A short review on physiological monitoring during working activities

Introduction: Temperature extremes, load carriage, inadequate sleep, information overload, dehydration, and impaired nutrition, are common risks associated with many occupational activities, including those for whom optimal functioning is critical at all times. These safety-sensitive occupations include firefighters, first responders, police officers, physicians, airline pilots, soldiers, and those operating heavy machinery. In any of these cases, the resulting interaction between occupational stress and individual susceptibility to illness demands careful management. This represents a dual challenge to organizations responsible for the well-being of personnel who engage in strenuous tasks, imposing requirements to be vigilant for or, even, curtail situations that result in high physiological strain. The emergence of wearable physiological monitoring technologies could prove advantageous in this regard. Purpose: To our knowledge, no review gathering the applicability of these systems within occupational groups has been conducted. Therefore, this review aims to summarize current progress in the development of wearable physiological monitoring systems for occupational applications. Methodology: Five databases were accessed (SCOPUS, PubMed, Science Direct, Academic Search Complete and Web of Science) and a total of 12 keywords were combined to develop a search on journal articles from January 2014 to January 2019. Study eligibility based on active workers participants and assessment methods not interfering with normal tasks development and involving harmless procedures. Furthermore, investigations conducted with prognostic health-related goals were filtered. Results and Discussion: Nineteen studies were analyzed in this review. In general, their goals were directed to quantifying the impact of specific physically demanding tasks or validating newly proposed methods for classifying the effects of different levels and workloads of occupational tasks based on workers’ physiology. Identified occupational groups mostly included construction workers, drivers, and firefighters. Retrieved papers highlighted the importance of field monitoring to provide a chance to timely detect any abnormal condition in the worker’s physiology that might be affected by working conditions or environmental stresses. Conclusions: Wearable sensors proved to be a valid tool for assessing physiological status in simulated and real working environments. Future research perspectives should be focused on validation of standardized procedures within bigger samples and involving a variety of safety-sensitive professions. Finally, based on physiology and novel computational techniques, it was observed that further developments should be concentrated in the algorithms that allow low-cost sensors to be used in operational settings to provide the continuous subjects’ status promoting to sustain their given tasks in a safer and healthier way.


INTRODUCTION
Athletes must compete with very high metabolic demands in outdoor temperature extremes. Miners and steelworkers are exposed to high heat conditions (Butlewski, Dahlke, Drzewiecka, & Pacholski, 2015;Chen, Chen, Yeh, Huang, & Mao, 2003). Firefighters, first responders, and soldiers often wear personal protective equipment that imposes additional thermal burdens from insulation and extra carried weight (Buller, Welles, & Friedl, 2018) while exposed to extreme environments, inadequate sleep, information overload, dehydration and even impaired nutritional status (Lieberman et al., 2005). As a result, decrements in workplace performance, health, and safety are typically encountered. This represents a dual challenge to organizations responsible for the well-being of personnel who engage in strenuous tasks, imposing requirements to be vigilant for, or even curtail, situations that may result in high physiological strain in healthy personnel and also to identify and protect vulnerable individuals. The emergence and increasing interest in wearable physiological monitoring devices can help to address this challenge but requires that the right questions are asked in sourcing, developing, validating and applying such technologies (Stacey, Hill, & Woods, 2018). Wearable physiological monitoring can provide predictions about an individual's health and performance from their real-time physiological state (Raskovic, Martin, & Jovanov, 2004). However, available systems mostly do not satisfy the requirements for occupational use. Even when they offer more than raw physiological data, computed information is usually based on proprietary algorithms that cannot be properly reviewed and validated. The critical component of a real-time physiological monitoring system is the algorithm that turns data into useful and actionable knowledge for a worker or a small unit leader. Useful information from these systems is defined as vitally important alerts that can be acted on to affect the outcome of operations and improve safety and effectiveness (Friedl, 2018). To our knowledge, no comprehensive search of the literature has been developed in this regard. Thus, a review is proposed to find relevant information about the current progress of these physiological monitoring systems and their potential applications for occupational settings.

METHODOLOGY
This review was limited to research articles and articles in press published in peer-reviewed journals in the English language. It was conducted in Scopus, PubMed, Science Direct, Academic Search Complete and Web of Science databases and narrowed to articles published between January 2014 and January 2019. The 12 identified keywords were combined as follows: ( ( ( "physiolog*monitor*" ) OR ( "noninvasive monitor*" ) OR ( "medical monitor*" ) OR ( " wearable sens*" ) ) AND ( ( assessment )

OR ( occupational ) OR ( model ) OR ( fatigue ) OR ( algorithm ) OR ( worker ) OR ( training ) OR ( "physical exertion" ) ) )
The search focused on investigations developed within working-age active participants and incorporated both females and males with no additional restrictions. Study selection was based on three phases of exclusion: applying filters from databases, eliminating repeated records and analyzing each article individually to remove studies in which no prognostic health objectives were pursued, procedures were not developed within active working-age subjects, or no noninvasive methods were used. Finally, inclusion criteria were those investigations in which noninvasive objective physiological assessment methods were applied, and measurements were developed during real or simulated working activities.

