Leveraging Large Linked Datasets and Novel Analytic Methods to Understand the Long-term Effectiveness and Value of Cardiac Rehabilitation in Understudied Populations - PROJECT SUMMARY Cardiac rehabilitation (CR) combines supervised exercise, risk factor modification, and psychosocial support in a structured program that may lower cardiovascular morbidity, mortality, and healthcare spending in selected patients with cardiovascular disease. Yet only 1 in 4 eligible patients in the U.S. receives CR, with marked disparities by sex, race/ethnicity, socioeconomic level, and rurality. Barriers to uptake include uncertainty of long-term benefit among clinicians and patients, limited availability of local CR facilities, and high out-of-pocket costs. Most randomized clinical trials of CR in the U.S. did not enroll diverse populations, were too small to evaluate heterogeneity, and could not examine long-term outcomes. Moreover, major advances in cardiovascular therapies – such as transcatheter valve replacement or new treatments for heart failure – may have altered the effectiveness of CR, limiting the generalizability of past trial data. Observational analyses of real-world data may provide contemporary insights, but commonly used statistical methods are challenged by immortal time bias and confounding, particularly when treatments, like CR, need to be sustained over time. This project applies novel causal inference methods that have been specifically developed to address these limitations to large, diverse datasets incorporating administrative claims, electronic health records, and a cardiovascular registry, all linked with national health system and community-level socioeconomic data. Aim 1 will evaluate the effectiveness of CR in diverse populations of Medicare, Veterans Health Affairs, and commercial insurance beneficiaries. It will use the target trial framework (which uses observational data to emulate the hypothetical trial that would definitively answer the research question of interest) and quasi- experimental designs (which exploit random variation in CR use). It will examine both in-person and virtual CR, the uptake of which has increased since the COVID-19 pandemic. Aim 2 will examine individual-level variability in response to CR using novel risk score- and effect score-based analyses to derive personalized estimates of treatment benefit. The final validated score-based model will be used to develop an online tool so that clinicians can incorporate this information into shared decision-making regarding CR use. Aim 3 will use time-driven activity-based costing to estimate the cost of delivering in-person or virtual CR, and use nationally representative simulation models to evaluate the population health benefit and cost-effectiveness of CR. IMPACT: This research, which will be the largest and most comprehensive contemporary evaluation of CR in the U.S., will directly address key knowledge gaps identified by the Centers for Disease Control's Million Hearts Initiative. Our multidisciplinary team will evaluate the long-term effectiveness of CR in diverse populations to increase evidence-based CR uptake, produce personalized estimates of treatment benefit to facilitate shared decision-making, and project long-term cost-effectiveness to stimulate strategic investments needed to enhance equitable access to this underutilized guideline-directed therapy.
