PROJECT SUMMARY/ABSTRACT Unnecessary tests and treatments represent two of the greatest sources of preventable patient harm and excess healthcare spending in the United States. Robust evidence demonstrates routine preoperative tests such as electrocardiograms, blood and urine tests, and chest radiographs before low-risk surgery do not prevent adverse events or improve outcomes. In addition, unnecessary preoperative testing can trigger care cascades leading to additional specialty consultations and invasive procedures, potentially resulting in significant patient harm, delays in surgery, and patient financial toxicity. Numerous professional societies recommend against routine preoperative testing prior to low-risk surgery but eliminating use has proven difficult. In the US, few investigators have moved beyond describing the prevalence of low value testing or testing limited strategies targeting a single procedure (i.e., cataract surgery) or hospital. As a result, there are no broadly scalable de-implementation strategies and unnecessary, wasteful, and harmful testing remains commonplace. This research will test a theory-informed and broadly scalable multi-level, multi-component de- implementation strategy to reduce unnecessary preoperative testing before low-risk surgery across diverse procedures and practice types. In the Michigan Value Collaborative (MVC), a consortium of 87 Michigan hospitals, routine preoperative testing occurs before low-risk surgery in more than 52% of patients with inter- facility variation ranging from 8% to 85%. Our preliminary data has identified important, yet modifiable determinants of unnecessary testing. Based on these findings, we will test a multi-level, multi-component de- implementation strategy across Michigan hospitals. Using facilities jointly participating in the Michigan Surgical Quality Collaborative (MSQC), the largest surgical collaborative quality improvement program in the United States, we will evaluate the effectiveness of our intervention through a stepped wedge cluster randomized trial.