Positive predictive value of congenital heart disease diagnostic codes in the Danish National Patient Registry

Background: Congenital heart disease (CHD) is a public health concern as the most common birth defect with growing prevalence in adults. Valid CHD data are critical in clinical epidemiological CHD research. Objectives: To evaluate the positive predictive value (PPV) of having a CHD diagnosis and related surgery registered in the population-based Danish National Patient Registry (DNPR) according to the International Classification of Diseases, 8th and 10th revision, using predefined exclusion criteria to increase diagnostic validity. Methods: Using medical record review as reference standard, we computed the PPV of registered diagnostic codes for CHD and CHD subtypes including 95% confidence intervals (CIs) as well as the PPV of presence of CHD-related surgery. A sample of 480 patients diagnosed with CHD at any hospital in the Central Jutland region, during 1977–2013, was randomly selected for the medical record review. Reviews were performed by two paediatric cardiology nurses, blinded to the specific CHD diagnosis, and a paediatric cardiologist was consulted in cases of uncertainty. Results: We located medical records of 463 of the 480 patients included (96%), and found an overall PPV of the CHD diagnosis of 95.5 (95%CI: 93.3–97.1). Subtype specific PPVs were generally lower. In example PPV for atrial septal defect was 69%. It increased to 78% after inclusion of persistent foramen ovale in the reference standard subtype definition. The overall PPV of CHD-related surgery codes was 92.7% Conclusions: The overall PPV of CHD diagnoses in the DNPR is high. Data indicated lower PPVs for CHD subtypes.