Analysis of prevalent and incident cases
Prevalent cases utilized members who contributed at least six months of
benefits with the health plans to examine existing and new onsets of AF
cases(6). The analysis of incident cases utilized members with at least
30 months of benefits with the health plans (suppl fig S1). This period
ensured at least six months of benefits similar to members in the
prevalence study together with allowing a period of 24 months without
filing any medical and pharmacy claims. Piccini et al(2) used a similar
methodology for the analysis of incident cases in administrative
databases.
For inclusion, prevalent or incident cases had to contribute two or more
medical claims and/or two or more medication claims as detailed in suppl
fig S1. Prevalence was calculated over the period of the study with the
denominator equal to the population count at the beginning of the study
and the numerator to the unique persons filing claims during the
analysis period as specified previously. An incidence rate was computed
for new cases, with the denominator equal to the sum of person-years
contributed by those in the incidence study with at least 30 or more
months of benefits. The count and rates of prevalent and incident cases
were computed for individual medical criteria (ER, hospital admission,
office visit, office observation, or outpatient procedure), combined
medical (ER, hospital admission, office visit, office observation, or
outpatient procedures), all claims with AF ICD 10 codes, pharmacy
criteria (anticoagulant or rhythm control medications), and combined
medical/pharmacy.
The comorbidity profiles for prevalent and incident AF targets were
analyzed for congestive heart failure, hypertension, diabetes mellitus,
ischemic stroke, transient ischemic attack, thrombo-embolic events,
myocardial infarction, peripheral artery disease, valvular disease,
coronary artery disease, obstructive sleep apnea, and chronic kidney
disease. The details of diagnosis codes for the comorbidity profiles are
given in suppl table S3. We also determined the prevalence and incidence
of multimorbidity (defined as ≥2 chronic long-term conditions – but for
the purposes of this study, we also looked at AF with ≥2 chronic long
term conditions).