JMIR Publications

ABSTRACT

Background:

The health conditions of Japanese salespersons may be adversely affected by their lifestyle. Face-to-face or on-site health intervention are not convenient for salespersons because of their tendency towards out-of-office sales. Previous studies showed that mobile health (mHealth) interventions (compared to the usual), has a great potential to promote physical activity (PA). For Japanese salespersons, mHealth can offer additional convenience to change their PA habits because they can access mHealth contents anytime and anywhere. However, it is still unclear what specific elements are important to maintain the PA level using an mHealth approach.

Objective:

We aimed to identify elements which account for both the high average PA level and the prevention of a long-term decrease in PA.

Methods:

Salespersons were recruited from 11 Japanese companies. Team-based walking events were held from October to December 2018 (total, 9 weeks), during which, the walking step data were recorded by smartphone application (apps). Average walking steps of each participant during the event and the difference in walking steps between the initial to the final week were respectively used as dependent variables. We performed the analyses to estimate the effects of team characteristics (i.e., the frequency of communication with team members and team size) and behavioral characteristic (i.e., number of days with recorded steps on the apps) on the average walking steps, and the difference in walking steps between the initial and the final week. Multiple and multilevel regression analyses were performed.

Results:

Of 416 participants, walking step data of 203 participants who completed post-event assessments were included in the analyses. Multiple regression analysis of the average walking steps showed that the number of days with recorded steps was positively correlated with the log-transformed average walking steps (β = 0.01, P < .001). Multilevel analysis of the average walking steps considering the company level estimated that the intra-class correlation coefficient was 37%. This means that belonging to the same company largely affected an individual’s average walking steps. Multiple regression analysis of the difference in walking steps showed that communication with team members once or twice a week correlated with preventing a decrease in walking steps from the initial to the final week (β = 1539.4, P < .05), and being on a larger team also correlated with a decrease in walking steps from the initial to the final week (β = -318.4, P < .05).

Conclusions:

This study showed that the elements which accounted for high average walking steps and those preventing the decrease in walking steps from the initial to the final week differed. Behavioral characteristic correlated positively with average walking steps. Team characteristics (regular communication and smaller team size), significantly correlated with preventing a decrease in walking steps.