Background: Despite evidence for single gene–drug interactions (e.g. DPYD, CYP2C19, TPMT), pharmacogenomics (PGx) has yet to achieve routine NHS implementation. Existing trials have focused on isolated drug decisions, without addressing the system-wide clinical, economic, and operational benefits of integrating panel PGx into real-world acute care. The PHOENIX trial was designed to generate causal, health-economic, and equity evidence for panel-based pre-emptive PGx within NHS workflows.

Methods: PHOENIX is a pragmatic, two-group, randomised controlled trial recruiting acutely admitted adults across all specialties in the Queen Elizabeth University Hospital, Glasgow. Participants are randomised (1:1) to a 15-gene PGx-guided prescribing arm or usual care. Primary outcomes include adverse drug events/treatment failure/mortality at 3-months, EQ-5DL and quality-of-life. Embedded cost-utility and cost-effectiveness analyses assess both value and equity.

Results: An automated recruitment and eligibility algorithm reduced screening workload by 92% (3.4hrs to 17 mins per patient), saving 174 person-days in 4-months and screen-fail to enroll ratio: 2.4:1. Among the first 625 participants, prescribing exposure was highly concentrated: six drug–gene pairs—Clopidogrel–CYP2C19, Omeprazole–CYP2C19, Ibuprofen–CYP2C9, Atorvastatin–SLCO1B1, Codeine–CYP2D6, and Flucloxacillin–HLA-B—accounted for 86% of all PGx interactions. CYP2C19 and CYP2D6 emerged as linchpin genes, influencing 17 drugs across cardiology, general medicine, surgery, and orthopaedics.