Using Wearable Technology to Quantify Physical Activity Recovery: Secondary Report From the AFTER (App-Facilitated Tele-Rehabilitation) Program for COVID-19 Survivors Randomized Study

Document Type


Publication Date



JMIR Publications

Source Publication

JMIR Rehabilitation and Assistive Technologies

Source ISSN


Original Item ID

DOI: 10.2196/43436


Background: Knowledge on physical activity recovery after COVID-19 survival is limited. The AFTER (App-Facilitated Tele-Rehabilitation) program for COVID-19 survivors randomized participants, following hospital discharge, to either education and unstructured physical activity or a telerehabilitation program. Step count data were collected as a secondary outcome, and we found no significant differences in total step count trajectories between groups at 6 weeks. Further step count data were not analyzed.

Objective: The purpose of this analysis was to examine step count trajectories and correlates among all participants (combined into a single group) across the 12-week study period.

Methods: Linear mixed models with random effects were used to model daily steps over the number of study days. Models with 0, 1, and 2 inflection points were considered, and the final model was selected based on the highest log-likelihood value.

Results: Participants included 44 adults (41 with available Fitbit [Fitbit LLC] data). Initially, step counts increased by an average of 930 (95% CI 547-1312; P<.001) steps per week, culminating in an average daily step count of 7658 (95% CI 6257-9059; P<.001) at the end of week 3. During the remaining 9 weeks of the study, weekly step counts increased by an average of 67 (95% CI −30 to 163; P<.001) steps per week, resulting in a final estimate of 8258 (95% CI 6933-9584; P<.001) steps.

Conclusions: Participants showed a marked improvement in daily step counts during the first 3 weeks of the study, followed by more gradual improvement in the remaining 9 weeks. Physical activity data and step count recovery trajectories may be considered surrogates for physiological recovery, although further research is needed to examine this relationship.


JMIR Rehabilitation and Assistive Technologies, Vol. 10 (2023). DOI.