3.2 Incident heart failure hospitalizations by ethnicity 141 this study. First, there was the lower data coverage in 2012 and 2013 due to some loss of information from several centres during a data migration process. This data loss did not occur at random hence imputation was not a viable option. To get a complete picture of the prevalence of HF in Malaysia, outpatient visits for HF should ideally be included but data for outpatient and specialist clinics’ data was only available in aggregate form by ICD-10 codes which encompassed a broader scope of circulatory disease, not HF specifically. Another limitation was the underestimation of absolute numbers of hospitalizations because only a primary discharge diagnosis of HF was considered for inclusion and cases came from only MOH hospitals. Additionally, the use of overall population as denominator will also underestimate the rate of hospitalization. The primary discharge diagnosis was coded with ICD-10 code, however only the category level of the code was reliable, limiting the possibility of identifying the subtypes and aetiology of HF using the ICD-10 code. In addition, the lack of information on drug therapies meant that we were not able the quantify the effect of newer HF drug therapy on hospitalization during the study period. Nevertheless, this was the best available data that allowed uniform assessment of trend changes and this was not influenced by the selection criteria. Lastly, information on left ventricular function, comorbid conditions and other relevant risk factors were also not available for differentiating between aetiologies and subtypes of HF. CONCLUSION This population-based analysis has shown a steady decrease in incidence of HF hospitalizations over a period of ten years. Incidence was lower in women than men and persons of Indian ethnicity compared to other ethnic groups. These findings highlight disparities in occurrence of HF between sex and ethnic groups in the population that were attributable to underlying population risks and offers opportunities for targeted interventions in reducing the risk of de novo HF hospitalization.
RkJQdWJsaXNoZXIy MjY0ODMw