Validation of an ICD-10-based algorithm to identify stillbirth in the Sentinel System

Area of Science:

• Public Health
• Health Informatics
• Epidemiology

Background:

• Accurate identification of stillbirth events is crucial for public health surveillance and monitoring medical product safety.
• Electronic healthcare data offers a potential resource for large-scale epidemiological studies, but requires robust case-finding methods.
• Existing methods for identifying stillbirths in electronic health records may lack sufficient accuracy or efficiency.

Purpose of the Study:

• To develop and validate an algorithm using International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis codes to identify stillbirth cases.
• To assess the accuracy of the developed algorithm in identifying definite or probable stillbirth events using electronic healthcare data.
• To evaluate the agreement between claims data and medical charts for key event characteristics like date and gestational age.

Main Methods:

• A retrospective study utilized claims data from multiple healthcare systems within the Sentinel Distributed Database.
• Algorithms were developed based on ICD-10-CM codes for females aged 12-55 years between July 2016 and June 2018.
• A sample of medical charts underwent physician adjudication to confirm stillbirth events and determine positive predictive values (PPVs).

Main Results:

• Of 110 potential cases, 54 were confirmed as stillbirth events by adjudicators.
• The highest performing algorithm achieved a PPV of 82.5% (95% CI, 70.9%-91.0%).
• High agreement (≥90% within 7 days) was observed between claims data and medical charts for outcome date and gestational age.

Conclusions:

• Electronic healthcare data, when analyzed with a validated ICD-10-CM based algorithm, can effectively identify stillbirth cases.
• This approach demonstrates utility for signal detection of potential adverse effects of medical products on pregnancy outcomes.
• The developed algorithm provides a reliable tool for enhancing pharmacovigilance and public health monitoring related to stillbirth.