Background Digital sensors are gaining popularity not only for health and wellness, but for remote tracking of key health indicators. Long-term adoption and longitudinal sensor data collection remain challenging. In the general US population, one third of owners of wearable devices discontinue use within 6 months of first use [Endeavour Partners, 2014]. Less is known about the continuation of activity monitoring sensors in patients with chronic medical conditions. Methods The present study reviewed the continuation of activity tracking in two cohorts of multiple sclerosis (MS) patients from an online patient research network (www.patientslikeme.com; PLM). First, we considered data from a June 2014 study in which MS patients received activity trackers (Fitbit One) and were engaged for a 3-week pilot study. Second, an additional cohort was derived from non-study MS patients who joined the site and linked their Fitbit brand activity trackers to PLM. With both cohorts, patient tracking data was voluntarily shared through the PLM platform and was enabled through the manufacturer's application program interface (API). Participant engagement was enhanced through social media support (community moderation, patient support, peer-to-peer forum discussion). Combining these cohorts yielded a database with activity data from MS patients spanning over 2 years. Results Two hundred forty-eight (248) patients were recruited for the pilot study of which 82% enabled data sharing and the mean 3-week adherence was 86% (18.2 out of 21 days). Following the pilot study, 23 patients continued to link and provide real time activity data over time (range: 1 month to 24 months). A total of 20 patients contributed activity data for a period spanning 12 months (12 of those contributed activity data for a period spanning 24 months). An additional 62 MS patients who were not part of the pilot study also synced their Fitbit activity trackers with their PLM profile. A total of 39 non-pilot study patients contributed activity data for a period spanning 12 months (18 of those contributed activity data for a period spanning 24 months). Including both pilot study participants and non-pilot study participants, a total cohort of 59 patients contributed activity data spanning a one-year period; of those, 30 contributed activity data spanning a period of two years. Conclusion Free-living assessment of ambulation can be challenging due to varying degrees of user continuation. A community engagement model may be a useful framework for long-term data collection.