Economic Evaluation of the Use of Artificial Intelligence in Stress Echocardiography: Cost-Consequence and Cost-Effectiveness Analysis

OBJECTIVES: To assess the economic impact of Artificial Intelligence (AI)-assisted stress echocardiography (SE) in the National Health Service (NHS) diagnostic pathway for coronary artery disease (CAD) using cost-effectiveness analysis and cost-consequence analysis alongside a multicenter randomized controlled trial. METHODS: This evaluation was embedded in the PROTEUS trial, a 20-hospital NHS randomized controlled trial enrolling 2213 patients with suspected CAD undergoing SE. Patients were randomized to standard care (control) or AI-assisted SE (intervention). Outcomes were assessed using EuroQol 5-dimension 5-level (EQ-5D-5L) and Seattle Angina Questionnaire (SAQ-7), with quality-adjusted life-years (QALYs) estimated over 6 months. NHS costs were obtained from a costing study. Cost-consequence analysis compared within- and between-group outcomes. Cost-effectiveness analysis used a decision tree incorporating AI cost scenarios and clinician time savings. Probabilistic sensitivity analysis assessed uncertainty. RESULTS: Diagnostic accuracy was high in both groups (97.2% control; 96.9% intervention). SAQ-7 scores improved significantly within both groups (P < .001), with no between-group differences. EQ-5D-5L showed no significant within-group changes, except self-care (control: P = .017; intervention: P = .032), and no between-group differences. QALYs were similar (0.392 intervention vs 0.390 control; P = .818). Mean NHS costs were £376.72 (intervention) vs £366.08 (control). In the base case without AI costs, the incremental cost-effectiveness ratio was £6939 per QALY. The intervention remained cost-effective for AI costs up to £35.35 per case or £45.93 with clinician time savings. Sensitivity analyses supported the results. CONCLUSIONS: AI-assisted SE may be a cost-effective addition to NHS CAD diagnostics, with value driven by costs, accuracy, and efficiency.