Restricted Mean Survival Time to Interpret Clinical Trials for Treatment Decision-Making in Older Adults
Funded Grant
Overview
Affiliation
View All
Overview
description
PROJECT SUMMARY/ABSTRACT Patient-centered care of older adults with multimorbidity involves shared decision-making based on accurate communication of evidence. In clinical trials, treatment effect is conventionally summarized in terms of relative risk reduction using hazard ratios and absolute risk reduction. Despite widespread use, these conventional measures based on probabilities are not well understood by clinical community and often misleading in treatment decision-making. We have recently proposed restricted mean survival time (RMST) as a patient-centric outcome metric that can be intuitively interpreted as the average event-free survival time up to a pre-specified time point. Treatment effect can be summarized using RMST difference, which means “gain or loss in event-free survival time due to treatment in a pre-specified period”. Since it is expressed on a time scale that most clinicians and patients can relate to, it has great potential to facilitate shared decision-making in older adults, which involves assessment of benefit within a defined time frame, such as the remaining life expectancy. The objective of this application is to determine the usefulness of RMST in treatment decision-making, using an example of intensive vs. standard blood pressure lowering strategies in older adults. Our hypothesis is that presenting evidence using RMST difference, rather than absolute and relative risk reduction, would be more effective in reducing patients' uncertainty about treatment choice. To test our hypothesis, we will accomplish 2 specific aims to: 1) determine the benefit of intensive vs. standard blood pressure lowering strategies using RMST difference in older adults and identify characteristics associated with a greater RMST benefit; and 2) evaluate the effect of evidence com- munication formats based on RMST difference vs. conventional measures on patients' uncertainty in treatment decision-making. For Aim 1, we will analyze 2 publicly available datasets from the Action to Control Cardiovas- cular Risk in Diabetes Trial and Systolic Blood Pressure Intervention Trial to estimate RMST difference. For Aim 2, we will conduct an online survey of 200 community-dwelling older adults and an in-person survey of 100 residents at a local senior housing site. Participants will be randomized to one of the 2 evidence presentation formats, RMST vs. conventional measures, to assess the effect on reducing uncertainty in decision-making. We will conduct focus group interviews of 20-30 survey participants to further understand the reasoning and context behind their choice. Our approach is innovative, because we apply an intuitive and methodologically robust RMST to analyze and interpret clinical trial data, and rigorously evaluate the acceptability and understanding from the patient's perspective by adopting quantitative and qualitative design. The impact of this research is expected to be significant, because our method has great potential to be generalized to other clinical trials with time-to-event endpoints to facilitate shared decision-making in older adults.
