Validity of a Hypertrophic Cardiomyopathy Diagnosis in Adult Patients in the Danish National Patient Register

PURPOSE: To assess the positive predictive value (PPV) of the diagnostic codes for hypertrophic cardiomyopathy (HCM) in the Danish National Patient Register (DNPR) and identify factors affecting the PPV.

PATIENTS AND METHODS: We randomly sampled 200 patients registered in DNPR with the diagnostic codes DI421 (obstructive) or DI422 (non-obstructive) HCM, from Eastern Denmark, between December 1st, 2017, and September 16th, 2021. We assessed patients' medical records and classified whether the diagnosis of HCM was correct, incorrect, or uncertain according to the European Society of Cardiology (ESC) guidelines. Overall PPVs were estimated for all hospitals irrespective of specialty and for dedicated centers for inherited cardiac diseases.

RESULTS: Of the 200 patients, seven were excluded; six were younger than 18 years when initially diagnosed and one adult patient had requested that their medical records be unavailable for research purposes. Of the 193 patients (median age 61 years, Interquartile range (IQR) 50-70, 40% female) with medical records available for assessment, 148 (77%) patients were registered correctly, 41 (21%) were incorrectly registered and for four (2%) patients there was insufficient data to determine if the diagnosis was correct. The overall PPV was 0.77 (95% confidence interval (CI) 0.70-0.82). The PPV for patients diagnosed in dedicated centers for inherited cardiac diseases was 0.91 (95% CI 0.84-0.95). For patients diagnosed in a dedicated center for inherited cardiac diseases with more than one clinical visit, the PPV increased to 0.99 (95% CI 0.93-1), however with the risk of a reduced sensitivity as these patients constituted only 52% of the correctly registered patients.

CONCLUSION: The overall PPV of the HCM diagnostic codes in the DNPR was 0.77 (95% CI 0.70-0.82). The validity of the diagnosis was higher at dedicated centers for inherited cardiac diseases. These findings may prove useful for future register-based research.