A Multilevel Longitudinal Analysis of the Relationship Between Therapist Use of Practices Derived from the Evidence Base and Outcomes for Youth with Mood Disturbance in Community-Based Mental Health

Abstract

Depressive disorders impact one in four youth before the age of 18. They are associated with numerous short- and long-term negative sequelae. The current study utilized a common elements methodology to identify 15 discrete practice elements derived from the evidence base (PDE) treatment literature on youth depression and examine their impact on progress towards mood-related treatment targets during the first six months of intensive in-home (IIH) UC treatment. Fifteen mood-related PDEs were identified by consulting the PracticeWise Evidence-Based Services database (2013). Other client-level predictors (e.g. age, gender) were also examined. Study data were drawn from archival records in the Hawaiʻi Child and Adolescent Mental Health Division system. Information on therapist-identified treatment targets (TT), progress ratings (PR), and practice elements were obtained using the Monthly Treatment Progress and Summary (MTPS; CAMHD, 2008). This study focused on PRs for five mood-disturbance related TTs identified via factor analysis. Monthly PRs were examined using a two-level multilevel longitudinal modeling approach for 402 youth (52% male) age 10 to 18 years old (M= 14.87, SD= 2.03) with at least 90 days of IIH and the TT Depressed Mood endorsed one or more times in the first six months of treatment. The average mood-related PR at each measurement occasion served as the outcome variable. The repeated measurement of time was modeled at Level 1, with the intercept placed at the last measurement occasion in order to compare PRs at the sixth month endpoint. Total number of PDEs received during the study window was the primary predictor of interest at Level 2. There was significant variability in PRs at the sixth month to be modeled. However, no significant slope for change over time emerged (perhaps because there was no pretreatment baseline measure from which to measure trajectory of change). The best-fitting model had an intercept of 3.63 on the MTPS rating scale. On average, youth improved 0.764 or three-fourths of an MTPS rating scale point from beginning of treatment to the end of the study window. There was a positive, significant relationship between the total PDEs a youth received and his/her PRs on moodrelated TTs in the sixth month (stdz coeff= 0.19). Three other variables also predicted intercept: functional impairment (stdz coeff= -0.11), gender (stdz coeff= 0.10), and number of diagnoses (β= 0.387 for youth with two disorders). Collectively, these predictors explained 12.2% of the variability in PRs at the sixth month. Results speak to the importance of ongoing implementation efforts for empirically supported practices, and add to the growing body of research pointing to the possible utility of a modular approach to UC practice and research.