Differential Impact of Wraparound Facilitation Best Practices on Youth Outcomes: A Latent Class Analysis

Background & Purpose:

The wraparound best practice model outlines ten key principles and four distinct phases of work.  Research has shown that doing wraparound with fidelity to this model is important to produce optimal outcomes. The Wraparound Fidelity Index-EZ measures adherence to these principles and phases.  This study used Latent Class Analysis (LCA) to examine WFI-EZ data to determine if the population of wraparound facilitators consisted of distinct subpopulations characterized by their use of certain best practice indicators.  We then examined whether class membership predicted youth outcomes. This study found four distinct classes of providers and class membership differentially predicted youth outcomes.

Methods:

Latent Class Analysis was used to determine and define subpopulations characterized by  integration of best practice indicators among 1,970 wraparound facilitators.  Facilitators were classified into their most likely baseline class based on the posterior probabilities retained form the LCA model.  A regression analysis was conducted to estimate the differential impact of class membership on youth outcomes that include on-going placement in a community setting, and the absence or presence of contact with law enforcement, treatment in an emergency room, and school expulsion.  Additional outcomes assess the degree to which the youth has experienced problems that disrupt home life, interfere with school success, friendships, or participation in community activities. MPlus software was used in this analysis. Specific latent class solutions were determined by the agreement between the statistical indices, substantive theory, and empirical evidence. Full Information Maximum Likelihood (FIML) estimation was employed to make use of all available data, including cases with missing data. 

Results:

LCA results indicated that a four class model was the best fit in determining classes of facilitators (BIC = 53,712) and had a low level of class misspecification (Entropy = .89).  The four classes of facilitators were distinguished as: 1) A High Fidelity, High Confidence (HFHC) class (35%); 2) A Low Fidelity, Low Confidence (LFLC) (24%); 3) a Moderate Fidelity, Moderate Confidence (MFMC) group (24%); and 4) a Low Fidelity, Moderate Confidence (LFMC) class (16%).  The regression analysis revealed that on most outcomes of interest, youth receiving wraparound services from facilitators in classes 2, 3, and 4 did significantly less well than youth receiving services from class 1 – or the High Fidelity, High Confidence Class. 

Conclusion & Implications:

This study found that this sample population of wraparound facilitators consisted of four separate classes of providers, characterized by their individual adherence to the wraparound best practice model.  Facilitators who practice with the highest degree of fidelity to the model and confidence achieved the best outcomes, as compared to other subpopulations of providers with varying degrees of fidelity and confidence. Results of this study are consistent with the literature indicating achievement of positive outcomes is contingent to some extent on adherence to the established wraparound best practice model. Findings of this study extend this evidence by demonstrating the differential impact of wraparound facilitator practice types. Our findings support previous research and highlight specific areas in which providers could target skill development in order to achieve better youth outcomes.