terminating treatment prematurely. Conducting a scholastic assessment with the child was associated with a lower risk of premature termination, while there was a trend towards a higher risk of premature termination when individual child therapy was the recommended treatment. Conclusions: This study offers recommendations for how these findings could assist South African clinicians to enhance client retention in child and family mental health services, and suggestions for future research.
Introduction Although attrition of children and families from mental health services has received little attention relative to attrition among adult clients, existing studies report high rates of drop out for children and their families across different stages of treatment (Staudt 2003). For example, McKay, McAdam and Gonzales (1996) report that 39% of the clients attending a child and family clinic failed to attend the first scheduled appointment; dropout rates of 45% (Armbruster and Fallon 1994) and 49% (Johnson, Mellor and Brann 2008) have been found amongst all clients attending treatment at child and family clinics over a one-year period; and Kazdin and Mazurick (1994) report that nearly 48% of child and family cases terminated treatment prematurely after psychotherapy had commenced. Dropping out of treatment presents a significant obstacle to treatment outcome research, as the loss of cases during the course of treatment, and resulting changes in group composition, threaten all aspects of validity through the introduction of potential sampling bias, statistical power reductions and limits to the generalisability of research results (Kazdin, Holland and Cowley 1997). In terms of clinical practice, those who terminate treatment prematurely are less likely to benefit from treatment than those who do not (Reitzel et al. 2006). Finally, premature termination of treatment affects the efficient delivery of clinical services, as frequent cancellations or no-shows increase service delivery costs and reduce the opportunity for waiting list clients to receive treatment (Kazdin et al. 1997). In Journal of Child & Adolescent Mental Health is co-published by NISC (Pty) Ltd and Routledge, Taylor & Francis Group
156
Taylor, Kaminer and Hardy
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
light of these barriers to effective treatment practice, service delivery and research, research must identify factors that predict premature termination of treatment in child and family mental health clinics (Johnson et al. 2008). Factors associated with premature termination of treatment in child and family mental health clinics Some research on predictors of premature termination of treatment amongst children and families attending mental health services has focused on samples that have completed the intake or assessment stage and commenced psychotherapy (Kazdin, Mazurick and Bass 1993, Kazdin and Mazurick 1994); other studies have explored dropout across different stages of clinical contact. Although research findings are somewhat inconsistent, possibly due to variations in sampling, methodology and definitions of “dropout” or “premature termination” (Staudt 2003), administrative, child and family factors have all been identified as predictors of premature termination of child and family mental health treatment. Administrative factors Some studies have found that children and families who are waitlisted for an extended period before their first appointment tend to be more at risk for pre-intake attrition (Sherman et al. 2008). Length of time on the waiting list does not appear to be predictive of treatment dropout amongst adult clients once treatment has started (Reitzel et al. 2006), however, this relationship has not been explored in child and family client populations. Another administrative factor that may be associated with premature termination of child and family treatment is referral source. Reis and Brown (1999) suggest that clients who are involuntarily referred for treatment, such as through the court or schools, are more likely candidates for premature termination than those who seek treatment voluntarily. Children and families may therefore be more likely to remain in treatment if parents have sought the treatment themselves, as opposed to being involuntarily referred by institutions. However, a study by Armbruster and Fallon (1994) found that coercive referral sources such as schools, courts, and social services agencies were not associated with premature termination of treatment. Family factors Family factors appear to consistently account for the greatest amount of variance in premature termination of child and family mental health treatment. Family socio-economic status, family structure and parental mental health history have frequently been found to be predictors of premature termination. Children from socio-economically disadvantaged families, and whose parents have a low level of education or are unemployed, have been found to be more likely premature termination candidates than children from high socio-economic status families (Armbruster and Fallon 1994, Kazdin et al. 1997). It is possible that for families with fewer resources, receiving therapeutic treatment might be accorded less importance than meeting more basic survival needs. For example, families from groups with lower socio-economic status may not have the financial resources for mental health services when they can ill afford other necessities, may be less able to negotiate time off work to attend sessions, and may struggle to pay for transport to and from a clinic. The children of parents who have little social and emotional support from intimate partners or other family members may be more likely premature termination candidates. For example, children who have young mothers, and children who come from single parented or non-biologically headed homes, are at greater risk for premature termination (Kazdin and Mazurick 1994, Kazdin et al. 1997), as are children whose families have low levels of cohesion (Armbruster and Fallon 1994). Both maternal and paternal anti-social behaviours have been found to be predictors of premature termination of treatment at child and family clinics (Armbruster and Fallon 1994, Kazdin and Mazurick 1994, Kazdin et al. 1997). However, the influence of other forms of parental psychopathology on premature termination of child and family therapy is unclear.
