OBJECTIVES

Pharmacogenomic (PGx) test results delivered by clinical decision support (CDS) alerts guide drug therapy, however, evidence of its value is lacking. Thus, we estimated the value of developing and implementing PGx-CDS alerts over a 20-year time horizon, using a cost-utility analysis from the perspective of Learning Health Systems (LHSs). To generate customizable estimates, we present modifiable parameters in a R Shiny app to guide development of alerts.

METHODS

We compared PGx-CDS alerts to no alerts in an age-based cohort of 500,000 individuals who had received preemptive PGx testing for four pharmacogenes (CYP2C19, CYP2C9, CYP4F2 and VKORC1) to guide potential therapy with clopidogrel for acute coronary syndrome (ACS) or warfarin for atrial fibrillation (AF). We identified literature-based payoffs for PGx testing, including incremental costs, incremental quality-adjusted life-years (QALYs), and risk differences for adverse events. Incident prescription use by age was estimated from IBM® MarketScan Research Databases 2015-2019. To mirror real-world practice, alert fatigue, start-up- and maintenance costs were included. A 3% annual discount was applied, and one-way and probabilistic sensitivity analyses were conducted.

RESULTS

Over 20 years, 3,173 alerts would be fired, at an average cost of $92/alert fired. The incremental cost of PGx-CDS alerts was $292,029, and the incremental QALYs were 10.29, producing an incremental cost effectiveness ratio (ICER) of $28,381/QALY gained in favor of alerting. For every two alerts fired for clopidogrel, one adverse event of non-fatal MI and one cardiovascular death would be averted; for every two alerts fired for warfarin, three events of bleeding and one event of clotting would be averted. The most influential parameter was the risk of initiating clopidogrel or warfarin therapy. Estimated ICERs were below $50,000/QALY in 99% of 1,000 probabilistic simulations.

CONCLUSIONS

Our results suggest that implementing PGx-CDS alerts reduces the number of drug-related adverse events in a cost-effective manner.