Novel approaches to improve comparative effectiveness research of medical and surgical weight reduction strategies in clinical practice

The obesity epidemic will affect 1 in 2 adults in the US by 2030. While several medical and surgical weight reduction strategies are available, with newer, transformative medications recently approved, little is known about their safety and effectiveness in clinical practice. Real-world evidence (RWE) studies based on routinely

generated health insurance claims from millions of patients have the size and the breadth of data capture needed to provide critical information on the safety and effectiveness of weight loss strategies in clinical practice, complementing evidence from RCTs. In this setting, however, the adiposity measure that guides

weight loss treatment, i.e., body mass index (BMI), although captured with high specificity, is often missing. Yet, BMI information is necessary to accurately identify study populations, adjust for confounding, and assess effect modification on a large scale in claims data. Similarly, confounding and other biases may be present

in RWE studies. Thus, identifying methodological approaches that increase the validity of RWE studies comparing weight loss strategies is critical to generate unbiased findings that can validly address questions not answered or answerable by RCTs. Our overarching goal is to develop, implement, and test approaches

to produce large scale, high-quality RWE on the comparative safety and effectiveness of medical and surgical weight loss strategies to complement RCTs. To achieve this goal, we will leverage (1) the Nurses’ Health Studies (NHS) I and II and the Health Professionals Follow-up Study (HPFS), large longitudinal cohort

studies with rich lifestyle and dietary data, linked with their insurance claims data, and (2) large U.S. national (federal and commercial) claims databases, resulting in a heretofore untapped new data infrastructure. We will improve existing algorithms to measure BMI in national claims data using primary data from longitudinal cohort

studies (NHS I and II, and HPFS) linked with their claims data (Aim 1); assess the success of approaches to reduce confounding, including comparator choice and propensity score adjustment, with respect to measures of achieved comparability between medical and surgical weight loss strategies identified in claims-based RWE

studies with regard to information only measured in the linked data (Aim 2); evaluate when RWE studies of weight reduction strategies emulating target trials provide causal conclusions (Aim 3); and fill gaps in the RCT evidence in targeted questions regarding the safety and effectiveness of medical and surgical weight reduction

strategies in clinical practice (Aim 4). This work will develop more accurate, methodologically rigorous approaches to conduct RWE studies of weight reduction strategies using large scale, real-world data and will create an invaluable infrastructure to conduct studies on the comparative safety and effectiveness of various cardiometabolic treatments.