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164
157
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
The risk created by family variables appears to be cumulative, with premature termination of individual child therapy more likely to occur among families headed by a single parent who is young and of a lower socio-economic status (Campbell, Baker and Bratton 2000). Diagnostic factors The child’s diagnosis has also been reported to be a predictor of premature termination. Kazdin et al. (1997) found that children presenting with conduct problems are more likely premature termination candidates. Johnson et al. (2008) found that attrition was more common amongst children with diagnoses of conduct disorders or attention deficit/hyperactivity disorder (ADHD) and less common among those with diagnoses of anxiety disorders or with no diagnosis. The Johnson et al. (2008) study was one of the first to concentrate solely on child diagnosis as a predictor of premature termination of child treatment, and the authors note that further research is needed on how and why certain diagnoses are associated with a greater prevalence of premature termination. Type of treatment Child and family mental health clinics in South Africa and elsewhere tend to offer a range of possible treatment options, including scholastic assessments, parent counselling, family therapy, individual child therapy or some combination of these. However, the relationship between type of treatment and premature termination has seldom been explored in child and family mental health settings. It is possible that treatments involving parents or families may have a higher risk of premature termination than individual child therapy, due either to logistical barriers such as taking time off work, transport costs and lack of child care arrangements or to psychological resistance on the part of family members. However, it is also possible that treatments that focus only on the child and do not provide parallel therapeutic holding for family members may be more at risk of premature termination. Aims of this study While international research has identified multiple administrative, family and child diagnostic predictors of premature termination of treatment in child and family mental health service settings, there is a lack of South African research addressing this topic. Since the South African context may be characterised by specific and unique challenges with regard to patient retention in mental health settings, local research identifying factors that are associated with premature termination of treatment at child and family mental health clinics is warranted. Such research will enable clinicians to identify families at high risk for terminating treatment prematurely, and to put measures in place to enhance retention. This study therefore examined administrative, family, child and type of treatment factors that may be associated with premature termination of treatment at a child and family clinic in South Africa. Method Study design and setting A retrospective archival analysis of clinic files was conducted at the University of Cape Town (UCT) Child Guidance Clinic (CGC). The CGC is a postgraduate teaching centre for training clinical psychology masters students. It is also an outpatient treatment clinic for children and families who are unable to afford private clinical services. The therapeutic process at the CGC begins when a parent calls to make an appointment, identifying in the process which child they may be concerned about. When an appointment is available, the whole family (or as many members of the household as possible) is seen by a student psychologist for one or two intake sessions, and the child is also interviewed alone where appropriate. Thereafter, the student psychologist determines what further information needs to be gathered, for instance further history-taking meetings with one or both parents, collateral from the child’s teacher, and scholastic or other forms of assessment with the child. After the assessment process is completed, feedback is provided to the family and a treatment plan is recommended. This treatment
158
Taylor, Kaminer and Hardy
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
plan may entail parent counselling with one or both parents, family therapy, individual child psychotherapy, referral of the child or other family members to other practitioners or a combination of these. There is no limit to the number of sessions that may be offered, and treatment contracts are negotiated with families on a case-by-case basis. Sessions are usually offered once a week but this may vary depending on the family’s circumstances. Sample The sampling frame included all cases over the 8-year period from 2002–2009 where the identified client was a child (that is, younger than 18 years of age). The clinic also sees a small number of adult patients, however, these were not included in the study. Those case files in which several of the variables under study were not recorded (n = 61) were excluded from the sample while case files which contained only one or two missing variables were retained. The final sample comprised 332 case files. Data collection The definition of premature termination used in this study was based on that used by Johnson et al. (2008:520), namely “the unilateral decision of the client to terminate treatment, either explicitly, through stating their desire to discontinue and not making any more appointments, or implicitly, by failing to attend sessions even when followed up. Completion of treatment was defined as mutually agreed discontinuation of treatment”. To determine whether or not each case met the definition for premature termination, the termination/discharge summary and clinical notes on the last contact with the client were examined. The independent variables selected for inclusion were based on current premature termination literature as well as the anecdotal clinical experience of