Estimating the cost impact of atrial fibrillation using a prospective cohort study and population-based controls

• Published in: BMJ open Vol. 13; no. 9; p. e072080
• Main Authors: Aebersold, Helena, Foster-Witassek, Fabienne, Serra-Burriel, Miquel, Brüngger, Beat, Aeschbacher, Stefanie, Beer, Jürg-Hans, Blozik, Eva, Blum, Manuel, Bonati, Leo, Conen, David, Conte, Giulio, Felder, Stefan, Huber, Carola, Kuehne, Michael, Moschovitis, Giorgio, Mueller, Andreas, Paladini, Rebecca E, Reichlin, Tobias, Rodondi, Nicolas, Springer, Anne, Stauber, Annina, Sticherling, Christian, Szucs, Thomas, Osswald, Stefan, Schwenkglenks, Matthias
• Format: Journal Article
• Language: English
• Published: London BMJ Publishing Group LTD 14.09.2023; BMJ Publishing Group
• Series: Original research
• Subjects: Algorithms; Cardiac arrhythmia; Codes; Cohort analysis; Cost estimates; Diagnosis related groups; DRGs; Health care expenditures; Health Economics; Health insurance; Illnesses; Insurance claims; Outpatient care facilities; Patients; Population; Population-based studies; Reimbursement; United States > US; Switzerland; Europe
• ISSN: 2044-6055; 2044-6055
• DOI: 10.1136/bmjopen-2023-072080

Aims Atrial fibrillation (AF) costs are expected to be substantial, but cost comparisons with the general population are scarce. Using data from the prospective Swiss-AF cohort study and population-based controls, we estimated the impact of AF on direct healthcare costs from the Swiss statutory health insurance perspective. Methods Swiss-AF patients, enrolled from 2014 to 2017, had documented, prevalent AF. We analysed 5 years of follow-up, where clinical data, and health insurance claims in 42% of the patients were collected on a yearly basis. Controls from a health insurance claims database were matched for demographics and region. The cost impact of AF was estimated using five different methods: (1) ordinary least square regression (OLS), (2) OLS-based two-part modelling, (3) generalised linear model-based two-part modelling, (4) 1:1 nearest neighbour propensity score matching and (5) a cost adjudication algorithm using Swiss-AF data non-comparatively and considering clinical data. Cost of illness at the Swiss national level was modelled using obtained cost estimates, prevalence from the Global Burden of Disease Project, and Swiss population data. Results The 1024 Swiss-AF patients with available claims data were compared with 16 556 controls without known AF. AF patients accrued CHF5600 (EUR5091) of AF-related direct healthcare costs per year, in addition to non-AF-related healthcare costs of CHF11100 (EUR10 091) per year accrued by AF patients and controls. All five methods yielded comparable results. AF-related costs at the national level were estimated to amount to 1% of Swiss healthcare expenditure. Conclusions We robustly found direct medical costs of AF patients were 50% higher than those of population-based controls. Such information on the incremental cost burden of AF may support healthcare capacity planning.