Validation of intracranial hemorrhage in the Norwegian Patient Registry

Abstract

Objectives: Administrative health registries need to have accurate diagnoses and sufficient coverage in the population they serve in order to be useful in research. In this study, we investigated the proportion of discharge diagnoses of intracranial hemorrhage (ICH) that were coded correctly in the Norwegian Patient Registry (NPR). Materials and Methods: We reviewed the electronic medical records and diagnostic imaging of all admissions to St. Olavs University Hospital, Trondheim, Norway, between January 1, 2008, to December 31, 2014, with a discharge diagnosis of ICH in the NPR, and estimated positive predictive values (PPVs) for primary and secondary diagnoses. Separate calculations were made for inpatient and outpatient admissions. Results: In total, 1,419 patients with 1,458 discharge diagnoses of ICH were included in our study. Overall, 1,333 (91.4%) discharge diagnoses were coded correctly. For inpatient admissions, the PPVs for primary discharge codes were 96.9% for hemorrhagic stroke, 95.3% for subarachnoid hemorrhage, and 97.9% for subdural hemorrhage. The most common cause of incorrect diagnosis was previous stroke that should have been coded as rehabilitation or sequela after stroke. There were more false-positive diagnoses among outpatient consultations and secondary diagnoses. Conclusion: Coding of ICH discharge diagnoses in the NPR is of high quality, showing that data from this registry can safely be used for medical research.