the CGC staff, and were divided into four categories: administrative factors, family factors, child factors and type of treatment. Administrative factors The administrative variables of “days on waitlist” and whether the child was “school referred” were included, because they have been reported as possible predictors of premature termination in the literature (Reitzel et al. 2006, Sherman et al. 2008). The continuous variable of “days on waitlist” was calculated by looking at the date of the first phone call on the referral card and the date of the first intake session on the file contact sheet. Information about whether or not the child was referred by his or her school was obtained from the referral card. Family factors Socio-economic status. Although families attending the CGC do report their monthly income on their application form, in order to establish which fee they should pay on a sliding scale, the reliability of this report may be limited since there is some incentive for families to under-report their income. Consequently, in line with previous studies (Armbruster and Fallon 2004, Sherman et al. 2008), “maternal and paternal education level” (whether the parent completed high school) and “current employment status” were selected as indicators of family socio-economic status. This information was found in the clinical history completed by the clinician. Family structure Four independent variables were used to reflect family structure: the presence or absence of a “single parent home”, the presence or absence of a “multiple fathered home” (that is, the children living in the home have the same mother but different fathers), the “number of siblings” of the child client, and the “total number of people living in the household”. This information was also found in the clinical history. Family history “Maternal psychiatric diagnosis” (whether the child’s mother has a past or current psychiatric diagnosis) and “maternal trauma history” (whether the child’s mother has experienced a traumatic
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164
159
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
event) were selected as indicators of maternal psychological history. This information was again found in the clinical history. Paternal psychiatric and trauma histories could not be included due to the high number of single parent households in the sample (n = 146; 44.5%) and the consequent amount of missing data about fathers in the case files. Child factors The child’s “gender” and their “age at the time of referral” were examined as independent variables. Since the child’s diagnosis has emerged as an important predictor of premature termination in previous studies (Kazdin et al. 1997, Johnson et al. 2008), the presence or absence of a diagnosis of: 1) ADHD/attention deficit disorder (ADD); 2) depression; 3) conduct disorder/oppositional defiant disorder; and 4) anxiety disorder were also included as independent variables. These diagnoses were selected on the basis of the premature termination literature and through an examination of common presenting problems in the case files at the CGC. Diagnostic information was taken from the DSM-IV-TR (APA 2000) multi-axial diagnosis provided by the clinician. Type of treatment Cases were coded for whether or not a scholastic assessment with the child had been conducted. For cases where some form of psychotherapy was recommended, treatment was categorised as: 1) individual child therapy; 2) parent or family counselling; or 3) a combination of 1) and 2). Decisions about treatment categorisation were based on the management plan and clinical session notes in each case file. Scholastic assessment and psychotherapy were not mutually exclusive, since intervention at the CGC may include both assessment and psychotherapy. Procedure The files were examined manually by two researchers, and were coded according to a coding schedule. Before the coding started, both researchers coded several files together to ensure consistency. During the coding process there was ongoing communication between the two researchers to ensure continued consistency of coding. Data analysis The statistical package PASW Statistics (SPSS) version 18.0, with alpha set at 0.05, was used to analyse the data (PASW 2010). Descriptive data were first examined, making use of frequencies and percentages to describe the categorical variables, and means, standard deviations and ranges to describe the continuous variables. Following this, Fisher’s exact test and Pearson’s chi-square test of independence were performed to examine associations between the categorical independent variables and premature termination. The phi and Cramer’s V measures of the strength of association and the odds ratio were also calculated to evaluate the size of effect. Independent sample two-tailed t-tests were used to establish whether or not the mean scores of the continuous independent variables differed significantly across the premature termination and non-premature termination groups. Ethical considerations Before the first appointment at the CGC, parents or legal guardians are asked to sign a consent form giving UCT permission to use all case information, except names or other identifying data, for research. This form had been signed in all 332 case files retained in the sample. Because of the archival nature of the data, this study presented no obvious risks or benefits to participants. Results Sample characteristics Tables 1 and 2 summarise the characteristics of families presenting for treatment at the CGC over the period 2002–2009. Frequencies and percentages for the categorical independent variables are
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
160
Taylor, Kaminer and Hardy
