Age and sex differences in the sensitivity of ICD-10 diagnostic codes for identifying patients with out-of-hospital cardiac arrest in the British Columbia cardiac arrest registry: a retrospective cohort study.
Resuscitation · 2025-12-11 · Method & validation
Abstract
BACKGROUND: Cardiac arrest researchers frequently use administrative health databases to study out-of-hospital cardiac arrest (OHCA) incidence, but their sensitivity to identify OHCA is not well known. METHODS: We included Emergency Medical Services (EMS)-treated cases from the BC Cardiac Arrest Registry that survived to hospital admission between 2012 and 2016, and linked to their admission records in the hospital admissions database (Discharge Abstract Database). We calculated sensitivity as the proportion of cases with OHCA-related International Classification of Disease 10 (ICD-10) codes that were listed as a preadmission comorbidity or the primary reason for hospital stay (coded as arising preadmission). In cases without OHCA-related codes, we identified other recorded ICD-10 codes and diagnosis types. RESULTS: Of 6009 eligible OHCAs during the study period, 2032 (33.8 %) patients who survived to hospital admission were included. OHCA-related ICD-10 codes were recorded in 1357/2032 (66.8 %; 95 % CI 64.7, 68.8) of cases. Sensitivity did not differ across sex and age groups. Among cases without OHCA-related codes, acute myocardial infarction was the most frequently assigned pre-admission diagnosis (225/675; 33.3 %), coded as most responsible for hospital admission, and coronary artery disease-related ICD-10 codes were the most common pre-admission diagnoses (176/675, 26.1 %). CONCLUSION: One third of patients in our patient population suffering from OHCA and admitted to hospital were not coded as having a cardiac arrest. Using administrative databases to identify OHCA patients results in underestimation of cardiac arrest incidence, highlighting the need for ongoing dedicated OHCA registries.