Objective: The aim of this validation study was to assess the completeness of the registrations of chest X-rays (CXR) in two different versions of the Danish National Patient Registry (DNPR).
Material and methods: We included electronic record data on CXR performed on patients aged 40 to 99 years from nine radiology departments covering 20 Danish hospitals. From each department, we included data from three randomly selected weeks between 2004 and 2011 (reference standard). In two versions of the DNPR from the State Serum Institute (SSI) and Statistics Denmark, respectively, we investigated the proportion of registered CXR compared to the reference standard. Furthermore, we compared the completeness of the recorded data according to the responsible department (main department).
Results: We identified 11,235 patients and 12,513 CXR in the reference standard. The data from the SSI contained 12,265 (98%) CXR, whereas the data from Statistics Denmark comprised 9,151 (73.1%) CXR. The completeness of the SSI data was fairly constant across years, radiology departments, medical specialties, and age groups. The data from Statistics Denmark was almost complete in 2011 (95.8%). However, for the remaining study period, the data with radiology departments registered as the main department were lacking in the version from Statistics Denmark, and so the overall completeness was 73.1%.
Conclusion: The completeness of CXR registrations varied between 98% and 73% depending on the information source, and this should be considered when investigating radiology services in the basis of DNPR.
Keywords: Danish National Patient Registry; chest X-ray; diagnostic procedures; general practice; primary health care; validation study.