Predictors of treatment outcome in a child and adolescent psychiatry clinic: A naturalistic exploration

https://doi.org/10.1016/j.childyouth.2011.10.004Get rights and content

Abstract

We present data on predictors of treatment outcome for 3200 consecutive referrals to a child and adolescent psychiatry clinic. Using Reliable Change Index (RCI) methodology, we divided children into those who, between intake and discharge, improved, stayed the same, or got worse according to clinician-rated impairment. Most predictors of improvement were related to parent variables (marital status, maternal anxiety, and ethnicity), while those associated with deterioration were tied to child status (extent of psychiatric comorbidity, history of placement in a self-contained classroom, and a prior trial of psychotropic medications). The implications of these findings for data-driven program development, clinic management, treatment planning, and systems of care are discussed.

Highlights

► We explored treatment outcome predictors for 3200 youth receiving psychiatric services. ► Parent-related variables (such as maternal anxiety) most predicted improvement. ► Variables associated with deterioration in functioning were more tied to child status. ► We discuss the benefits of data-driven clinic management and treatment planning.

Introduction

While a research literature exists on predictors of outcome for the treatment of specific childhood mental health disorders or circumstances (Brent et al., 1998, Cohen and Mannarino, 2000, Crawford and Manassis, 2001, Dow et al., 2007, Green et al., 2001, Liber et al., 2008, Merlo et al., 2009, Reyno and McGrath, 2006, Sallows and Graupner, 2005), few studies explore predictors of outcome for general clinic populations. Indeed, most studies report data from interventions delivered with a high degree of fidelity to a fixed treatment protocol. Those few studies that have considered moderator variables in real-life clinical settings have generally involved relatively small numbers of children and have considered only child or parent variables (Target & Fonagy, 2005). Therefore, the ecological validity of even these prior studies can fall into question.

The lack of solid data on predictors of treatment outcomes in general clinical populations is unfortunate because identifying such bellwethers could drive the development of more targeted clinic policies and procedures. Knowing up front what factors might predict outcome could allow for implementation of clinic processes that maximize the chances of providing effective care. For example, if it were to be determined that maternal depression was a significant predictor of a child's treatment outcome, a mother's psychiatric history would become a prime area for diagnostic inquiry and perhaps clinical management. Likewise, a finding that the distance between home and clinic was a predictive factor might pave the way for establishing more satellite clinics or perhaps home-based services. It also might provide important clues to the need for interagency collaboration around case management.

A systematic analysis of outcome predictors could also have a substantial impact on how clinic administrators assign and supervise cases. In an effort to avoid staff burnout, managers could rank cases on their likelihood of success based on empirically-derived algorithms. A clinician's caseload would therefore become more balanced between cases at risk for failure and those more likely to bear therapeutic fruit. Actual data on which cases might be most frustrating can help minimize caseload decision based more on bias or inaccurate presumptions.

The current study presents data on what predicts treatment outcome in a large number of patients from a child and adolescent psychiatric clinic. Using Reliable Change Index (RCI) methodology (Atkins, Bedics, McGlinchey, & Beauchaine, 2005), we divided children from a large sample of consecutive clinic referrals into those who, between intake and discharge, improved, stayed the same, or got worse according to clinician-rated impairment. In our analyses of variables that discriminated amongst the three groups, we included both child variables (e.g., age, sex, diagnostic status, and initial level of impairment) and parental factors (e.g., marital status, history of psychopathology, demographic factors). Based on our prior research (Gordon, Antshel and Lewandowski, 2010, Gordon, Antshel, Lewandowski and Seigers, 2010), we hypothesized that parent factors would emerge as predictive as those associated with the child. We also predicted that group membership would be predicted by the extent of comorbidity and the initial level of impairment.

Section snippets

Participants

We amassed data on consecutive referrals over a 6-year period (January 1, 2002–December 31, 2007) to a child and adolescent psychiatry clinic in an urban academic medical center. Given our focus on treatment outcomes, only those children who attended more than three sessions were included. Those children who attended two or fewer sessions (n = 512) were not included in the analyses. The sample consisted of 3200 patients (1884 males, 1316 females) between the ages of 3 and 17 years, 11 months (Mean

Group status

Using RCI statistics, 406 children (12.7%) were deemed worse, 1382 children (43.2%) were deemed the same and 1412 (44.1%) were deemed better.

Predicting group status

Significant predictors of group membership are presented in Table 1, Table 2. In both Table 1, Table 2, analyses were performed using “Worse” as the reference category; thus variables in Table 1, Table 2 reflect the odds of being in the dependent variable category (for Table 1, “Better”, for Table 2, “Same”) versus the comparison category (“Worse”)

Discussion

We found evidence that parent variables were at least as predictive as those associated with the child, and that predictors varied according to treatment outcome. For example, of the variables that predicted who improved, almost all were unrelated to child variables. Instead, outcome was more likely to be associated with parent variables such as being married, having a mother with a history of anxiety, and ethnicity. These findings mirror those of our previous studies in which we documented the

References (15)

There are more references available in the full text version of this article.

Cited by (7)

  • Systematic Review and Meta-analysis: Outcomes of Routine Specialist Mental Health Care for Young People With Depression and/or Anxiety

    2020, Journal of the American Academy of Child and Adolescent Psychiatry
    Citation Excerpt :

    For example, six publications55-68 were based on only one study sample; therefore, data from the original paper55 were included in the final meta-analysis. Seven studies were excluded because additional information was not provided by the corresponding authors on request.28,29,69-73 A total of 38 studies remained, relating to 44 independent study samples (Figure 1).

  • Do Parental Psychiatric Symptoms Predict Outcome in Children With Psychiatric Disorders? A Naturalistic Clinical Study

    2018, Journal of the American Academy of Child and Adolescent Psychiatry
    Citation Excerpt :

    Baseline child’s symptoms also were not always incorporated in a similar way as in the present study. Sometimes, a child’s change score was analyzed as an outcome measure6,10,22,33 or whether remission of a diagnosis was achieved.5,13 These analyses did not account for the higher symptom scores at baseline in children whose parents have psychopathology.

  • Therapist-Specific Factors and Psychotherapy Outcomes of Adult and Youth Clients Seen in a Psychology Training Clinic

    2017, Administration and Policy in Mental Health and Mental Health Services Research
View all citing articles on Scopus
View full text