Rapid learning during a public health emergency: COVID-19 related medication treatment patterns in Medicare

SUMMARY/ABSTRACT Medical practice is continuously evolving, with substantial resources allocated to developing medical innovations and enhancing clinical evidence on new and existing treatments. In most cases, new evidence supporting approval by the Food and Drug Administration (FDA) demonstrates efficacy of new treatments; after approval, new evidence for existing treatments may show them to be ineffective or unsafe. Understanding whether and how new clinical evidence is integrated into practice is critical from the perspectives of a) improving patient safety and health outcomes; b) addressing racial and socioeconomic equity in access and use; c) designing programs and policies to encourage use of high value, effective treatments and abandon less effective and harmful treatments; d) containing health care costs by allocating limited health care budgets to their most effective use. The COVID-19 pandemic provides a unique opportunity to study the diffusion of evidence into practice by observing adoption and de-adoption behavior of physicians in response to rapidly changing information about potential treatments and risks of continued use of common drugs in the context of COVID-19. This proposal builds on our prior work on de-adoption of harmful or ineffective medical practices by introducing additional factors that could be associated with the rates and timing of adoption and de-adoption of pharmaceutical treatments. We will examine prescription fills and claims-based use of a targeted set of drugs related to COVID-19 using data from Medicare claims for Fee-for-Service (FFS) enrollees. COVID-19 has particularly affected older adults and the disabled and chronically ill, many of whom face increased risk of severe morbidity and mortality from the disease, and potentially from lack of access to medical care during the pandemic. We will investigate how prescribing of COVID-19-related drugs responds to state-level drug policies, FDA safety communications, and COVID-19 pressures on the healthcare system. We will assess how patient characteristics (e.g., race and ethnicity, age, and key comorbidities), as well as physician and medical practice organization characteristics are associated with these prescribing patterns. Three in-depth COVID-19 case studies will assess uptake and de-adoption of treatments associated with lower quality evidence and more rapid turnaround in a context where public attention is closely attuned to every development and pre-print biomedical manuscripts are discussed at length in the popular press. Exploring adoption and de-adoption of treatments in a context of high uncertainty and high burden and prevalence of disease will provide important insights into the role of information quality in physician decision making. In addition, these cases will allow us to explore the effectiveness of state-level regulation of treatments. In particular, hydroxychloroquine was regulated in some states in response to fears that not enough would be available for people using the drug to treat non-COVID-19 conditions.