OBJECTIVES: Reimbursement authorities commonly require budget impact analyses (BIA) to support formulary decision-making. However, unlike cost-effectiveness analyses, which rely on the use of cost-effectiveness thresholds for interpretation, proper interpretation of BIA results is hampered by the lack of an accepted gold standard against which they can be evaluated. We conducted a systematic review of full-text, peer-reviewed, US-based pharmacotherapy BIAs to establish benchmarks in reported per-member per-month costs (PMPMCs) and assess how these have been qualitatively interpreted.

METHODS: Systematic PubMed searches (01/2003-10/2018) were conducted to identify published US peer-reviewed BIAs that reported PMPMCs associated with the introduction of new pharmacotherapies. All costs were inflation-adjusted to 2018 US dollars. Distribution of reported outcomes (PMPMCs and BIA authors’ qualitative judgments of their results) and study characteristics (indication and publication year) were analyzed descriptively.

RESULTS:We identified 49 BIAs reporting 133 PMPMC estimates (some studies reported PMPMCs of multiple populations/indications/payers/years). Of these 49 studies, 55% were published in/after 2015 and 86% provided qualitative conclusions of BIA results. Most BIAs were conducted in oncology (33%), autoimmune disease (14%), or cardiovascular disease (10%). The median (interquartile range [IQR]) PMPMC across all 133 estimates was $0.01 ($0.00-$0.04). Among PMPMC estimates ≥$0 (n=104), 80% were reported along with interpretations. “Modest,” “small,” or “minimal” were the most commonly used terms to describe result acceptability ($0.00≤PMPMC<$0.01: 27%, 36%, 23%; $0.01≤PMPMC<$0.04: 34%, 31%, 9%; PMPMC ≥$0.04, 33%, 0%, 13%, respectively).

CONCLUSIONS: Published PMPMC estimates are clustered within a few cents (median: $0.01; IQR: $0.00-$0.04) and have been invariably interpreted as acceptable, despite a 4-factor change separating the median from the 75th percentile PMPMC. Without pre-specified benchmarks, BIA authors should report actual values and refrain from providing qualitative judgement regarding PMPMC acceptability. Future research and guidance are needed to develop a set of US benchmarks to help interpret BIA results.