Statistical analysis
Detailed age treatment effect was conducted for several groups, namely,
< 45, 45-54, 55-64, 65-74, 75-84 and > 85
years. AF prevalence rates (%) and incidence rates (per 1000
person-years) were computed for individual and combined medical and
pharmacy service criteria. Duplication was eliminated whenever 2 or more
criteria are integrated. This was performed via use of unique,
non-duplicative identifiers for patients in the claim system.
Comparative analyses were performed relative to the youngest age group
(< 45 years). In these instances, prevalence and incidence
rate ratios were calculated together with 95% confidence intervals.
Analysis of individual comorbidity profiles was performed for each
comorbid condition as a binary outcome with gender and age as input
variables in logistic regression models. Computations were made for
prevalent and incident cases (% of total prevalent or incident cases)
as a function of AF + multi-comorbidity count. Cumulative prevalence and
incidence were additionally performed. Multi-morbidity were defined in
the conventional way as AF + 1 or more-comorbidity (i.e.,> 2 chronic conditions) as well as AF + 2 or more
co-morbidities. All descriptive and inferential statistical analyses
were conducted using the Statistical Analysis Software(7).