Epidemiology of Bloodstream Infections

Crude incidence rates (IR) were determined by dividing the number of BSI episodes with the population of Skåne. To increase comparability and to adjust for changes in population structure over time, age-standardised rates were estimated using the direct standardisation method and the 2013 European standard population [15 ]. Due to the focus on the ageing population, age-specific rates were described in 0–49, 50–64, 65–79 and ≥ 80 year strata. Crude rates were used for sex stratified IRs. Differences in IRs were estimated using incidence rate ratios (IRR). All IRs are presented as cases per 100,000 person-years, and IR and IRRs are presented with 95% confidence intervals.
To model change in overall IR over time, we fitted segmented regressions with the number of BSI episodes as the dependent variable, age and year as independent variables and the natural log of population as offset [16 (link)]. The results are expressed as annual percent change (APC) for each segment as well as the average annual percent change (AAPC) for the entire period [17 (link)]. For hypothesis testing of differences in the proportion of resistant isolates across sexes, the chi-square test was used. To test trends in the proportion of resistant isolates, Cochrane-Armitage trend tests were used, using Bonferroni correction for multiple testing. A p value of <0.05 was considered significant.