Purpose: Routinely collected clinical data are a valuable resource for epidemiological research in infectious diseases. We examined the validity of the ICD-10 diagnosis code K75.0 for pyogenic liver abscess (PLA) from hospital discharge registries.
Patients and methods: This validation study was conducted in the North Denmark Region, using data from Aalborg University Hospital and the North Denmark Regional Hospital, along with their smaller regional satellite hospitals. The study period extended from January 1, 2010, to June 30, 2022, covering a catchment population of approximately 590,000 inhabitants. We identified patients with a first diagnosis (primary or secondary) of PLA (ICD-10 code K75.0) recorded in the Danish National Patient Registry and estimated the positive predictive value (PPV) of the PLA diagnosis using medical records as the reference standard. Subanalyses of PPV were conducted based on the department setting (emergency, medical, or surgical).
Results: A total of 297 patients received an ICD-10 diagnosis code of K75.0 during the study period. Five (2.0%) patients were excluded due to initial hospitalization outside the North Denmark Region, and 67 (23%) were misclassified. The overall PPV for the K75.0 diagnosis code during the study period was 77% (95% CI: 72-82%). The highest PPV, 88% (95% CI: 81-93%), was observed in patients diagnosed in medical departments, while the lowest PPV, 56% (95% CI: 30-80%), was observed in patients diagnosed in emergency wards. The PPV for surgical departments was 69% (95% CI: 61-77%).
Conclusion: The overall PPV of the ICD-10 diagnosis code K75.0 for PLA was 77%. Variability in PPVs across departments suggested differences in diagnostic accuracy, with medical departments demonstrating the highest PPV.
Keywords: epidemiology; hepatic abscess; positive predictive value; pyogenic liver abscess; validation.
© 2024 Dudina et al.