Introduction

It is estimated that around 1 billion workers endure excessive heat exposure [1] and more than one in three workers exposed to heat suffer from heat strain [2]. Heat strain is broadly defined as the physical and psychological responses of the body caused by heat exposure [1]. Heat strain increases the risk of work injuries and severe heat illness, that can ultimately cause death [3]. With ongoing climate change, it is expected that the prevalence of heat strain will increase, reinforcing the need for effective prevention strategies [1].

Among the range of heat preventive measures implemented in occupational settings, closely monitoring workers’ health is essential, not only to avoid severe health issues and support individual well-being, but also to deepen our understanding of the effects of heat on human health. In the literature, workers health under heat strain is usually measured through self-reported questionnaires and sometimes by measuring physiological parameters (e.g., heart rate, core temperature). These health outcomes are often collected in cross-sectional studies, with heat strain symptoms or health markers typically recorded only before and after a work shift. For example, a detailed review of the studies, included in a meta-analysis examining the prevalence of heat strain in occupational settings [2], revealed that 22 of the 33 included studies used only two or fewer health observation windows, primarily taken before and after work shifts. Only eleven studies conducted three or more assessments, out of which, only six used intensive monitoring of workers’ heart rate and/or temperature. However, none of these intensive studies monitored workers for longer than one shift. The longest monitoring period of all the included studies extended to just two weeks.

This opinion argues that current methods used to estimate worker’s health under heat strain are suboptimal, given the range of technological solutions now available. Two promising approaches are passive sensing and ecological momentary assessments (EMAs). Passive sensing refers to the continuous collection of physiological and behavioural outcomes through wearable devices. EMAs, on the other hand, are brief surveys, usually prompted on smartphones, that capture subjective state at relevant moments of the day (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Conceptual illustration of intensive monitoring of outdoor workers via activity monitors and EMAs across a summer.

Illustration made by the authors using images designed by Freepik (www.freepik.com). Image of farmer retrieved from Freepik (https://www.freepik.com/free-vector/farmer-using-technology-digital-agriculture_16310216.htm#fromView=search&page=1&position=0&uuid=dc82a45c-337b-4ae6-90a4-1a771ffffa25&query=farmer±smartphone); image of smartwatch retrieved from Freepik (https://www.freepik.com/free-vector/fitness-trackers-flat-design_7971283.htm#fromView=search&page=1&position=15&uuid=c2d0691f-40af-4ddd-8d9a-74d105ff5a76&query=smartwatch±illustration).

https://doi.org/10.1371/journal.pclm.0000795.g001

The benefits of intensive monitoring

From a methodological standpoint, intensively monitoring heat strain in outdoor workers would help limit recall bias, when an individual does not accurately remember a past experience. Typically, participants are asked to remember heat related symptoms that have occurred several days or even weeks before the assessment [2], increasing the risk of measurement error [4]. Timely methods such as EMAs can prompt symptom reporting based on workers’ schedules, mitigating these limitations and misleading conclusions about the effectiveness of specific interventions [5].

At a more conceptual level, some specific questions could arguably be better understood through individual, continuous, time-series data, such as heat acclimatisation [6]. Most studies on heat acclimatisation rely on pre- and post-assessments, typically before and after the summer [7]. Time series data collected across a whole summer via passive sensing and EMAs could be valuable to study heat acclimatisation, enabling the analysis of time-varying associations, and offering a much more fine-grained understanding of how vulnerability to heat evolves over time. A systematic review of studies on occupational heat stress in critical sectors (industry, construction, mining and agriculture) [8], underlines the scarce in-field longitudinal research, limiting conclusive findings about adaptation.^. Having relatively long individual time series also provides the opportunity to perform person-specific analyses [9], and, relevant to the concept of acclimatisation, to identify which participants are more resilient to heat, and why.

At the interventional level, continuously monitoring heat strain in outdoor workers would pave the way for the development of just-in-time adaptive interventions (JITAIs). JITAIs are a type of adaptive intervention which aim to provide the right type and intensity of support to individuals at the right time [10]. For example, in the context of occupational heat strain, relatively simple algorithms combining temperature and heat strain data could trigger the decision to provide an intervention, ranging from a simple smartphone’s notification or call reminding workers of heat prevention measures, or to adjust workload in the next days. These decisions could even be anticipated some days using tailored early warning signals (see an example for physical activity [11]). We are aware of only one trial using real-time feedback via health alerts (based on workers’ heart rate and core temperature) to both construction workers and their supervisors to prevent heat stroke [12].

The feasibility of continuously monitoring health outcomes under heat stress in outdoor workers

Only a few studies used continuous health monitoring among outdoor workers for extended periods, such as over the course of one or more seasons. One study, for example, was conducted in rural Burkina Faso [13], collecting 11 months of heart rate (HR), sleep and daily steps data via activity monitor. Missing data were high: out of around 300 measurement days, valid days fluctuated from around 50%-55% for daily steps and sleep to less than 20% for HR. Some issues related to the setting were unreliable electricity and internet access, as well as environmental challenges affecting communication between the research team and the workers and obstructing the replacement of malfunctioning devices and the loading of recorded data. While offering valuable insights, the study focused on passive sensing measures, in a very specific challenging context. In a similar vein, a shorter, two-week, study conducted in a rural context in Kenya [14] reported high percentage of days for which usable data were recorded (85.7%-100%) of research-grade activity monitors in a group of farmers with no particular challenges regarding devices’ installation, data synchronization or devices’ damages related to the field work. Another study conducted among farmers in the United States being prompted EMAs for 24 randomly selected days over 6 months, showed a median response rate of 75% and a median of 18 days of daily farming tasks reported per worker [15]. Finally, our group also conducted a study in Catalonia during the summers of 2022 and 2023 in two different group of outdoor workers (N = 21) combining EMAs and passive sensing with activity monitors (unpublished descriptive statistics related to protocol’s feasibility can be found on OSF https://osf.io/zq4x8/).

Conclusion

To our knowledge, only a handful of studies have tested continuous monitoring of workers health outcomes over extended periods, despite the significant benefits outlined. Technical challenges (e.g., device malfunction, data loss, limited connectivity) and participant burden (e.g., survey fatigue, discomfort with certain devices) can reduce data completeness. However, research in other domains has more consistently demonstrated the feasibility and acceptability of EMA and passive sensing over several months. Engaging participants in the co-design of the study protocol, with a focus on selecting wearable devices that align with their preferences, is essential. There are several alternatives to the wrist-worn accelerometers, that could be considered less invasive to the participants, such as connected sleep mattress or digital rings. Second, potential motivational strategies include rewards, such as positive feedback for consistent participation, personalized information on health status, or financial incentives for each completed survey or for each valid day of monitoring. Finally, the use of “measurement bursts”, periodic assessments phases dispersed over time (such as five consecutive days every two weeks), can reduce participants’ burden and prevent respondents’ fatigue, while allowing the collection of longitudinal intensive workers’ health outcomes [8].

We argue that occupational health research could benefit from adopting intensive monitoring studies using EMAs and passive sensing. After gaining a clearer understanding of the feasibility of such protocols in different workplace contexts, the adoption of intensive monitoring related tools should help advance knowledge on heat acclimatisation, reduce bias in current measurement approaches, and provide an empirical foundation for JITAIs specifically designed for preventing heat strain.