Cost-effectiveness of screening for paroxysmal atrial fibrillation in patients undergoing echocardiography

Background Screening for atrial fibrillation is recommended for patients > 65 years on current guidelines. Targeted screening may be more efficient, however the appropriate location for screening programs has not been well defined. Our aim was to compare the cost-effectiveness of unselected electrocardiographic (ECG) screening for atrial fibrillation (AF), and selective screening based on an abnormal echocardiogram. Methods Two strategies of portable ECG screening for AF were compared in the base case of a hypothetical asymptomatic 65-year-old man (CHA(2)DS(2)-VASC = 3 based on hypertension and diabetes mellitus) with previous echocardiography but without a cause for AF (e.g. mitral valve disease, left ventricular (LV) dysfunction). With age-based screening (AgeScreen; 3% AF detection rate) all patients underwent ECG. With imaging-guided screening (ImagingScreen; 5% detection rate), only patients with left atrial (LA) volume >= 34 mL/m(2) and LA reservoir strain -18% underwent ECG screening. A Markov model was informed by published transition probabilities, costs and quality-adjusted life years (QALY). Costs, effects and incremental cost-effectiveness ratio were assessed for each screening strategy over a 20-year period. The willingness-to-pay threshold was $53 000/QALY. Results ImagingScreen dominated AgeScreen, with a lower cost ($54 823 vs $57842) and better outcome (11.56 vs 11.52 QALY over 20 years). Monte Carlo simulation demonstrated that 61% of observations were more efficacious with ImagingScreen, with cost below willingness to pay. The main cost determinants were annual costs of stroke or heart failure and AF detection rates. ImagingScreen was more cost-effective for AF detection rates up to 14%, and more cost-effective across a range of annual stroke ($24 000-$102 000) and heart failure ($4000-$12 000) costs. Conclusion In patients with a previous echocardiogram, AF screening of those with baseline clinical and imaging risk parameters is more cost-effective than age-based screening.