— Variation in Key Informant Reporting of Primary Care/Mental Health Integration Implementation

Benzer JK, COLMR, VA Boston Healthcare System; Cramer IE, COLMR, VA Boston Healthcare System; Burgess JF, COLMR, VA Boston Healthcare System;

Objectives:
Implementation research methodology varies from single source key informant interviews to detailed ethnographies. This study will demonstrate how variation in key informants can provide unique implementation data for interventions that cross professional boundaries.

Methods:
An operational partner requested evaluation of the implementation of primary care (PC) mental health (MH) integration (PC/MHI) in selected clinics one year after the PC/MHI implementation mandate. The current study focused on co-located MH care at two sites that each included a hospital-based PC/MHI clinic and a large (10,000+ unique patients) affiliated CBOC. Key informants were MH leaders, MH staff co-located in the PC setting, and PC staff.

Results:
Compared to PC and MH staff co-located in the PC setting, MH leaders accurately reported both the procedures used to coordinate MH care in the PC setting and identified the major barriers to implementation. MH and PC staff informants elaborated on these themes, providing important operational details not reported by leaders. MH staff reports detailed the variation in how PC/MHI was implemented at sites. PC reports facilitated inferences regarding the working relationships between services that were not available from any single source. Both hospital-based PC/MHI clinics had barriers to collaboration. Initial PC resistance to co-location appeared to be the source of difficulty for one hospital, whereas resource-protective MH service agreements were the source of difficulty for the other. Collaboration was more apparent in CBOCs than in the hospital-based clinics.

Implications:
Results support the validity of single source key informant interviews for establishing the general structure of an intervention, but multiple key informant interviews were found to provide unique insights not available from any single source. In a multi-disciplinary intervention such as PC/MHI, key informants should be drawn from multiple professional groups and from both leaders and frontline staff.

Impacts:
Variation in perceptions across professions and hierarchy can be important in understanding the challenges to intervention implementation. This variation is likely to be most important for interventions that span disciplinary boundaries such as PC/MHI and Patient Aligned Care Teams (PACT). PACT implementation studies in particular may benefit from inclusion of key informants from both primary care and specialty care.