A VALIDATION OF REGISTER DERIVED DIAGNOSES OF INTERSTITIAL LUNG DISEASE IN PATIENTS WITH INFLAMMATORY ARTHRITIS. DATA FROM NOR-DMARD

S. Aarrestad Provan, F. Ahlfors, G. Bakland, H. Ye, E. Kristianslund, T. K. Kvien, E. Ikdahl, T. M. Aaløkken, A. M. Hoffmann-VoldDiakonhjemmet Hospital, Center for Treatment of Rheumatic and Musculoskeletal Diseases (REMEDY), Oslo, Norway Inland Norway University of Applied Sciences, Elverum, Public Health, Elverum, Norway Sahlgrenska University Hospital, Radiology, Gothenburg, Sweden University Hospital of North Norway HF, Rheumatology, Tromsø, Norway Revmatismesykehuset, Rheumatology, Lillehammer, Norway Oslo University hospital, Department of Radiology, Institute of Clinical Medicine, Oslo, Norway Oslo University hospital, Rheumatology, Oslo, Norway  Background There is a lack of knowledge concerning the validity of International classification of diseases (ICD) diagnoses extracted from health registers although this information is often used in epidemiological studies. Objectives To assess the validity of register derived diagnoses of interstitial lung disease (ILD) in patients with rheumatic diseases, using ILD identified on computed tomography (CT) and/or medical journal as gold standard. Methods The Norwegian Anti-Rheumatic Drug Register (NOR-DMARD) is a multi-centre prospective observational study that includes patients with a diagnosis of inflammatory arthritis who started treatment with disease modifying anti-rheumatic drugs. NOR-DMARD was linked to the Norwegian Patient Registry (NPR) and Cause of Death Registry. We searched both registers for ILD ICD-10 codes (J70, J84 or J99) and extracted CTs from participants at four hospitals. CTs taken three months before or within one year after the first ICD-10 diagnosis in the registries were considered relevant to the study. An expert thoracic radiologist (FA) scored all examinations according to presence of ILD and categorised the finding into; no ILD, possible ILD and confirmed ILD. Possible ILD CT images were re-examined by a second expert radiologist (TMA). In addition, medical records were searched for ILD diagnoses given by specialist clinicians on several occasions. Presence of confirmed ILD was assessed in patients with available CTs, registry ICD-code for ILD at ≥2 time points and across ILD diagnoses subgroups. Results We identified 80 cases with an ILD diagnosis given in a register (ICD-10 n (%); J70 10 (12.5), J84 52 (65.0) and J90 18 (22.5)). CTs were available in 72/80 (90%) patients, 60/72 (83%) within the pre-specified time-window. The ILD diagnosis was confirmed on CT in 31/80 (38.8%) and in 9/80 (11.3%) by medical records only. ILD was validated in 34 (56.7%) of 60 patients with available CT within the relevant time-period and in 28/43 (65.1%) patients who had received an ILD code at ≥2 time points and had a HRCT within the relevant time-period (Table 1). The diagnosis was confirmed in 20 (69.9%) of the 29 patients of this group who had received a J84 diagnoses. ILD diagnoses that were not validated were most frequently given under investigations for respiratory symptoms 20/38 (52.6%). Conclusion The validity of registry-based diagnoses of ILD must be carefully considered in epidemiological studies. Table 1. Positive predictive value of ILD diagnoses from patient registers in cohorts with progressive selection criteria Number of patients in cohort Number of patients (%) with an ILD diagnosis confirmed by CT and/or medical records 80 patients with ILD in register 42 (52.5) 60 patients with ILD in register and available CT taken within the prespecified time-period 34 (56.7) 43 with two or more ILD diagnoses in register and available CT taken within the prespecified time-period 28 (65.1) 45 with two or more ILD diagnoses in register and available CT taken within the prespecified time-period across diagnoses ICD10: J702 (33.3) ICD10 J8420 (69.0) ICD10 J996 (75.0) Footnotes: ILD: interstitial lung disease REFERENCES: NIL. Acknowledgements: NIL. Disclosure of Interests Sella Aarrestad Provan Consultant of: Boehringer Ingelheim, Novartis, Grant/research support from: Boehringer Ingelheim, Fredrik Ahlfors: None declared, Gunnstein Bakland Speakers bureau: Abbvie, Consultant of: Novartis, UCB, Pfizer, MSD og Celgene, Hu Ye: None declared, Eirik kristianslund: None declared, Tore K. Kvien Speakers bureau: Grünenthal, Sandoz, UCB, Consultant of: AbbVie, Amgen, Celltrion, Gilead, Novartis, Pfizer, Sandoz, UCB , Grant/research support from: AbbVie, Amgen, BMS, Galapagos, Novartis, Pfizer, UCB. , Eirik Ikdahl: None declared, Trond M Aaløkken Speakers bureau: Boehringer Ingelheim, Anna-Maria Hoffmann-Vold Speakers bureau: Boehringer Ingelheim, Jannsen, Medscape, Merck Sharp & Dohme and Roche, Consultant of: ARXX, Boehringer Ingelheim, Genentech, Jannsen, Medscape, Merck Sharp & Dohme and Roche, Grant/research support from: Boehringer Ingelheim, Jannsen. Keywords: Diagnostic tests, Comorbidities DOI: 10.1136/annrheumdis-2023-eular.3276Citation: , volume 82, supplement 1, year 2023, page 2000Session: Diagnostics and imaging procedures (Publication only)