Identifying Ectopic Pregnancy in a Large Integrated Health Care Delivery System: Algorithm Validation.

Area of Science:

• Reproductive Health
• Epidemiology
• Health Informatics

Background:

• Electronic health records (EHRs) are crucial for public health surveillance.
• Accurate case ascertainment is vital for ectopic pregnancy (EP) surveillance.
• No studies have validated EHR-based EP case-finding algorithms.

Purpose of the Study:

• To assess the validity of an enhanced algorithm for ectopic pregnancy (EP) case ascertainment.
• To compare the enhanced algorithm's performance against a previously validated method.
• To improve the accuracy of EP identification in electronic health databases.

Main Methods:

• Reviewed 500 women's medical records (2009-2018) with potential EP.
• Developed an enhanced algorithm using ICD-10 codes, telephone visits, and excluding abdominal-only codes.
• Calculated sensitivity, specificity, PPV, NPV, Youden index, and F-score.

Main Results:

• The enhanced algorithm achieved 97.6% sensitivity and 92.9% PPV for EP.
• Overall performance metrics (Youden index 82.5%, F-score 95.2%) demonstrated high accuracy.
• The enhanced algorithm outperformed the previous version in sensitivity and NPV.

Conclusions:

• The enhanced algorithm demonstrates adequate performance for EP case ascertainment in integrated healthcare databases.
• This algorithm offers improved capture of true ectopic pregnancy cases compared to prior methods.
• The findings support the use of this enhanced algorithm in future epidemiological studies.