Real‐World Study to Evaluate Change in Clinical, Wearable Biosensor, and Patient‐Reported Outcomes Among Patients With Rheumatoid Arthritis Initiating Treatment With a Janus Kinase or Tumor Necrosis Factor Inhibitor

Objective To evaluate rheumatoid arthritis (RA) symptoms and disease activity among patients initiating new treatment, we examined clinical, patient‐reported outcome (PRO), and wearable biosensor data, and explored longitudinal associations among these three data types. Methods RA patients with moderate‐to‐high disease activity (Clinical Disease Activity Index (CDAI) >10) and prescribed upadacitinib or adalimumab were enrolled (03/2021–10/2022) during in‐office visits at 28 community rheumatology clinics in the United States and followed for ≥12 weeks. At‐home data collection included disease activity and symptom measures from the PRO Measurement Information System (PROMIS) using the PatientSpot app. Biosensor activity data were collected with a wrist‐worn Fitbit device. Changes over time and associations among data types were analyzed with Pearson's R correlations and linear regression models. Results Of 299 patients recruited, 204 (68%) started upadacitinib or adalimumab within 30 days; 150 (50%) provided clinical and PRO data, the main study population of whom 95 (63%) provided biosensor data. Mean CDAI improved 13 units and PROMIS Pain Interference, Fatigue, and Sleep Disturbance scores improved by 4–5 units from baseline to follow‐up. Activity improved for patients considered sedentary at baseline. Moderate correlations between CDAI and PROs ranged from 0.38–0.56. Linear regression model showed that changes in PROs explained 52% of the variability in change in CDAI. Conclusions Treatment response is observed across diverse data types and tracking RA disease activity outside of the clinic for patients initiating new treatments may be useful in evaluating disease status and optimal changes in therapy, even in the absence of physician‐derived measures.