Introduction
Atrial fibrillation (AF) is the commonest cardiac rhythm disorder
globally, and confers a large healthcare burden from mortality and
morbidity from stroke/systemic thromboembolism, heart failure, dementia
and hospitalisations(1). Accurate, population-level data that can enable
ongoing monitoring of AF epidemiology, quality of care at affordable
cost, and complications are needed.
In the US, data on AF prevalence/incidence are typically identified in
administrative medical databases using inpatient/outpatient/physician
claims with ICD 9/10 codes(2, 3). Nonetheless, identification does not
include claims from other healthcare services (e.g., nursing skilled
facilities) or procedures peculiar to AF (e.g., catheter ablation). It
is also common for fully insured individuals with medical/pharmacy
benefits, who are on medications such as anticoagulants and heart rhythm
control, not to have medical claims with AF ICD 9/10 codes. Indeed, a
Canadian study by Tu et al(4) voiced similar concerns and found that an
algorithm relying on a combination of criteria yielded higher prevalence
than those simply based on hospital admissions, emergency room visits or
office visits. Collectively, the aforementioned issues may underestimate
our knowledge of the true AF prevalence and incidence and the associated
co-morbidity profiles.
In a systematic review of validated methods for identifying AF using
administrative data, Jensen et al(5) found that 10 of 16 studies used
only inpatient data. We hypothesized that prevalent/incidence data would
vary by type of population studied, and associated comorbidities would
vary accordingly. The aim of this study was to examine AF
prevalence/incidence and associated comorbidity profiles for a large US
adult cohort spanning across a wide age range (< 65 and> 65 years) for both males/females based on both
medical/pharmacy claims. The prevalence and incidence were comparatively
analyzed for different healthcare parameters (i.e., emergency room
visit, hospital admission, office visit, office observation,
cardioversion, catheter observation, all medical claims, anticoagulant
medication, heart rhythm control medication, both medical and pharmacy
claims). Finally, we assessed co-morbidity profiles, AF + co-morbidity
counts, and multi-morbidity in AF cohorts.