The Selection of Optimal Anchor-based Minimal Important Difference to Enhance Trustworthy Interpretation of Patient-reported Outcomes in Clinical Research and Evidence-based Decision Making

Abstract

Patient centered research encourages incorporating patient perspectives to inform the pursuit of clinical questions, the conduct of clinical research, the benefit-risk assessments and decision-making, as well as the delivery of appropriate health care. Researchers have thus increasingly adopted Patient-reported outcomes measurements (PROMs)—the instruments that capture patient-reported outcomes (PROs)—to gather the important outcomes from patients’ viewpoint without outside interpretation from anyone else, such as quality of life, mental health, and physical function. Typically, researchers use pre- and post-event PROM results to measure the impacts of an intervention, informing benefit-risk assessments and decision-making. The interpretation of the PROM results involves deciding whether a particular treatment effect is trivial, small but important, moderate or large. To aid such interpretation, researchers proposed anchor-based minimal important difference (MID) MID): the smallest important difference or change, either beneficial or harmful, that patients perceive as important important. The trustworthy interpretation, however, relies largely on the choice of an optimal anchor-based MID. With the widespread recognition of the usefulness of anchor-based MIDs, the number of published anchor-based MIDs for PROMs has grown rapidly. Though all the published MIDs have gone through peer review, considerable difficulties for optimal MID selection exist because the MIDs for a given PROM vary widely and are not equally trustworthy in one way or another. This thesis aims to address the use of anchor-based MIDs in enhancing the interpretation of PROMs, with particular focuses on the methodological issues related to selecting an optimal MID and the development of a systematic approach to selecting an optimal MID. To start with, we conducted a systematic survey of the literature addressing the issues related to selecting optimal MIDs MIDs. Subsequently, based on the survey information, we refined the existing anchor-based MID credibility instrument by adding the construct proximity assessment, which supplemented the MID methodology assessments. Then, informed by the work above, we developed the systematic approach to selecting an optimal anchor-based MID for a given PROM, which is geared by both the methodology rigor and application contexts. Finally, this thesis concludes with insights to the application of the selection approach and the opportunities for future research.