Sensitivity of Administrative Coding for Heart Disease in Pregnancy

Area of Science:

• Cardiology
• Obstetrics
• Health Informatics

Background:

• Heart disease in pregnancy is a significant cause of maternal mortality globally.
• Accurate data is crucial for understanding the true burden of pregnancy-related heart conditions.
• Electronic medical records offer potential but require validation.

Purpose of the Study:

• To evaluate the accuracy of administrative coding (ICD-10-AM) for identifying pregnancies complicated by heart disease.
• To assess the sensitivity of coding during childbirth admissions in an Australian tertiary hospital.

Main Methods:

• Compared ICD-10-AM codes from hospital administrative data against a cardiologist-adjudicated standard.
• Utilized a clinical database of 155 women receiving specialized cardio-obstetric care.
• Conducted case-level and individual code-level comparisons.

Main Results:

• Overall sensitivity for detecting any cardiac disease was 66.5%.
• Only 25.2% of cases had complete diagnostic matching; 41.3% were partially or incorrectly coded.
• 33.5% of cases lacked any cardiac-related code, with low sensitivity for arrhythmias (19%) and un-coded congenital heart disease.

Conclusions:

• Administrative coding significantly underestimates the prevalence of heart disease in pregnancy.
• Coding limitations, such as capturing only actively managed diagnoses, contribute to underestimation.
• Epidemiological studies and service planning require enhanced case-ascertainment strategies for accuracy.