Gossec L, Guyard F, Leroy D, et al. - Via a prospective observational study (ActConnect) that involved 155 patients with definite RA or axial spondyloarthritis, experts evaluated longitudinally the correlation between patient-reported flares and activity-tracker–provided steps per minute, using machine learning. The disease was well-controlled, although flares were common. The model created by machine learning did well against patient-reported flares. Sensitivity analyses were affirmative. In conclusion, nevertheless, these pilot verdicts would have to be validated, the conventional discovery of flares by machine-learning processing of activity tracker data gives a framework for prospective studies of remote-control monitoring of disease activity, with high accuracy and minimum patient burden.