Validating peritonsillar abscess drainage rates using the Pediatric hospital information system data.

Abstract

To evaluate the accuracy of the Children's Hospital Association's Pediatric Health Information System (PHIS) registry data in determining surgical drainage rates as compared to a chart review on the same cohort of children with diagnosis of peritonsillar abscess. Retrospective analysis. Our analysis included 200 children, ages 2 to 17 years, treated for a peritonsillar abscess from 2011 to 2016. The primary outcome was to determine the sensitivity, specificity, predictive values, receiver operating characteristics (ROC), and likelihood ratios of surgical drainage rates comparing the PHIS database to manual chart review of the same patients. One hundred and fifteen (58%) children underwent drainage by chart review, whereas 87 (44%) had a drainage procedure by PHIS data. Age was a significant predictor of abscess drainage by chart review (age coefficient = 0.10; standard error = 0.04; 2 = 5.8; P = 0.02; odds ratio = 1.1; 95% confidence interval [CI] = 1.01-1.19). When using the chart review as the reference value for surgical drainage, the PHIS data had a sensitivity of 76% and 100% specificity. The positive and negative predictive values were 100% and 75%, respectively. The ROC area was 0.88 (95% CI, 84 to 92). Cases that lacked a clear procedure note composed the false negative cases in the PHIS. At our institution, the PHIS Administrative data was adequate at predicting surgical drainage of the peritonsillar abscess when compared to chart review. 4 Laryngoscope, 130:238-241, 2020.