Satisfaction with telemedicine versus in-person visits in rheumatology: a noninferiority randomised controlled trial

We tested whether a rheumatology telemedicine visit was noninferior to an in-person visit for patient satisfaction and care effectiveness, including for subgroups of age, sex, race and ethnicity, income, and employment status. This multicenter noninferiority trial randomised patients to 1 in-person or telemedicine visit. The primary outcome was the proportion of high visit satisfaction (9 or 10 on a 0-10 scale). The primary analysis tested whether patient satisfaction with telemedicine was noninferior to in-person care (10% noninferiority margin). Secondary and exploratory outcomes included preference for next visit type, satisfaction in subgroups, and immunosuppressant toxicity monitoring. We performed modified intent-to-treat (mITT) and per-protocol analyses. Among 651 randomised patients (in-person, n = 323; telemedicine n = 328), 500 (76.8%) had a visit defining the mITT population. Satisfaction with telemedicine was not noninferior to in-person; 77.0% telemedicine, 90.1% in-person, difference 13.1% (90% CI, 7.7%-18.5%). Per-protocol analysis results were consistent. More participants in the in-person group compared with those in telemedicine preferred the same visit type for their next visit vs a different visit type/no preference (55.6% in-person, 19.1% telemedicine, P < .0001, mITT analysis). Men were equally satisfied with both visit types (90.0%), while women were more likely to be satisfied with in-person visits (90.2% vs 74.7%). Toxicity monitoring rates were higher in in-person vs telemedicine (eg, hepatic function: 92.1% vs 66.3%, P = .0001, per-protocol analysis). Among a large group of diverse patients, satisfaction with telemedicine was not noninferior to in-person rheumatology visits. More participants preferred in-person visits in the mITT and per-protocol analyses. Appropriate toxicity monitoring was lower in telemedicine vs in-person groups.