Validation of hospital discharge coding for neonatal abstinence syndrome

Abstract

Aim: To validate the diagnostic discharge coding of neonatal abstinence syndrome (NAS) (International Classification of Diseases [ICD]-10-AM, P96.1).

Methods: Retrospective record review of infants diagnosed with NAS (P96.1) in a non-tertiary Australian hospital between 2000 and 2016. NAS criteria were predetermined to include the following: (i) maternal opioid use; (ii) infant requiring NAS medication and (iii) at least one score of ≥8 on the Finnegan Neonatal Abstinence Scoring Tool (FNAST).

Results: Of the 253 infants coded with P96.1, 82/146 (56%) opioid-exposed infants and 9/107(18%) infants exposed to non-opioid drugs only received withdrawal medication: sensitivity 56.2 (95% confidence interval: 47.7-64.3), specificity 91.6 (84.2-95.8%), positive predictive value (PPV) 90.1 (81.6-95.1%) and negative predictive value (NPV) 60.5 (52.5-68.0%) for all three criteria. Using the criterion of ≥1 FNAST score ≥8 resulted in 58.0 (51.3-64.4%) sensitivity, 63.6 (40.8-82.0%) specificity, 94.4 (88.8-97.4%) PPV and 12.6 (7.3-20.6%) NPV for identifying need for NAS medications.

Conclusion: A diagnosis of P96.1 is highly specific and predictive but poorly sensitive for identifying opioid-exposed infants requiring medications for withdrawal.

Keywords: Discharge code; Neonatal abstinence syndrome; Opioids; Validation.