Acute coronary syndrome cohort definition: Troponin versus ICD-9-CM codes

Aim: To determine the accuracy of troponin laboratory results versus International Classification of Diseases clinical modification codes (ICD-9-CM) in identifying acute coronary syndrome (ACS) rule-out (R/O) patients who present to emergency departments (EDs). Materials & methods: Retrospective data analysis and chart review (to establish gold standard) were conducted on ED patients. Data retrieved from a clinical data warehouse were reviewed to identify patients with two or more troponins within 24 h of ED registration and ICD-9-CM codes consistent with ACS R/O. Results: Of 329 charts reviewed, 17 were determined to be ACS R/O. A total of 31 out of 329 (9.42%, 95% CI: 6.26-12.58%) had two or more troponins with a sensitivity of 100% (95% CI: 77.08-100%) and specificity of 95.51% (95% CI: 92.42-97.43%). A total of 32 out of 329 patients (9.73%, 95% CI: 6.53-12.93%) had R/O ICD-9-CM codes with a sensitivity of 76.47% (95% CI: 49.76-83.00%) and specificity of 93.91% (95% CI: 90.50-96.19%). All 17 gold-standard ACS R/O patients were identified using troponins while ICD-9-CM identified 13 out of 17. Conclusion: Clinical data (two troponins) availability is timelier and compares well with billing data (ICD-9-CM codes) in ACS R/O patient identification. Clinical data use may be generalized to identify other disease specific cohorts for clinical research.