Else Helene Ibfelt, Jan Sørensen, Dorte V Jensen, Lene Dreyer, Berit Schiøttz-Christensen, Pia H Thygesen, Ada Colic, Johnny L Raun, Natalia Manilo, Anne Rødgaard, Uta E Poulsen, Claus Rasmussen, Torben Hansen, Babara Unger, Randi Pelck, Anita Kincses, Henrik Nordin, Tove Lorenzen, Ali Theibich, Inger Marie Jensen Hansen, Jakob Espesen, Jolanta Grydehøj, Mette Holland-Fischer, Anne Gitte Loft, Merete Lund Hetland

In Denmark, patients with rheumatoid arthritis (RA) are registered in the nationwide clinical DANBIO quality register and the Danish National Patient Registry (DNPR). The aim was to study the validity of the RA diagnosis and to estimate the completeness of relevant RA cases in each registry. Patients registered for the first time in 2011 with a diagnosis of RA were identified in DANBIO and DNPR in January 2013. For DNPR, filters were applied to reduce false-positive cases. The diagnosis was verified by a review of patient records. We calculated the positive predictive values (PPVs) of the RA diagnosis registrations in DANBIO and DNPR, and estimated the registry completeness of relevant RA cases for both DANBIO and DNPR. Updated data from 2011 to 2015 from DANBIO were retrieved to identify patients with delayed registration, and the registry completeness and PPV was recalculated. We identified 1,678 unique patients in DANBIO or in DNPR. The PPV (2013 dataset) was 92% in DANBIO and 79% in DNPR. PPV for DANBIO on the 2015 update was 96%. The registry completeness of relevant RA cases was 43% in DANBIO, increasing to 91% in the 2015 update and 90% in DNPR. DANBIO held a high proportion of true RA cases (96%) and was found to be superior to the DNPR (79%) with regard to the validity of the diagnosis. Both registries were estimated to have a high completeness of RA cases treated in hospital care (~90%).