Data-driven discovery of candidate predictors of future rheumatoid arthritis diagnosis in the UK biobank
Authors' conclusion
Our data-driven analyses identified lifestyle, psychosocial, clinical and biomarker features predictive of RA years before diagnosis.
Study context
- Population
- UK Biobank participants; incident rheumatoid arthritis (RA) diagnosis4,510 incident cases
- Participants
- 445,515 participants
- Outcome
- Candidate predictors of incident RA diagnosis
Methodological detailshypothesis-free data-driven machine learning, logistic regression adjusted for known confounders, Bonferroni corrected p-value of p<3.0 × 10-4
- Hypothesis-free data-driven machine learning
- Logistic regression adjusted for known confounders
- Bonferroni corrected p-value of p<3.0 × 10-4
Main finding
As reported in the abstract
Candidate predictors of RA diagnosis
Features identified by the model as potentially important
200
Neuroticism score, highest vs lowest, OR
1.44 (1.24-1.67) (95% CI)
CRP Q5 vs Q1, OR
1.92 (1.71-2.16) (95% CI)
Cystatin C Q5 vs Q1, OR
1.55 (1.38-1.74) (95% CI)
Observational data: associations, not effects.
Original publication
Data-driven discovery of candidate predictors of future rheumatoid arthritis diagnosis in the UK biobank. Semin Arthritis Rheum. Published 21 August 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →