Almut G Winterstein, RPh, PhD, FISPE
Distinguished Professor & Chair In POP, Director For CoDES, Director For Consortium For Medical Marijuana Clinical Outcomes Research
About Almut G Winterstein
Almut Winterstein received her pharmacy degree from Friedrich Wilhelm University in Bonn, Germany and her PhD in Pharmacoepidemiology from the Humboldt University in Berlin, Germany. She holds the position of distinguished professor and chair in the Department of Pharmaceutical Outcomes and Policy at the College of Pharmacy and is the founding Director of the Center for Drug Evaluation and Safety, both at the University of Florida. She has an affiliate appointment in the Department of Epidemiology at the Colleges of Medicine and Public Health and Health Professions. In 2017, she was named the Dr. Robert and Barbara Crisafi Chair for Medication Safety in recognition of her research on drug safety and on devising ways to improve medication use.
Since joining the University of Florida in 2000, Winterstein has served as principal investigator on more than 25 extramurally funded grants and contracts and published more than 400 manuscripts and conference abstracts. Her research interests have centered on the post-marketing evaluation of drugs in pediatrics and perinatal care, infectious disease and psychiatry and the evaluation and improvement of quality surrounding medication use using real-world data. As expert in drug safety, she has chaired the Food and Drug Administration’s Drug Safety and Risk Management Advisory Committee from 2012-2018. Recognizing her contributions in pharmacoepidemiology, Dr. Winterstein was inducted as a fellow of the International Society of Pharmacoepidemiology in 2013 and currently serves as president of the society. Before she became department chair in 2016, Dr. Winterstein served as graduate program director in her department. She has chaired more than 20 PhD committees and was awarded the UF-wide Dissertation Advisor/Mentoring award for her excellence in graduate training.
My scholarly work has contributed to public health through three mechanisms: (a) by generating evidence on drug safety and effectiveness in real-world populations, (b) by evaluation of policy or programs that affect medication use and (c) by providing data-driven support of decision-making related to medication use (pharmaceutical data analytics/population-based care). While these three research areas may appear quite diverse, they are united by similar methodology that employs pharmacoepidemiologic methods on large data repositories for causal inference or prediction, and by the same goal to improve clinical outcomes through most effective and safe use of medications. As an epidemiologist, I have also contributed to the methodological advancement of my discipline in general and specifically, via development of new resources to enhance the assessment of medications in maternal and child health.
Contribution of new evidence on drug safety and effectiveness – Population-based observational studies of drug safety and effectiveness are critical when randomized trials are not feasible (e.g., due to ethical or logistical constraints such as for investigation of drug effects during pregnancy or of rare side effects requiring sample sizes that cannot be accomplished with prospective enrollment), or if treatment effects are modified by a range of parameters that require examination of heterogeneous populations. Pharmacoepidemiology focuses on the assessment of drug effects in real-world populations using large health databases and complements clinical trial evidence. Examples of such evaluations that I have led include assessments of rare safety outcomes such as the cardiovascular (Winterstein, BMJ 2012, journal impact factor [IF] 27.6, Altmetric score [AS]: 38) and suicidal risk of central nervous system stimulants in the treatment of attention-deficit/hyperactivity disorder (Linden, Pediatrics 2016, IF: 5.4, AS: 43), the effect of quinolone ear drops on tympanic membrane perforations (Alrwisan, 2017, IF: 9.1, AS: 528), and the effect of combinations of opioids with skeletal muscle relaxants on the risk for opioid overdose (Li, Clin Pharm Ther 2020, IF:7.3 AS: pending). I am currently funded by the FDA to evaluate the safety of gadolinium-enhanced MRI procedures during pregnancy on the risk for stillbirth and sudden infant death, which will inform regulatory decisions regarding warnings or restrictions on the use of gadolinium in pregnancy. I have also conducted a variety of effectiveness studies that looked at scenarios where indications were expanded in clinical practice without clinical trial evidence to support efficacy. Since 2019, I have served as the principal investigator of the state-funded Medical Marijuana Clinical Outcomes Consortium, which aims to develop a repository of linked marijuana dispensing and medical records, to evaluate both safety and effectiveness outcomes. Several of my safety and effectiveness studies have informed current treatment guidelines, including for example, work on treatment and risk of respiratory syncytial virus in infants, which has been included in the American Academy of Pediatrics Vaccination guidelines and my work on stimulant safety, which have been included in Canadian and American guidelines and several position statements. In recognition of my contribution to pharmacoepidemiologic research, I was inducted as fellow of the International Society of Pharmacoepidemiology in 2013.
Evaluation of policies and programs that affect medication use – With our understanding of drug effects consistently improving, appropriate and timely adoption and integration of such emerging information in clinical practice requires similar research support as the study of drug effects itself. Motivated by my long-standing work on the FDA Drug Safety and Risk Management Advisory Committee, I have developed great interest in the effectiveness of regulatory action aimed at affecting provider or patient behavior to mitigate drug risk and optimize patient outcomes. I collaborated with the FDA to evaluate educational efforts such as the quality of consumer medication information dispensed in retail pharmacies (Winterstein, JAMA Intern Med 2012, IF: 20.8, AS: not available). The most recent expansion of this stream of research unites my interest in risk mitigation strategies and perinatal drug safety in the evaluation of FDA’s efforts to reduce fetal exposure to teratogens. Our first paper found suboptimal effectiveness of FDA’s risk mitigation program for mycophenolate (Sarayani 2019, BMJ Qual Safety, IF: 7.2; AS: 11), pointing to the need for enhanced efforts to prevent fetal exposure. Finally, a whole body of work over the past decade has evaluated drug utilization pattern and etiologies for suboptimal and disparate care. Examples include studies on psychotropic medication use in pediatrics, such as my paper on psychotropic polypharmacy trajectories in young children (Winterstein, J Clin Psych 2017, IF: 5.5, AS: 30).
Population-based care and quality improvement in medication use – Data are powerful tools to guide quality improvement efforts both through assessments of clinical care for benchmarking purposes as well as by delivering real-time guidance to support clinician and patient decision-making in concrete clinical scenarios. Though no longer a practicing clinician, I consider my work to support population-based care my clinical practice. My first funding as faculty at UF, shortly after release of the sentinel Institute of Medicine Report on Medical Errors, was focused on the development of EHR-based tools to characterize and measure medication errors in hospitals (Kanjanarat, AJHP 2003; IF: 1.9 ; AS: 3, 348 citations). In parallel I began to support the UF Health pharmacy service and the hospital Medication Safety Committee with EHR-based evaluations of ongoing quality improvement efforts. My experience in quality assessment and improvement resulted in more than a decade of collaboration with the Health Service Advisory Group to support the Centers for Medicare and Medicaid (CMS) Quality Improvement programs. Responsible for the scientific development and validation, my team developed 14 performance measures that are endorsed by the National Quality Forum (NQF) and used in CMS reporting and pay-for-performance programs (see for example Winterstein, Medical Care 2020, IF: 3.1, AS: 5). More recently, with funding from the ASHP Research Foundation, my team developed automated scoring systems that predict hospitalized patients’ risk for 16 preventable adverse drug events (Winterstein, AJHP 2018, IF: 1.9, AS: 12). We received funding from the FDA Safe Use Initiative to implement our hypo- and hyperglycemia prediction algorithms, which dynamically score risk and post a prioritized list of patients for intervention. I am particularly excited about this ongoing work because it expands process-focused clinical decision support to a patient-centered approach that aims to capture the totality of risk factors that might cause a preventable adverse drug event.
Methodological contributions to pharmacoepidemiology – Central to my research program is the use of existing real-world data to inform medication use. Critical for the validity of such work is the validity of the underlying data sources, i.e., the data’s ability to support causal inferences by providing appropriate comparison groups with similar risk for the examined outcome, valid capture of drug exposure (over time) and determinants thereof and accurate capture of relevant outcomes. While cumbersome, assessment of data’s fit for purpose is critical to avoid bias and ought to be shared with the larger research community. If significant issues are discovered, methodological workarounds might be feasible (but require testing) or data needs to be enhanced, e.g., through additional linkages. Several of my scientific papers have focused on the development and validation of measures of outcomes or risk factors for application in pharmacoepidemiologic causal inference studies such as my work on misclassification of diabetogenic risk (Winterstein, PDS 2014, IF: 2.9 , AS: not available) I have also extensively worked on database linkages and variable cross-validation such as the development of algorithms to evaluate whether encounters reimbursed under capitated payment schemes are fully captured in claims data (Choi, PDS 2109, IF: 2.9, AS: 3). To enhance our foundation for pharmacoepidemiologic work in pregnancy, my group has established and validated mother and infant linkages within public and private insured administrative data (Knox, PDS 2019, IF: 2.9, AS: not available) and cross-linked these data to Vital Statistics records to build algorithms to estimate gestational age and time of conception. These algorithms facilitate accurate timing of in utero exposure to enhance our ability to evaluate teratogens such as exemplified on our recent evaluation of mycophenolate (Thai, PDS 2020, IF: 2.9, AS: 10), and are a central component of my ongoing collaboration with colleagues at Harvard on a large, FDA-funded study aimed at enhancing this methodology.
- Patient Safety
- Program evaluation
Impact of the Transition from ICD–9–CM to ICD–10–CM on the Identification of Pregnancy Episodes in US Health Insurance Claims Data
Medical Costs and Productivity Loss Due to Mild, Moderate, and Severe Asthma in the United States