Medical reversals: De-implementing ineffective and unsafe treatments

摘要

Abstract Understanding whether and how robust clinical evidence is integrated into practice is critical from the perspectives of: a) improving patient safety and health outcomes; b) designing programs and policies to accelerate the use of high value, effective treatments and abandon less effective and harmful treatments; c) containing health care costs by allocating limited health care budgets to their most effective use. In most circumstances, physicians and health care delivery organizations (HCDOs) are the key agents in determining whether a patient receives a given medical treatment. However, physician and HCDO integration of clinical evidence into practice is not well studied or understood. In general, the term “integration of evidence into practice”, could mean both the adoption of new treatments and de-implementation of established treatments based on new evidence related to effectiveness or safety. In this proposal, we will focus on the latter. Our primary focus is to understand how physician networks, HCDOs and physician market environment influence the de-implementation of ineffective and unsafe treatments in practice. As such, we have the following aims: Aim 1: To describe variation in the de-implementation of ineffective and unsafe treatments across physicians and HCDOs. Aim 2: To investigate how characteristics of the physician (i.e. age, gender, years since medical school or residency, patient-mix), HCDO (i.e. practice size, specialty mix, ownership, level of integration), physician's patient sharing network (i.e. betweenness centrality), and physician's market environment (competition, malpractice environment) influence de-implementation of ineffective and unsafe treatments by physicians. Aim 3: To assess the influence of de-implementation of ineffective or unsafe treatments in the physicians' network (defined by HCDO affiliation and patient-sharing) on physicians' likelihood to de-implement ineffective and unsafe treatments. We will focus on different case studies concerning treatments with implications for cardiovascular outcomes and safety. Our study sample will include the Medicare Fee-For Service (FFS) population (from CMS) as well as the commercially insured and Medicare Advantage (MA) populations (OptumLabs).