Spondyloarthritis & psoriatic arthritis
Reducing selection bias while maintaining precision through an integrated analysis: 2-year longitudinal analysis of imaging outcomes in the SPondyloArthritis Caught Early cohort
Authors' conclusion
Multilevel models retain level variability by integrating information across patients, time points, readers and rounds to reduce selection bias and improve precision of change, which would otherwise be lost in simpler models.
Study context
- Population
- Patients with axial spondyloarthritis (axSpA) from the SPondyloArthritis Caught Early cohort followed for 2 years; imaging of sacroiliac joints and spineSPACE cohort, axSpA
- Participants
- 332 vs 279
- Exposure
- Integrated analysis vs completers-only analysis; aggregated scores vs scores per reader
- Integrated analysisn = 332
- Completers-onlyn = 279reference
- Outcome
- Precision of change estimates in imaging outcomes, integrated vs completers-only analysis
- Follow-up
- 2 years
Methodological detailsgeneralised estimating equations, multilevel models, ≥2 central readers in two rounds
- Generalised estimating equations
- Multilevel models
- ≥2 central readers in two rounds
Main finding
As reported in the abstract
Patients included and precision of change estimates
Integrated analysis included more patients than multilevel completers-only analysis, with similar precision
12%-16% more patients
Observational data: associations, not effects.
Further findings
Aggregated vs reader scores (completers-only), continuous outcomes
- Change estimates similar
Aggregated vs reader scores (completers-only), dichotomous outcomes
- Aggregated scores underestimated spinal changes
- maximum 82% less
- Aggregated scores overestimated sacroiliac joint changes
- maximum 37% more
Original publication
Reducing selection bias while maintaining precision through an integrated analysis: 2-year longitudinal analysis of imaging outcomes in the SPondyloArthritis Caught Early cohort. RMD Open. Published 26 March 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →