OBJECTIVES: The Institute for Clinical and Economic Review’s (ICER) Unsupported Price Increase (UPI) reports aim to identify drugs with substantial price increases without adequate evidence to justify the increases. ICER’s methodology and criteria for accepting evidence have received critique, and there is limited research on how different types of manufacturer-submitted evidence are appraised. Our objectives were to review ICER’s determinations for manufacturer-submitted evidence and identify acceptance and rejection trends.

METHODS: We assessed the 3 UPI reports published in 2019 to 2021 by developing a codebook to categorize each type of evidence along with ICER’s determination. We examined ICER’s reasons for accepting or rejecting evidence and identified study characteristics for accepted evidence.

RESULTS: Manufacturers submitted evidence for 26 of the 31 drugs reviewed in the 3 reports. Of 991 pieces of evidence, 976 (98%) were not accepted by ICER. Nearly two-thirds (n=658) did not meet ICER’s UPI review criteria; the remaining one-third (n=318) did not meet ICER’s criteria for new evidence to be moderate to high quality. Only 15 pieces of evidence were accepted as high-quality evidence demonstrating important new information. All evidence deemed high quality was from randomized controlled trials (RCT) in phase 3 (n=14) or phase 4 (n=1), with the majority double-blinded (n=12). Six studies supported Food and Drug Administration (FDA) label expansion for a new (n=2) or existing (n=4) indication, 5 demonstrated improved longer-term outcomes, 2 extended the evidence base to new populations, and 2 supported accelerated approval.

CONCLUSIONS: Our findings indicate the vast majority of evidence (98%) was not accepted by ICER, calling into question the restrictiveness of their criteria. Accepted evidence was typically from a late-phase double-blinded RCT that demonstrated information leading to an FDA label expansion, knowledge about improved outcomes or a new population, or strengthening the accelerated approval evidence base.