OBJECTIVES: The Institute for Clinical and Economic Review (ICER) positions its Unsupported Price Increase (UPI) reports as relevant policy tools. We aimed to identify trends in manufacturer evidence submissions and ICER’s rejection/acceptance decisions.

METHODS: A codebook was developed to compile and categorize manufacturer-submitted evidence and ICER’s reasons for rejecting or accepting evidence across 5 national UPI reports published from 2019 to 2023. We identified submission and rejection trends, as well as study characteristics for accepted evidence.

RESULTS: Manufacturers submitted evidence for 42 of the 57 drugs reviewed across the 5 reports, totaling 1,237 pieces of evidence and averaging 29 pieces per drug. The total evidence per UPI report and average pieces of evidence per drug declined over time, ranging from n=67 pieces per drug in 2019 to n=12 in 2023. Overall, 96% of evidence submissions were rejected by ICER, with a slight downward trend, ranging from 99% in 2019 to 91% in 2023.

Only 21 distinct randomized controlled trials were accepted by ICER, spanning phases 2-4, with the majority double-blinded. In 2019, ICER described the accepted evidence using a single category: longer-term data with improved outcomes (n=5). In subsequent years, ICER continued to introduce new descriptive categories—including evidence that supported Food and Drug Administration (FDA) label expansion for a new (n=5) or existing (n=4) indication, extended the evidence base to new populations or outcomes excluded in previous trials (n=3), supported FDA accelerated approval (n=2), supported FDA approval requirements for longer-term safety data (n=1), and strengthened the existing evidence base and guideline recommendations (n=1).

CONCLUSIONS: Manufacturers appeared to become increasingly selective over time with evidence they submitted to ICER’s UPI reports, while ICER accepted slightly more evidence with each UPI report. However, our findings demonstrate that ICER continues to reject the majority of UPI evidence submissions (96%), calling their policy-relevance into question.