The study comprised a retrospective audit of data from 118 patients diagnosed with paranoid
SZ (F20.0) and 71 patients with SAD (F25). We used the mean total costs as well as
partial cost, i.e., mean costs for medication products, mean personal costs and mean infrastructure
costs from each patient for the statistical analysis. We tested for differences in
the four variables between SZ and SAD patients using ANCOVA and confirmed the results
with bootstrapping.