(185) Validity of Out-of-Hospital and In-Hospital Cardiac Arrest Algorithms in the Danish National Patient Registry

Background: Cardiac arrest is registered in the Danish National Patient Registry (DNPR) with the International Classification of Diseases 10th revision code I46. However, it does not distinguish between out-of-hospital cardiac arrest (OHCA) and in-hospital cardiac arrest (IHCA).

Objectives: To validate an algorithm to identify cardiac arrest subtypes in the DNPR.

Methods: From Aarhus University Hospital, Denmark, we sampled patients with a primary or secondary cardiac arrest discharge diagnosis during 2019–2023. The algorithm categorized these patients as OHCA if they (1) only had a single department course during their hospitalization or (2) had multiple department courses during their hospitalization but were discharged with a cardiac arrest diagnosis from the first department course. The algorithm categorized the remaining patients as IHCA. We randomly sampled 200 patients with algorithm-based OHCA (n=100) and IHCA (n=100). Using medical record review as the reference, we calculated positive predictive values (PPVs) with 95% confidence intervals (CIs).

Results: Cardiac arrest was confirmed in 192 of 200 cases, yielding a PPV for cardiac arrest overall of 96% (95% CI: 92–98%). The PPV was 87% (95% CI: 79–92%) for OHCA and 61% (95% CI: 51–70%) for IHCA. The results were robust in age and sex strata, as well as in a worst-case scenario excluding cardiac arrests categorized as probable, indicating that the cardiac arrest was likely but could not be confirmed due to insufficient information in the medical records [PPV=95% (95% CI: 90–97%) overall; 86% (95% CI: 78–91%) for OHCA; and 61% (95% CI: 51–70%) for IHCA].

Conclusions: The validity of the cardiac arrest diagnosis in the DNPR was overall high. Furthermore, our algorithm had a high PPV for OHCA but a poor PPV for IHCA.