Online Program

Return to main conference page

Thursday, January 11
Thu, Jan 11, 9:00 AM - 10:45 AM
Crystal Ballroom B
Pragmatic Cluster Randomized Trials

Statistical challenges in the design of a pragmatic trial of primary care-based treatment for opioid use disorders: the PROUD trial (304166)

*Jennifer Bobb, Kaiser Permanente Washington Health Research Institute 
Katharine Bradley, Kaiser Permanente Washington Health Research Institute 
Abigail Matthews, The Emmes Corporation 
Jennifer Bacik McCormack, The Emmes Corporation 
Hongxiang Qiu, University of Washington 

Keywords: opioid epidemic, opioid use disorders, pragmatic clinical trial, cluster-randomized trial, electronic health record data, pharmacy data

Addiction to opioids is an epidemic needing new approaches to ensure access to and retention in evidence-based treatment. The PRimary care Opioid Use Disorders (PROUD) trial is a pragmatic, cluster-randomized implementation trial to evaluate the effectiveness of a collaborative care model for treating opioid use disorders (OUDs) within primary care in 6 diverse health systems (NIDA CTN-0074). Pragmatic trials often have several advantages over more traditional explanatory trials, including speed, large sample sizes, and generalizable study populations. Here we describe challenges in the design of the PROUD trial. Issues include combining complex, electronic health record (EHR) data across diverse health systems, developing a clinically meaningful, new study outcome that is defined using administrative and EHR pharmacy data that were not collected for research purposes, and incorporating multiple observational sub-aims to address study hypotheses. A particular challenge is the need to assess health outcomes among a latent population of individuals (those with OUDs, often undiagnosed), where the identification of this population may be affected by the care model being tested.