The non-anticoagulation costs of atrial fibrillation management: findings from an observational study in NHS Primary Care

BACKGROUND Atrial fibrillation (AF) management represents a significant burden on the UK NHS. Understanding this burden will be important in informing future health care planning and policy development. AIM To describe the non-anticoagulation costs associated with AF management in routine UK clinical practice. MATERIALS PATIENTS AND METHODS A retrospective observational study of 825 patients with AF undertaken in eight UK primary care practices. Data collected from routine clinical and prescribing records of all eligible, consenting patients, for a period of up to 3 years. The first 12 weeks following diagnosis was defined as the 'initiation phase'; the period after week 12 was defined as the 'maintenance phase'. RESULTS Mean (SD) total cost of AF management was £941 (£1094) per patient in the initiation phase and £426 (£597) per patient-year in the maintenance phase. AF-related inpatient admissions contributed most to total costs; the mean (SD) total cost per patient in the initiation phase was £2285 (£900) for admitted and £278 (£252) for non-admitted patients. Mean maintenance phase costs per year were £1323 (£755) and £168 (£234), respectively, for admitted and non-admitted patients. Significant patient variables contributing to high cost in the initiation phase were hypertension and younger patient age, although only accounting for 6% of cost variability. Significant variables in the maintenance phase (18% of cost variability) were the presence of congestive heart failure, structural heart disease or diabetes and the frequency of day case admissions, ECGs and hospitalisations in the initiation phase. CONCLUSIONS Inpatient admissions contributed most to total AF management costs. Given the burden of hospital care, future work should focus on evaluating the appropriateness and reasons for hospitalisation in patients with AF and the factors affecting length of stay, with the aim of identifying opportunities to safely reduce inpatient costs. A number of significant patient characteristics and initiation phase variables were identified, which accounted for 18% of the variability in total maintenance phase costs. However, none of these could adequately predict high maintenance phase costs.