Population-based drug-related anaphylaxis in children and adolescents captured by South Carolina Emergency Room Hospital Discharge Database (SCERHDD) (2000-2002).

Journal Article (Journal Article)

PURPOSE: Anaphylaxis is a life-threatening condition; drug-related anaphylaxis represents approximately 10% of all cases. We assessed the utility of a statewide emergency department (ED) database for identifying drug-related anaphylaxis in children by developing and validating an algorithm composed of ICD-9-CM codes. METHODS: There were 1 314,760 visits to South Carolina (SC) emergency departments (EDs) for patients <19 years in 2000-2002. We used ICD-9-CM disease or external cause of injury codes (E-codes) that suggested drug-related anaphylaxis or a severe drug-related allergic reaction. We found 50 cases classifiable as probable or possible drug-related anaphylaxis and 13 as drug-related allergic reactions. We used clinical evaluation by two pediatricians as the 'alloyed gold standard'1 for estimating sensitivity, specificity, and positive predictive value (PPV) of our algorithm. RESULTS: ED-treated drug-related anaphylaxis in the SC pediatric population was 1.56/100,000 person-years based on the algorithm and 0.50/100,000 person-years based on clinical evaluation. Assuming the disease codes we used identified all potential anaphylaxis cases in the database, the sensitivity was 1.00 (95%CI: 0.79, 1.00), specificity was 0.28 (95%CI: 0.16, 0.43), and the PPV was 0.32 (0.20, 0.47) for the algorithm. Sensitivity analyses improved the measurement properties of the algorithm. CONCLUSIONS: E-codes were invaluable for developing an anaphylaxis algorithm although the frequently used code of E947.9 was often incorrectly applied. We believe that our algorithm may have over-ascertained drug-related anaphylaxis patients seen in an ED, but the clinical evaluation may have under-represented this diagnosis due to limited information on the offending agent in the abstracted ED records. Post-marketing drug surveillance using ED records may be viable if clinicians were to document drug-related anaphylaxis in the charts so that billing codes could be assigned properly.

Full Text

Duke Authors

Cited Authors

  • West, SL; D'Aloisio, AA; Ringel-Kulka, T; Waller, AE; Clayton Bordley, W

Published Date

  • December 2007

Published In

Volume / Issue

  • 16 / 12

Start / End Page

  • 1255 - 1267

PubMed ID

  • 17918773

International Standard Serial Number (ISSN)

  • 1053-8569

Digital Object Identifier (DOI)

  • 10.1002/pds.1502


  • eng

Conference Location

  • England