Rheumatology

CATCH registry shows real-world ERA benefits of early MTX and treat-to-target, but gaps remain

The Canadian Early Arthritis Cohort (CATCH) accrued >3,800 incident ERA patients across 25 sites and >35,000 visits, informing real-world care. CATCH data support early methotrexate initiation and treat-to-target strategies associated with measurable improvements in care delivery. Persistent disease activity, adherence problems, and disparities by sex, comorbidity, and social determinants of health remain common. The registry infrastructure also enabled pragmatic trials, guideline input, and outcome-measure validation useful for policy and practice. 1

Machine learning tools substantially cut SLR screening workload while preserving high recall

Replicating four SLRs for EULAR RA guidance, three vetted ML tools reduced screening workload by a mean 77.8% (SD 12.8%). ML-assisted screening captured over 95% of relevant records while reducing abstracts to screen by a mean 57.3% (SD 7.4%). Authors caution about limitations and emphasise need for prospective integration strategies to preserve methodological standards. ML screening appears useful for guideline teams but requires validation within each review's workflow before routine adoption. 2

PROs are common but incompletely reported in spinal disorder RCTs

In 42 RCTs (8,669 participants), 95.2% used PROs as primary outcomes but only 27.2% met high-quality PRO reporting standards. Investigators used 25 different PRO instruments across five domains, with VAS and ODI most frequent. Major gaps included lack of justification for instrument choice, unclear data collection methods, and inadequate missing-data handling. Standardizing PRO selection, collection, and statistical handling is needed to improve trial validity and comparability. 3