Atrial fibrillation prevalence and incidence varies by population definition: Association with co-morbidity profiles and multimorbidity in a large United States adult cohort

Background Identification of published data on prevalent/incidence of atrial fibrillation/flutter (AF) often relies on inpatient/outpatient claims, without consideration to other types of healthcare services and pharmacy claims. Purpose To examine AF prevalence/incidence and associated individual comorbidity and multi-morbidity profiles for a large US adult cohort spanning across a wide age range for both males/females based on both medical/pharmacy claims. Methods We studied a population of 8,343,992 persons across many geographical areas in the U.S. continent from 1 January /2016 to 31 October 2019. The prevalence and incidence of AF were comparatively analyzed for different healthcare parameters. Results Based on integrated medical and pharmacy claims, AF prevalence was 12.7% in the elderly population (> 65 years) and 0.9% in the younger population (< 65 years). These prevalence rates are different from estimates provided by the US CDC for those aged > 65 years (9%) and age < 65 years (2%); thus, the prevalence is under-estimated in the elderly population and over-estimated in the younger population. The incidence ratios for elderly females relative to younger females was 15.07 (95%CI 14.47-15.70), a value that is about 50% higher than for elderly males (10.57 (95%CI 10.24-10.92)). Comorbidity risk profile for AF identified on the basis of medical and pharmacy criteria varied by age and sex. The proportion with multimorbidity (defined as ≥2 long term comorbidities) was 10-12%. Conclusion Continued reliance only on outpatient and inpatient claims greatly underestimates AF prevalence and incidence in the general population by over 100%. Multimorbidity is common amongst AF patients, affecting approximately 1 in 10 patients. AF patients with 4 or more co-morbidities captured 20 to 40% of the AF cohorts depending on age groups and prevalent or incident cases. Our proposed methodology can guide future analysis of quality/cost of care for progressive medical conditions at the population level.