PCN390 ANALYSIS OF INDIRECT TREATMENT COMPARISONS IN ONCOLOGY-RELATED HTA DECISIONS IN GERMANY

Abstract

In the absence of a real comparator, an indirect treatment comparison (ITC) allows demonstrating the comparative value of a technology. This study aims to investigate the relationship between ITC methods/data type used in oncology-related health technology assessment (HTA) submissions and associated decisions in Germany. Submissions were observed from the G-BA website as of April 2019. Submissions with added benefit were interpreted as positive decisions. For each submission, ITC-related data alongside descriptive information (year of submission, indication, etc.) were hand-collected. Data on decisions were regressed on ITC characteristics and covariates in linear probability and logistic regression frameworks. The study included 153 oncology-related submissions on 17 indications and 69 distinct agents, including agent combinations (10/2011-03/2019). 37 submissions (24.2%) included an ITC method (14 with naïve-comparison, 4 with matching-adjusted indirect comparison (MAIC), 4 with propensity score matching (PSM), 18 with other methods). Regarding data type used for ITC purposes, 20 submissions used external data from randomized-control trials (RCT). Only 1 submission used real-world evidence (RWE) from claims data and 2 from registry data. 9 submissions used data from other sources (e.g. prospective observational studies). Controlling for dummy variables for each cancer indication, orphan status and submission year, on average, submissions with ITC increased the probability of a positive decision (odds ratio=3.3, p=0.048). However, method selection and data type matter: While naïve-comparisons and MAIC increased the probability of positive decision (p<0.01 each), no significant effect of PSM was observed (p=0.99). Using external RCT data increased the probability of positive decision (p<0.01 each), whereas the effect of using data from single-arm trials was not significant (p=0.63). The selection of ITC method and the nature of the data matter for positive G-BA HTA decisions. More data are needed to assess the usability of RWE in ITC from the perspective of HTA decision-making.