JMIR Publications

ABSTRACT

Background:

Workplace programs designed to improve the health and psychological well-being of employees are becoming increasingly popular. However, there are mixed reports regarding the effectiveness of such programs and little analysis of what helps people to engage with such programs.

Objective:

This evaluation of a particularly broad, team-based, digital health and well-being program uses mixed methods to identify the elements of the program that reduce work stress and promote psychological well-being, sleep quality, and productivity of employees.

Methods:

Participation in the Virgin Pulse Global Challenge program during May to September 2016 was studied. Self-reported stress, sleep quality, productivity, and psychological well-being data were collected both pre- and postprogram. Participant experience data were collected through a third final survey. However, the response rates for the last 2 surveys were only 48% and 10%, respectively. A random forest was used to estimate the probability of the completion of the last 2 surveys based on the preprogram assessment data and the demographic data for the entire sample (N=178,350). The inverse of these estimated probabilities were used as weights in hierarchical linear models in an attempt to address any estimation bias caused by the low response rates. These linear models described changes in psychological well-being, stress, sleep, and productivity over the duration of the program in relation to gender and age, engagement with each of the modules, each of the program features, and participant descriptions of the Virgin Pulse Global Challenge. A 0.1% significance level was used due to the large sample size for the final survey (N=18,653).

Results:

The final analysis suggested that the program is more beneficial for older people, with 2.9% greater psychological well-being improvements observed on average in the case of women than men (P<.001). With one exception, all the program modules contributed significantly to the outcome measures with the following average improvements observed: psychological well-being, 4.1%-6.0%; quality of sleep, 3.2%-6.9%; work-related stress, 1.7%-6.8%; and productivity, 1.9%-4.2%. However, only 4 of the program features were found to have significant associations with the outcome measures with the following average improvements observed: psychological well-being, 3.7%-5.6%; quality of sleep, 3.4%-6.5%; work-related stress, 4.1%-6.4%; and productivity, 1.6%-3.2%. Finally, descriptions of the Virgin Pulse Global Challenge produced 5 text topics that were related to the outcome measures. Healthy lifestyle descriptions showed a positive association with outcomes, whereas physical activity and step count tracking descriptions showed a negative association with outcomes.

Conclusions:

The complementary use of qualitative and quantitative survey data in a mixed-methods analysis provided rich information that will inform the development of this and other programs designed to improve employee health. However, the low response rates and the lack of a control group are limitations, despite the attempts to address these problems in the analysis.