Cluster Randomized Trials in Health Policy Research
Randall D Cebul, Case Western Reserve University
Neal V Dawson, Case Western Reserve University
*Thomas Ezra Love, Case Western Reserve University
Keywords: study design, balancing, sample size, decision support, electronic medical records, diabetes
Electronic medical records (EMRs) with sophisticated clinical decision support (CDS) functions are increasingly common in health systems that have affiliated clinical practice sites. Cluster-randomized trials (CRTs) of different approaches to CDS are facilitated by EMRs in these systems by: 1. enabling identification of patients and problem areas that might benefit from CDS; 2. providing rich clinical data for a priori balancing of practices on important characteristics before allocating practice clusters to intervention groups; and 3. providing platforms for developing different types of CDSs; for example, for patients or for providers. Course leaders will illustrate key points by highlighting an AHRQ-support CRT of CDS in diabetes across two organizations and 24 practice sites, and a small-group interactive task completed during the session will motivate the presentation.