3148 — Using a Modified Stepped Wedge Design to Evaluate TIDES-Collaborative Care Management (TIDES-CCM) Intervention on Depression Process of Care Measures

Huynh AK, Veterans Affairs; Lee ML, Veterans Affairs; Farmer M, Veterans Affairs; Rubenstein LV, Veterans Affairs;

Objectives:
Assessment of naturalistically spread programs in VA has been hampered by low researcher control over implementation and by lack of electronic measures of program outcomes. We modified the randomized stepped wedge design to fit our evaluation needs. This study seeks to demonstrate the feasibility of using a modified stepped wedge design and longitudinal electronic depression care measures to evaluate primary care population outcomes, specifically collaborative care management (CCM) of depression.

Methods:
The modified stepped wedge design does not require randomization of timing in intervention adoption. Instead we evaluate this timing, use site characteristics, and assess the patient populations yearly per site. In this design the number of sites adopting an intervention at any given time period is not specified. The modified stepped wedge design uses all of the data points available by considering a within site pre-post evaluation, and a between site cross-sectional evaluation by time point. We applied the modified stepped wedge design to evaluate a TIDES-CCM intervention on longitudinal electronic depression process care measures. Measures assessed minimum follow-up care and receipt of minimum treatment after detection of a new episode of depression. Analyses accounted for the clustering of individuals within site across time periods.

Results:
Implementing the design posed challenges primarily related to the multi-level model used, related software limitations, and the precision with which implementation could be timed. Evaluation of the TIDES-CCM intervention showed there were improvements in outcomes over time, and primary care practice sites that were early adopters or medical center-based rather than community based had significantly better outcomes.

Implications:
The modified stepped wedge design can be implemented for assessing programs to improve VA depression care. The method can generate results based on outcome impacts for the full population of patients cared for at each evaluated primary care practice site, and can account for site contextual differences as well as patient characteristics.

Impacts:
The modified stepped wedge design has potentially wide ranging applicability for evaluation of VA programs that have undergone non-randomized implementation, can track timing of site level implementation, and are associated with relevant electronically-based patient level quality measures.