Proefschrift

CHAPTER 3.2 122 median with interquartile range (IQR) while categorical data were reported as frequencies and proportions. Crude incidence rates by age group, sex and calendar year were calculated by dividing the number of hospitalizations in each category against the respective base population numbers. National population counts by age and sex were extracted from vital statistics data provided by the Department of Statistics Malaysia. To evaluate trends, overall age- and sex-standardised incidence for each calendar year were standardised to the 2016 Malaysian population using direct standardisation. In direct standardisation, the expected incident HF hospitalizations were calculated by multiplying each age-sex specific incidence rate with the proportion of that said agesex group in the standard population. The standardised incidence rate was then obtained by summing the total expected cases across age categories and dividing by the total in the standard population.[14] An overall age- and sex-standardised incidence rates for each year was also calculated using WHO standard population with direct standardisation. Finally, to determine the strength and statistical significance of these trends, we used quasi-Poisson models, which accounts for overdispersion of the count data[15]. The population size for each age, sex and ethnicity-specific group were used as the offset variable in the regression. A sensitivity analysis was done to examine the impact of adding ICD-10 codes I11 (hypertensive heart disease with or without HF) and I13 (Hypertensive heart and renal disease with or without HF) when selecting for eligible cases of HF hospitalization. A separate sensitivity analysis was conducted to study the effect of data loss in 2012 and 2013 on the time trends, by excluding these two years from the quasi-Poisson model. All statistical analyses were performed using R statistical software version 3.6.1.[16] RESULTS Incidence of HF hospitalization A total of 105 399 hospital admissions for HF were included from 2007 to 2016. The absolute number of index hospitalizations has increased by 52.3% from 8191 to 12 472 cases annually in this ten-year period. The demographic characteristics of the HF admissions were reported in Table 1. The median age of patients remained stable, from 65.8 (IQR 56.2-74.0) in 2007 to 64.3 (IQR 55.5-73.1) years in 2016. Overall, 55.9%

RkJQdWJsaXNoZXIy MjY0ODMw