Randomized controlled trials (RCTs) are considered the gold standard for assessing efficacy of treatments. However, two characteristics of typical RCTs limit their utility for patient, providers, and other decision-makers. First, RCTs often focus on narrow patient populations to increase internal validity. This limits generalizability of the RCT results to older and sicker target populations, especially when heterogeneity of treatment effect (HTE) is present. Second, many RCTs have short follow-up even though many treatments that they evaluate are intended to be used for a long time. Although RCTs provide evidence about treatment efficacy during the study follow-up, estimating the long-term impact of treatment is difficult. As patients age, their disease state can change and they can develop new and worsening conditions that modify the effect of a treatment. Therefore, the average treatment effect observed during an RCT may not accurately reflect long-term treatment outcomes. Lack of generalizability due to these limitations causes uncertainty in translating RCT findings in clinical practice, especially in the early post-marketing stages. There is an urgent need to develop methods that can make trial results more useful to patients, clinicians, and other decision-makers. We propose a novel approach to expand RCT results to broader patient populations and over longer periods in the presence of HTE. The approach uses evidence about HTE derived from individual-level RCT data and an individual-level simulation model to estimate effects applicable to the characteristics of external target populations, such as older individuals. Existing individual-level data from three landmark RCTs will be used as case studies to evaluate the feasibility and application of the proposed approach for generalizing RCT results to older patients identified in existing Medicare data: the Randomized Evaluation of Long-Term Anticoagulation Therapy (RE- LY) trial, the Prospective Comparison of ARNI With ACEI to Determine Impact on Global Mortality and Morbidity in Heart Failure (PARADIGM-HF), and the Efficacy and Safety of LCZ696 Compared to Valsartan, on Morbidity and Mortality in Heart Failure Patients With Preserved Ejection Fraction (PARAGON-HF). Our hypothesis is that information about HTE in the RCTs can be leveraged to predict treatment outcomes in a larger target population and over longer time periods than studied in the RCTs.
Generalizing randomized controlled trial (RCT) results to older and sicker target populations is often challenging, especially when heterogeneity of treatment effect (HTE) is present. There is an urgent need to develop methods that can make RCT results more useful to patients, clinicians, and other decision-makers. We propose a novel approach that uses information about HTE in the RCT data to predict outcomes in older target populations and over longer periods.
