SMART trial designs, perioperative data ecosystems, and collaboration

At the Collaborative Clinical Trials meeting in Prato, Italy, Mike Grocott and co-host Mark Priestley speak with Sachin Kheterpal, University of Michigan Medical School's Associate Dean for Research Information Technology and the Robert B Sweet Professor and Kelley M. Kidwell, Chair of the Department of Anesthesiology and Professor of Biostatistics at the University of Michigan School of Public Health and Senior Associate Dean for Faculty Affairs . Kidwell explains sequential multiple assignment randomized trials (SMART) as multi-stage designs that randomize at key decision points to evaluate treatment sequences, distinguishing them from adaptive trials and addressing misconceptions about sample size. Kheterpal describes the University of Michigan–led Multi-Center Perioperative Outcomes Group (MPOG), contrasting EHR-based automated extraction and individual provider feedback with manually collected programs like NSQIP, and discussing alignment with AQI/NACOR and inspiration from UK National Audit Projects to improve documentation for future inference. They note emerging use of large language models to extract outcomes from free-text notes, discuss pragmatic trial designs including THRIVE, and emphasize the value of close statistician–clinician collaboration and scientific humility, highlighting a BMJ talk on countering misinformation. For more information on EBPOM Dingle, go here now: www.ebpom.org