Open AccessSensitivity and Specificity of Administrative Medical Coding for Pediatric Eosinophilic Esophagitis
Author(s) -
Robson Jacob,
Korgenski Kent,
Parsons Karyn,
McClain Amber,
Barbagelata Carlos,
AllenBrady Kristina,
Patel Raza,
O'Gorman Molly,
Peterson Kathryn,
Guthery Stephen
Publication year - 2019
Publication title -
journal of pediatric gastroenterology and nutrition
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.206
H-Index - 131
eISSN - 1536-4801
pISSN - 0277-2116
DOI - 10.1097/mpg.0000000000002340
Subject(s) - medicine , eosinophilic esophagitis , retrospective cohort study , cohort , medical record , diagnosis code , pediatrics , epidemiology , false positive paradox , coding (social sciences) , population , disease , statistics , mathematics , environmental health
Objectives: Eosinophilic esophagitis (EoE) is a delayed‐type hypersensitivity with increasing rates among pediatric populations. Although studies have used International Classification of Diseases (ICD) coding to define local cohorts and report disease epidemiology, the accuracy of the EoE ICD code for pediatric EoE is unknown. Methods: We searched the Intermountain Healthcare Database for pediatric cases with the EoE ICD code over a 5‐year period. We cross‐referenced these results with a recently published pediatric EoE cohort from the same region and period, where incident cases were identified via retrospective review of pathology reports and medical records. Using the retrospective review cohort as the reference standard, we evaluated the accuracy of the EoE ICD code. Results: Via retrospective review, we identified 1129 new pediatric EoE cases in the Intermountain Healthcare system over 5 years. Six hundred ten of these had the EoE ICD code associated with their chart. Out of 878,872 unique pediatric records in the Intermountain Healthcare system, 219 had the EoE ICD code incorrectly applied. The specificity of the EoE ICD code in children was 99%, but sensitivity and positive predictive value were 61% and 79%, respectively. Conclusions: The EoE ICD code has strengths and weaknesses in pediatrics. The EoE ICD code is specific, with few false positives across a large population, but not sensitive. The low sensitivity is likely multifactorial and requires further evaluation. Compared to retrospective chart review, which allows for application of clinicopathologic EoE diagnostic criteria, sole use of ICD codes results in underascertainment of EoE cases and key misclassifications.