presented in Table 1, and means, standard deviations and ranges for the continuous independent variables are reported in Table 2. With regard to referral patterns, a minority of the sample (n = 95; 28.7%) was referred by the child’s school, but a scholastic assessment was conducted for almost two thirds of the sample (n = 182; 60%) on the basis of the presenting problems described by the family, the school or both. More boys were seen at the clinic than girls (n = 201; 60.5% vs. n = 131; 39.5%) in this period, and the average age at referral was 9.5 years. The average time spent on the waiting list was just under four months, however, the standard deviation indicates a large amount of variability in the waitlist variable. With regard to socio-economic indicators, only about half the mothers (n = 173; 57.1%) and fathers (n = 124; 46.4%) had completed high school, but the majority of both mothers (n = 236; 74.9%) and fathers (n = 227; 83.5%) were currently employed. A substantial portion (n = 146; 44.5%) of families attending the clinic lived in single-parent households and almost 20% (n = 65) lived in multiple-fathered households. The average household size was about four family members, and the average number of siblings was approximately one, suggesting that overcrowding and large family size are not common features of the households of families presenting to this clinic. With regard to maternal psychological history, a minority of mothers (n = 51; 16%) had a past or current psychiatric diagnosis while a greater proportion (n = 92; 29%) had experienced a traumatic event. Only a minority of children received a diagnosis of ADHD/ADD (n = 49; 14.8%), depression
Table 1: Characteristics of families presenting for treatment: frequencies and percentages for categorical variables Variable School referred Scholastic assessment Mother completed high school Father completed high school Mother employed Father employed Single-parent household Multiple-fathered household Presence of past or current maternal psychiatric diagnosis Presence of maternal trauma exposure Child’s gender Male Female ADHD/ADD diagnosis present Depression diagnosis present Conduct or oppositional defiant disorder present Anxiety disorder present Type of psychotherapy recommended Parent or family counseling Individual child therapy Combination
Frequency (Valid total) 95 (331) 182 (303) 173 (303) 124 (267) 236 (315) 227 (272) 146 (328) 65 (329) 51 (316) 92 (315)
Percentage (%) 28.7 60.1 57.1 46.4 74.9 83.5 44.5 19.8 16.1 29.2
201 131 49 (331) 34 (331) 28 (331) 34 (331)
60.5 39.5 14.8 10.3 8.5 10.3
80 (209) 34 (209) 95 (209)
38.3 16.3 45.5
Table 2: Characteristics of families presenting for treatment: means, standard deviations and ranges for continuous variables Variable Days on waitlist Number of people living at home Number of siblings Child’s age at referral
N 288 329 332 332
Mean 119.09 4.43 1.14 9.58
SD 110.37 1.72 1.05 3.29
Minimum 2 2 0 3
Maximum 720 12 7 17
Journal of Child and Adolescent Mental Health 2011, 23(2): 155–164
161
Downloaded by [Van Pelt and Opie Library] at 20:15 20 October 2014
(n = 34; 10.3%), conduct or oppositional defiant disorder (n = 28; 8.5%) or an anxiety disorder (n = 34; 10.3%). Where psychotherapy was recommended, the largest proportion of cases received a recommendation of a combination of individual child therapy and parent or family counselling (n = 95; 45.5%). In 80 (38.3%) cases only parent or family counselling was recommended, and in 34 (16.3%) cases only individual child therapy was recommended. Rates of premature termination The overall rate of premature termination in the sample was 27.1% (n = 90), which represents almost a third of all cases included in this study. A small number of cases (n = 23 (305); 7.5%) unilaterally withdrew from treatment immediately after the intake interview. In those cases where a scholastic assessment with the child was recommended (n = 182 (303); 54.8%), 17% (n = 31) terminated prematurely (that is, the assessment process was not completed, or the family did not return for feedback and recommendations after the assessment was completed). Of those cases for which some form of psychotherapy was recommended (n = 209 (305); 63%), 26.79% (n = 56) terminated the therapy prematurely. Premature termination rates were highest in cases where individual child therapy was the recommended treatment (n = 14; 41.2%), lower in cases where parent or family counselling was recommended (n = 22; 27.5%), and lowest in cases where a combination of individual child therapy and parent or family counselling was recommended (n = 20; 21.1%). Association between independent variables and premature termination Fisher’s exact test was performed to examine the associations between premature termination and each of the categorical independent variables. The p values yielded by Fisher’s exact test are summarised in Table 3. Results indicate that living in a single-parent household, and a child diagnosis of conduct or oppositional defiant disorder, were associated with an increased likelihood of premature termination of therapy. However, the presence of a maternal psychiatric diagnosis was associated with a decreased likelihood of terminating therapy prematurely. A larger proportion of children from single-parent homes terminated therapy prematurely than children from homes with both parents (33.6% versus 22%; Fisher’s Exact test, p = 0.024), although the effect size (ϕ = 0.129) was small. Based on the odds ratio, within this sample a child from a single-parent home was 1.8 times more likely to prematurely terminate therapy than children from homes with both parents. Table 3: Association between categorical independent variables and premature termination Variable School referred Scholastic assessment conducted Mother completed high school Father completed high school Mother employed Father employed Single-parent household Multiple-fathered household Presence of past or current maternal psychiatric diagnosis Presence of maternal trauma exposure Child’s gender ADHD/ADD diagnosis present Depression diagnosis present Conduct or oppositional defiant disorder present Anxiety disorder present * p value from Fisher’s Exact test
p* 0.49
