Is the quality of evidence in health technology assessment deteriorating over time? A case study on cancer drugs in Australia.

Abstract

This study aimed to assess whether there have been changes in the quality of clinical evidence submitted for government subsidy decisions on cancer medicines over the past 15years. We reviewed public summary documents (PSDs) reporting on subsidy decisions made by the Pharmaceutical Benefits Advisory Committee (PBAC) from July 2005 to July 2020. Information was extracted on the study design, directness of comparison, sample size, and risk of bias (RoB). Changes in the quality of evidence were assessed using regression analysis. Overall, 214 PSDs were included in the analysis. Thirty-seven percent lacked direct comparative evidence. Thirteen percent presented observational or single-arm studies as the basis for decisions. Among PSDs presenting indirect comparisons, 78 percent reported transitivity issues. Nearly half (41 percent) of PSDs reporting on medicines supported by head-to-head studies noted there was a moderate/high/unclear RoB. PSDs reporting concerns with RoB increased by a third over the past 7years, even after adjusting for disease rarity and trial data maturity (OR 1.30, 95% CI: 0.99, 1.70). No time trends were observed regarding the directness of clinical evidence, study design, transitivity issues, or sample size during any of the analyzed periods. Our findings indicate that the clinical evidence supplied to inform funding decisions for cancer medicines is often of poor quality and has been deteriorating over time. This is concerning as it introduces greater uncertainty in decision making. This is particularly important as the evidence supplied to the PBAC is often the same as that supplied to other global decision-making bodies.