Validation of the Synovo Digital Disease Activity Score (DDAS) Against Physician-Performed DAS28/CDAI: A Non-Clinical, Non-Interventional, Non-Academic Study

Overview

The goal of this observational study is to assess whether the Synovo digital assessment can measure rheumatoid arthritis disease activity in adults with rheumatoid arthritis. The main questions it aims to answer are: How closely does the Synovo digital disease activity score agree with the clinician-assessed DAS28 score? Can the Synovo assessment identify changes in disease activity between follow-up visits? Participants will: Complete a structured digital questionnaire Record guided videos of selected joint movements using a smartphone Undergo a standard DAS28 assessment by a clinician during the same visit Some participants may repeat the assessments at a later follow-up visit.

Rheumatoid arthritis is a chronic inflammatory disease that can lead to pain, functional impairment, joint damage and disability. Evidence from tight-control and treat-to-target studies shows that structured and repeated measurement of disease activity, followed by timely clinical action, improves disease control and remission outcomes. The Synovo DDAS combines patient-reported information with movement features derived from four guided smartphone videos. The system produces a continuous score calibrated to the DAS28 scale. The study evaluates the measured relationship between DDAS and clinician-derived DAS28. It does not assume interchangeability and does not test autonomous treatment decisions. The study is designed to determine whether the DDAS demonstrates sufficient association, agreement and longitudinal tracking to justify further independent and multicenter validation. This is an exploratory, single-centre study intended to establish an initial clinical evidence base rather than definitive clinical interchangeability or regulatory-grade performance. Primary objective: To assess the association between paired DDAS and clinician-derived DAS28 values. Secondary objectives * To quantify numerical agreement between DDAS and clinician-derived DAS28 using mean bias and limits of agreement, with appropriate adjustment for repeated measurements within participants. * To assess score scaling and agreement using regression analysis and a linear mixed-effects model with a participant-level random intercept. * To evaluate categorical agreement across the standard DAS28 disease activity bands using a confusion matrix and quadratic-weighted Cohen's kappa. * To estimate sensitivity, specificity, positive predictive value and negative predictive value at clinically relevant DAS28 thresholds, including high disease activity, treatment target and remission. * To assess the ability of DDAS to track the direction and magnitude of changes in clinician-derived DAS28 across repeat routine-care visits.