The stakes are considerable. In industrialized societies, roughly 15 to 20 percent of the workforce labours outside the conventional daytime window of 6 a.m. to 6 p.m., staffing hospitals, mines, factories and transport networks through early mornings, evenings and nights. This pattern of work is consistently linked to elevated risks of cardiovascular disease, breast cancer and type 2 diabetes. The mechanisms appear to be multifactorial, involving circadian disruption, disturbed sleep, psychosocial stress and lifestyle behaviours, but one piece of the puzzle has remained frustratingly blurry: how much shift workers actually move and sit during their unpredictable days.

Traditional research tools struggle with this question. Self-report recall questionnaires ask people to summarize weeks or months of behaviour, an approach prone to memory errors and one that rarely accounts for the wild variation in shift schedules. A nurse finishing a night shift on Tuesday and starting an early shift on Thursday experiences fundamentally different days, and a single retrospective questionnaire flattens that complexity into a misleading average. EMA takes the opposite approach, pinging participants on their phones at scheduled moments and asking what they were doing in the preceding few minutes, capturing behaviour as it unfolds in its natural context.

In the study, the team recruited 128 workers in and around Brisbane, Ipswich and Toowoomba in south-east Queensland, ultimately analysing data from 120 participants, 58 percent of whom were female with a mean age of 36 years. The cohort included 69 shift workers, most of them nurses and paramedics, alongside 51 non-shift workers drawn from office settings. After downloading the SEMA3 smartphone app, participants were divided into three groups: shift workers receiving prompts tailored to their individual shift schedules, shift workers receiving standardized prompts, and non-shift workers receiving standardized prompts.

The prompting protocol was technically precise. Every participant received five surveys per day at three-hour intervals over seven to ten days, with the duration adjusted for shift workers so that data covered all their work patterns and days off. Tailored prompts were calculated from each worker’s anticipated shift schedule and wake-sleep pattern, then adjusted whenever the schedule changed. For night-shift workers, prompts ran from 10 p.m. to 10 a.m.; for evening shifts, from 1 p.m. to 1 a.m. Standardized prompts, by contrast, arrived between 10 a.m. and 10 p.m. regardless of schedule, deliberately avoiding hours when workers were likely asleep. Each survey contained five questions asking about current activity, its duration, its location, and any sitting involved, and took an average of just 24 seconds to complete. Unanswered surveys vanished after 30 minutes.

The recruitment results were strikingly strong. About 78 percent of invited workers enrolled, and every one of the 128 enrolled participants completed the study, with retention defined as responding to at least one prompt on four of seven days for non-shift workers and seven of ten days for shift workers. Those who declined mostly cited unwillingness to travel to the university for in-person meetings. Across the study, 4,482 prompts were sent, and 64 percent of surveys were started and completed, slightly below the 70 percent target the researchers set based on prior EMA literature, but comfortably within the wide 58 to 92 percent range reported in comparable studies.

The comparison between prompting strategies produced a nuanced picture. Shift workers on standardized prompts showed the lowest compliance at 57 percent, while tailored-prompt shift workers and non-shift workers reached 64 and 68 percent respectively, a difference that was not statistically significant. Interestingly, tailored-prompt shift workers answered more surveys on their days off than on working days, and were more likely to respond during day and evening shifts than during night shifts. Non-shift workers, meanwhile, responded more reliably on weekdays than weekends. The first prompt of the day attracted the highest response rate at 26 percent, while the fifth and final prompt drew only 14 percent, suggesting that engagement fades as the day wears on regardless of occupation.

Notably, compliance showed no relationship with age, gender, body mass index or marital status, an encouraging sign that EMA does not systematically exclude certain demographic groups. The researchers also found that reported activities differed across groups, with work duties the most frequently logged activity at 16.4 percent of responses, followed by phone or computer use, and non-shift workers reporting the highest levels of physical activity and television watching. Caring for children was the least reported activity at just one percent.

The authors are candid about why compliance fell short of the goal. Unlike some comparable workplace studies, their protocol sent no reminders and offered no monetary incentives, only personalized feedback on participants’ responses and movement. Evidence from meta-analyses suggests both reminders and incentives can lift completion rates, and an office-worker study that paid participants 20 dollars achieved higher compliance. The 30-minute expiry on unanswered prompts may also have cost responses. There are practical limits, too: workers in mining, manufacturing or retail may simply be unable to reach their phones during shifts, and the standardized 10 a.m. to 10 p.m. window may have missed early-morning and late-night behaviour entirely.

Even so, the study stands as one of the few to adapt EMA prompting to actual shift schedules, and its message is cautiously optimistic. Enrollment and retention were high, participant burden was demonstrably low, and the method generated thousands of within-person data points that traditional questionnaires could never capture. Beyond measurement, the researchers note that repeated real-time monitoring may itself make workers more aware of their habits, potentially prompting them to break up long stretches of sitting. If future studies can boost compliance through reminders, incentives and smarter tailoring, EMA could become the foundation for health promotion programs that know precisely when, and why, shift workers need support most.

Subject of Research: Feasibility of ecological momentary assessment for measuring physical activity and sedentary behaviour in shift and non-shift workers

Article Title: Feasibility of ecological momentary assessment in measuring physical activity and sedentary behaviour in shift and non-shift workers

Article References: Monnaatsie, M., Biddle, S. J., & Kolbe-Alexander, T. (2024). Feasibility of ecological momentary assessment in measuring physical activity and sedentary behaviour in shift and non-shift workers. Journal of Activity, Sedentary and Sleep Behaviors, 3(1), Article 24. https://doi.org/10.1186/s44167-024-00063-7

Image Credits: AI Generated

DOI: 10.1186/s44167-024-00063-7

Keywords: shift work, ecological momentary assessment, physical activity, sedentary behaviour, smartphone surveys, occupational health, EMA compliance, circadian disruption, nurses, mHealth, study feasibility, Feasibility