DISCUSSION
This review focused on the assessment of continuous physiological responses of occupational tasks. Mostly, studies' goals were directed to quantifying the impact of physically demanding activities or validating newly proposed methods for classifying the effects of different levels and workloads of occupational tasks based on workers' physiology. In general, they highlighted the importance of field monitoring to timely detect any abnormal condition of the worker that might be affected by working or environmental stresses (Hwang et al., 2016). Since most of the investigations were developed within real working environments, the feasibility of performing continuous assessments during regular occupational situations was indeed demonstrated. Among retrieved articles, a variety of occupational groups were assessed. Studied professions included construction workers, drivers, firefighters, pilots, ironworkers, cleaning staff and law enforcement personnel of military sites. The encountered interest on some of these occupations was justified as they correspond to safety-sensitive professions, in which effective human performance is crucial to a successful outcome (Barger, Lockley, Rajaratnam, & Landrigan, 2009). Furthermore, a notable focus was evidenced on construction workers (included in five out of the 19 articles). Construction work typically involves physically demanding tasks performed in harsh environmental conditions, which can cause fatigue and lead to poor judgment, poor quality of work, increased risk of accidents and reduction in productivity (Aryal et al., 2017). Consistently, among retrieved papers, the clear interest was on assessing fatigue (Aryal et al., 2017;Hwang et al., 2016;Lee et al., 2017;L. Yang et al., 2018) and developing methods for preventing risks of accidents (Antwi-Afari et al., 2018). On the other hand, observing monitoring methods (Table  1), several wearable sensors were identified, and their simultaneous use for multivariable measurement was a clear tendency. Noteworthy is the fact that the only wearable system used in more than one selected study was the Equivital LifeMonitor (Hidalgo Ltd., Cambridge, UK). As Mehta et al. (2017) indicated, the validity, reliability, and applicability of this system for sleep and ambulatory monitoring of multiple physiological parameters during construction and firefighting work have been previously demonstrated (Gatti, Schneider, & Migliaccio, 2014;Liu, Zhu, Wang, Ye, & Li, 2013;Savage et al., 2014). Among selected studies, three included this system to obtain physiological data from operators in drillship (Mehta et al., 2017), wildland firefighters (Sol et al., 2018) and law enforcement personnel (Yokota et al., 2014). In all cases it proved advantageous for obtaining measurements in field conditions and, HR assessment was observed as the focus assessed variable. Thus, it can be inferred that the tendency on the usage of validated procedures is maintained and future perspectives could be oriented to test the other referred methods within bigger samples and during real-life operations. Finally, observing physiological variables, cardiac responses to specific occupational activities were the most considered assessment goal among retrieved papers. HR was included in 11 out of the 19 final articles. Based on current work physiology literature, this can be explained as this variable has shown to be sensitive to changes in physical and mental fatigue (Borg, Hassmén, & Lagerström, 1987;Hankins & Wilson, 1998), as well as sleep and circadian issues (Carney et al., 2014;Kang et al., 2015). Collectively, findings not only suggest the relevance of multivariable approaches that include monitoring of cardiac responses along with validated fatigue indicators (thermal responses and scales of perceived exertion) but also confirm the need of inclusion of other variables such as metabolic equivalents (Lee et al., 2017) and respiratory signals (Boon-Giin et al., 2014;Fu et al., 2016). Finally, despite being assessed in few of the selected articles, approaches such as foot plantar distribution (Antwi-Afari et al., 2018) also proved to be beneficial for prognostic goals in occupational settings and could be included in future research works. Lastly, observing data processing methods, support vector machine classifiers and algorithm-based applications for smartphones suggest the evolution of data management with the tendency of assessing traditional fatigue physiological indicators (HR, acceleration, respiration signals) but through newly available computational techniques.

CONCLUSIONS
In this review, the evidence of current progress in the development of physiological monitoring systems and their applications for occupational settings is compiled. Selected studies indicate the feasibility of devices using physiological signals to examine the impact of occupational tasks and improve the management of health-related negative consequences in the near future. With a basis in physiology and application of principled computational techniques, it was demonstrated that future perspectives should be focused in the algorithms that allow simple low-cost sensors to be used in operational settings and provide the continuous subjects' status promoting to sustain their given tasks in a safer and healthier way.

Funding
This work was accomplished during the period of a research scholarship granted by the Doctoral Program of Occupational Safety and Health of the University of Porto.