Clin Epidemiol. 2021 May 24;13:335-344. doi: 10.2147/CLEP.S298770. eCollection 2021.

Camilla Rye ^1 2, Katrine Hass Rubin ², Frederik Trier Moller ^3 4, Mette Julsgaard ⁵, Tine Jess ⁶, Vibeke Andersen ^1 2 7

Author information

• ¹Research Unit for Molecular Diagnostics and Clinical Research, Institute for Regional Health Research-Center of Southern Jutland, University Hospital of Southern Denmark, Aabenraa, Denmark.
• ²OPEN - Open Patient Data Explorative Network, Department of Clinical Research, University of Southern Denmark and Odense University Hospital, Odense, Denmark.
• ³Department of Epidemiological Research, Statens Serum Institute, Copenhagen, Denmark.
• ⁴Department of Infectious Disease Epidemiology and Prevention, Statens Serum Institute, Copenhagen, Denmark.
• ⁵Department of Hepatology and Gastroenterology, Aarhus University Hospital, Aarhus, Denmark.
• ⁶Center for Clinical Research and Prevention, Bispebjerg and Frederiksberg Hospital, Copenhagen, Denmark.
• ⁷Institute of Molecular Medicine, University of Southern Denmark, Odense, Denmark.

Abstract

Purpose: Inflammatory bowel diseases (IBD), including Crohn's disease (CD) and ulcerative colitis (UC), are chronic diseases. The aim was to validate diagnoses of IBD among patients aged 50+ years in the Danish National Patient Registry (NPR) by comparison with patient medical records.

Patients and methods: Men and women in the Diet, Cancer and Health (DCH) cohort were linked to NPR, and cases with a diagnosis of IBD and their respective hospital records were identified. Validation was performed by comparing patient medical records with information on discharge diagnoses of IBD from the NPR.

Results: Of 57,053 individuals in the DCH-cohort, 339 were registered with an IBD diagnosis in NPR, with 277 (82%) records available for review. Among 277 patients, the positive predictive values (PPVs) of one CD or UC registration in NPR were 78% for IBD overall, 51% for CD and 54% for UC. One hundred fifty-seven patients had at least two CD and/or UC registrations with PPVs of 90% for IBD overall, 65% for CD and 73% for UC. One hundred and two patients had at least three registrations with PPVs of 97% for IBD overall, 75% for CD and 88% for UC. 96% were diagnosed at a specialized department. Other diagnoses coded as IBD mostly included microscopic colitis, irritable bowel syndrome and cancer.

Conclusion: Validity of IBD diagnoses in the registry of individuals aged 50+ years increased with the number of registrations. It is recommended that these results are taken into consideration in future studies, especially in epidemiology research using NPR as a data source for patients diagnosed with IBD.