PCP60 - EVOLVING STRATEGIES FOR GENERATING, SYNTHESIZING, AND VISUALIZING REAL WORLD EVIDENCE IN RARE DISEASES

Abstract

RWE for rare diseases is often unavailable. While registries can provide valuable data, these are commonly used to satisfy post-marketing approval requirements and can have limited access and chart data completeness. Non-registry RWE are commonly in the form of case reports. Given the limitations with available RWE in rare diseases, decision-makers often resort to expert opinion. However, rigorously analyzed and synthesized published data can help provide a more objective and complete picture of disease burden. Our objective is to present approaches for generating, synthesizing, and visualizing literature-based evidence rare diseases. We began by developing a structured framework to guide approaches to using rare disease RWE from the literature. This included consideration of: how to identify and collect the data; which data are appropriate to include (given that more severe cases and outcomes are more likely to be reported); how to handle exposure time and number at risk; ways to visualize patterns in patient presentation; and how to integrate expert opinion. We applied and refined the framework using published RWE from case reports from >500 patients with hypophosphatasia. The framework, and how the case report data informed strategies of interest to HEOR practitioners, will be presented. Strategies included tabulating the occurrence of key clinical events; using Kaplan-Meier curves to estimate time-to-event, accounting for censoring; and approaches to identify patterns in clinical data. A checklist for analytic and visualization features useful in displaying these data will also be presented. In the absence of large-scale data collection, methods to understand the epidemiology, burden-of-illness, and outcomes in rare diseases are needed. Several methods exist for synthesizing and visualizing published data to answer HEOR-related research questions, to supplement data from registries and expert opinion. While these methods have promise, it is important to consider potential limitations inherent in using data collected in a non-standardized fashion.