— Identifying Critical Components of Behavioral Health Provider Fidelity in Co-located, Collaborative Care: Results from a Delphi Study

Beehler GP, VA VISN 2 Center for Integrated Healthcare; Funderburk JS, VA VISN 2 Center for Integrated Healthcare; Possemato K, VA VISN 2 Center for Integrated Healthcare; Vair CL, VA VISN 2 Center for Integrated Healthcare;

Objectives:
Developing accurate measures of fidelity to the Co-located, Collaborative Care (CCC) model in Primary Care-Mental Health Integration is vital as VA administrators/researchers begin to closely examine CCC processes and outcomes. This study aimed to build expert consensus regarding critical components of CCC to direct the development of a self-report measure of BHP fidelity.

Methods:
Using a modified Delphi approach (i.e., a structured iterative process using self-report surveys to identify consensus), we emailed a panel of 33 experts to identify the critical components of BHP behavior when working within a CCC model. Twenty-nine experts (88% response rate) completed the first survey rating 56-items as prohibited, essential, or compatible BHP behaviors. Respondents consisted primarily of doctoral-level psychologists (93%) from a variety of institutions including the VA (38%) and DoD (31%). Nineteen items met consensus after the first survey. Using expert feedback to revise items, 25 experts (86% response rate) completed the second survey rating 37 revised items and 2 new items suggested by the experts.

Results:
The 43 (74%) items that met consensus after the second survey reflected BHP behaviors across several domains, with the highest level of agreement on items reflecting the importance of accepting warm hand-offs from Primary Care Providers, providing feedback to Primary Care Providers about new referrals within 1 business day, accepting referrals for promoting healthy lifestyles or adjustment to medical illness, and documenting findings from functional assessments and screening instruments in progress notes. Participants' response patterns and qualitative feedback indicated that key contributors to low item consensus included 1) participants' concern about item wording, 2) reluctance to endorse key behaviors that would limit model flexibility and patient-centeredness, and 3) disagreement across sub-groups (VA, DoD, other) suggesting that the importance of some BHP behaviors vary by organizational setting.

Implications:
Our findings highlight the benefits and challenges of developing expert consensus to inform a coherent measure of CCC fidelity since some components previously thought to be vital to CCC met with low consensus.

Impacts:
Developing this fidelity measure will assist stakeholders in evaluating the implementation of CCC and the relation between BHP fidelity and Veteran health outcomes.