The importance of considering both primary and secondary diagnostic codes when using administrative health data to study acute coronary syndrome epidemiology (ANZACS-QI 47)

Abstract

Abstract Aims Routinely collected health administrative data have become an important data source for investigators assessing disease epidemiology. Our aim was to investigate the implications of identifying acute coronary syndrome (ACS) events in New Zealand (NZ) national hospitalization data using either the first (primary) or subsequent (secondary) codes. Methods and results Using national health datasets, we identified all NZ hospitalizations (2014–16) for patients ≥20 years with a primary or secondary International Classification of Diseases 10th Revision, Australian Modification (ICD10-AM) ACS code. Outcomes included 1-year all-cause and cause-specific mortality, hospitalized non-fatal myocardial infarction, heart failure, stroke, or major bleeding, and a composite comprising these outcomes. Of 35 646 ACS hospitalizations, 78.5% were primary and 21.5% secondary diagnoses. Compared to primary coding, patients with a secondary diagnosis were older (mean 77 vs. 69 years), more likely to be females (48% vs. 36%), had more comorbidity, and were less likely to receive coronary angiography or revascularization. Higher adverse event rates were observed for the secondary diagnosis group including a three-fold higher 1-year mortality (40% vs. 13%) and two-fold higher composite adverse outcome (54% vs. 26%). The use of primary codes alone, rather than combined primary and secondary codes, resulted in overestimation of coronary angiography and revascularization rates, and underestimation of the 1-year case fatality (13.1% vs. 19.0%) and composite adverse event rate (26% vs. 32%). Conclusion Patient characteristics and outcomes of ACS events recorded as primary vs. secondary codes are very different. These findings have important implications for designing studies utilizing ICD10-AM codes.