Identifying patients at high risk of rheumatoid arthritis diagnosis using electronic healthcare records: the RheumDetect study. (360G-Wellcome-319356_Z_24_Z)

Rheumatoid arthritis (RA) is an autoimmune disease. It significantly affects quality of life and is related to higher morbidity and mortality. Early diagnosis and treatment can prevent disease-related damage and improve outcomes. Recent studies have shown that patients present to their GP more frequently in the two years before RA diagnosis indicating that a prolonged ‘diagnostic window’ may exist in RA in which more timely detection may be achieved for some patients. While promising new approaches to disease prediction, which take account of trends in blood tests and the temporal order of clinical events (disease trajectories), have been applied in serious diseases such as cancer, they have not been applied in RA. This study aims to use data from the Clinical Practice Research Datalink (CPRD) to identify patterns in clinical events (including symptoms, blood tests results, prescriptions) occurring up to 5 years prior to rheumatoid arthritis diagnosis and to develop and validate dynamic risk prediction models (conventional and ML based) to support earlier RA detection. The results of this study are intended to inform policy and guide clinical decision making on the need for rheumatoid arthritis tests or referrals.