Fractional flow reserve (FFR) versus angiography in guiding management to optimise outcomes in non-ST segment elevation myocardial infarction (FAMOUS-NSTEMI) developmental trial: cost-effectiveness using a mixed trial- and model-based methods

BACKGROUND In the Fractional flow reserve (FFR) versus angiography in guiding management to optimise outcomes in non-ST elevation myocardial infarction (FAMOUS) clinical trial, FFR was shown to significantly reduce coronary revascularisation, compared to visual interpretation of standard coronary angiography without FFR. We estimated the cost-effectiveness from a UK National Health Service perspective, based on the results of FAMOUS. METHODS A mixed trial- and model-based approach using decision and statistical modelling was used. Within-trial (1-year) costs and QALYs were assembled at the individual level and then modelled on subsequent management strategy [coronary artery bypass graft (CABG), percutaneous coronary intervention (PCI) or medical therapy (MT)] and major adverse coronary events (death, MI, stroke and revascularisation). One-year resource uses included: material, hospitalisation, medical, health professional service use and events. Utilities were derived from individual EQ5D responses. Unit costs were derived from the literature. Outcomes were extended to a lifetime on the basis of MACE during the 1st year. Costs and QALYs were modelled using generalized linear models whilst MACE was modelled using logistic regression. The analysis adopted a payer perspective. Costs and outcomes were discounted at 3.5 %. RESULTS Costs were related to the subsequent management strategy and MACE whilst QALYs were not. FFR led to a modest cost increase, albeit an imprecise increase, over both the trial [£112 (-£129 to £357)] and lifetime horizons [£133 (-£199 to £499)]. FFR led to a small, albeit imprecise, increase in QALYs over both the trial [0.02 (-0.03 to 0.06)] and lifetime horizons [0.03 (-0.21 to 0.28)]. The mean ICER was £7516/QALY and £4290/QALY over the trial and lifetime horizons, respectively. Decision remained high; FFR had 64 and 59 % probability of cost-effectiveness over trial and lifetime horizons, respectively. CONCLUSIONS FFR was cost-effective at the mean, albeit with considerable decision uncertainty. Uncertainty can be reduced with more information on long-term health events.