Characteristics of VA hospitals by participation status in a large pragmatic embedded clinical trial

The Diuretic Comparison Project (DCP) was a multi-center, embedded pragmatic trial conducted within the VA healthcare system comparing chlorthalidone and hydrochlorothiazide in preventing major adverse cardiovascular events and non-cancer deaths in hypertensive patients. Study procedures were decentralized, and facility leadership first had to agree and accept study procedures before staff approached participants. Recent evidence suggests facilities that choose and choose not to participate in trials may differ and this study considered such differences within DCP.

A cross-sectional comparison of facilities participating in DCP was conducted with data from the study start (June 2016) including: 2016 American Community Survey, 2016 Strategic Analytics for Improvement and Learning reports, star ratings, and 2014 hospital complexity. Characteristics of participating and non-participating centers were compared using logistic regression, including county-level socio-economic features and hospital-level performance.

Of 144 VA medical centers, leadership at 75 centers (52 %) initiated participation. Facilities in highly-urban and higher median income counties were more likely to participate, as were higher-complexity facilities. Facilities with research experience were 2.18 times as likely to participate. No other hospital performance metrics or county-level demographics were associated with participation.

Overall, this study suggests research exposure and quality care metrics may impact a facility's decision to participate. These results highlight key considerations for recruitment to multi-site, and particularly pragmatic, clinical trials. Consideration of supporting facilities that have not historically participated in research may be fruitful for recruitment. These results emphasize the importance of education about pragmatic study design and its integration with clinical care.