P104: Can patients suffering traumatic cardiac arrest be identified using the National Ambulatory Care Reporting System (NACRS) database?

Abstract

Introduction: Trauma is a common cause of mortality across all age groups and is projected to become the third greatest contributor to global disease burden. Recent studies have demonstrated that survival from traumatic cardiac arrest (TCA) is more favourable than once believed and further research on this population is being encouraged. Currently, it is unclear whether existing databases, such as the National Ambulatory Care Reporting system (NACRS), which includes all emergency department visits, could be used to identify TCAs for population-based studies. We aimed to determine the accuracy of NACRS administrative codes in identifying TCA patients. Methods: This retrospective validation study used data acquired from NACRS and our institutional Patient Care System. We identified a number of International Classification of Diseases, tenth revision (ICD-10) diagnostic, procedural and cause of injury codes that we hypothesized would be consistent with TCA. NACRS was subsequently searched for patients meeting the diagnostic code criteria (January 1 - December 31, 2015). The following inclusion criteria were: an eligible ICD-10 diagnostic code or a qualifying Canadian Classification of Health Interventions (CCI) procedure code and an eligible ICD-10 external cause of injury code. Electronic medical records for these patients were then reviewed to determine whether true TCAs had occurred. Results: Eighty-five patients met the inclusion criteria and one was excluded from analysis due to inaccessible health records, leaving 84 patients eligible for chart review. Overall, 55% (n = 46) of patients were found to have true TCA, 35% (n = 29) sustained a cardiac arrest of non-traumatic etiology and 11% (n = 9) were considered “unclear” (i.e. could not determine whether it was a true TCA based on the medical records). We found that true TCA patients were most accurately identified using a combination of ICD-10 CA cardiac arrest and external cause of injury codes (Positive predictive value: 70.6%, 95% CI 46.9-86.7). Conclusion: TCA patients were identified with moderate accuracy using the NACRS database. Further efforts to integrate specific data fields for TCA cases within existing population databases and trauma registries is necessary to facilitate future studies focused on this patient population.