Background: Health registries may yield important data for epidemiological studies. However, in order to be a valuable source for information, the registered data have to be correct.
Objectives: The aim of the study was to validate data from the Norwegian Patient Registry (NPR) regarding multiple sclerosis (MS).
Materials and methods: We obtained data on individuals residing in Nordland County and registered with a MS diagnosis in the NPR or in local hospital records. The NPR data included a unique 11-digit personal identity number that made it possible to identify the individuals medical records. For each individual registered with MS in the NPR, the hospital record was scrutinized in order to confirm or rule out the diagnosis.
Results: In Nordland County, 657 individuals had MS 1 January 2017. Of these, 637 were recorded with a correct diagnosis of MS in the NPR, while 59 were recorded incorrectly. Incorrect registration was due to a diagnosis that did not fulfill the diagnostic criteria, later investigation had ruled out MS or it was an error in the diagnostic code registration process. Twenty individuals were not registered with MS in the NPR. These were patients who received their diagnosis before data in the NPR were person identifiable (before 2008), and who later had no MS-registered contact with public specialist healthcare services. The sensitivity is 0.97, and the positive predictive value is 0.92.
Conclusion: Data from the NPR gave a good estimate of the occurrence of MS, but nearly one in 10 registered diagnoses was not correct.
Keywords: epidemiology; health registries; multiple sclerosis; validation.
© 2019 The Authors. Brain and Behavior published by Wiley Periodicals, Inc.