Heat-related illness risk and associated personal and environmental factors of construction workers during work in summer

Heat-related illness (HRI) is a common occupational injury, especially in construction workers. To explore the factors related to HRI risk in construction workers under hot outdoor working conditions, we surveyed vital and environmental data of construction workers in the summer season. Sixty-one workers joined the study and the total number of days when their vital data during working hours and environmental data were recorded was 1165. Heart rate with high-risk HRI was determined using the following formula: 180 − 0.65 × age. As a result of the logistic regression analysis, age, working area, maximum skin temperature, and heart rate immediately after warming up were significantly positively related, and experience of construction was significantly negatively related to heart rate with high-risk HRI. Heart rate immediately after warming up may indicate morning fatigue due to reasons such as insufficient sleep, too much alcohol intake the night before, and sickness. Asking morning conditions may lead to the prevention of HRI. For occupational risk management, monitoring of environmental and personal conditions is required.

measure of HRI prevention 10,12 and WBGT is used as a heat stress index in many studies 5,20 . On the other hand, HRI involves complex interactions between environmental heat strain, clothing, and human thermal physiology 21 . Recently, a wearable sensor for monitoring workers' conditions has been developed 22 . US NIOSH recommends physiologic monitoring such as heart rate and/or core temperature of heat-exposed workers 16 . A wearable sensor enables workers to monitor their condition, making it possible to take a rest when they feel fatigue or to express that they feel unwell.
Continuous monitoring can make it possible to grasp the physical condition of all workers even from the remote place. Especially, its monitoring might be useful for construction workers because their working site is outdoor and they often work in small group or alone. To clarify HRI risk by using continuous monitoring data and personal background information is needed to lead effective management for construction workers. The aim of the present study was to explore the personal and environmental factors related to HRI risk using continuous monitoring data recorded by a wearable sensor in hot outdoor working conditions.

Results
Characteristics. The total number of person-time when vital data of the workers during working hours and environmental data were recorded was 1165. WBGT was measured for 42 days in Site 1 and 43 days in Site 2. On these 1165 person-time, a high heart rate indicative of HRI risk was observed on 102 person-time (8.7%).
We describe the workers' personal data in Table 1 and the environmental data in Table 2. The mean age (standard deviation) was 48.4 (14.0) years, and median number of working days (25-75 percentile) was 14   Table 3. When vital data as mean ± SD were compared between high and low risk groups, maximum heart rate (BPM) (165.9 ± 16.1 and 119.3 ± 16.0, p < 0.001), minimum heart rate (BPM) (58.1 ± 14.2 and 54.3 ± 6.4, p = 0.010) and heart rate immediately after warming up (BPM) (89.0 ± 23.2 and 81.8 ± 15.3, p = 0.004) were significantly higher in the high risk group. Skin temperature and energy consumption did not show a significant difference between the high and low risk groups. In a comparison of energy consumption among age groups, the age group of ≥ 60 years consumed significantly less energy (kcal), with a mean (SD) of 566.0 (176.4), compared to 744.9 (189.0), 771.0 (176.5), and 703 (184.5) in the age groups of < 40, 40-49, and 50-59 years, respectively.

Discussion
We investigated HRI risk in construction workers using their personal information, vital data and environmental data, and several factors were found to be risk factors of HRI. A previous study has recommended heat stress monitoring to prevent HRI 23 . Self-monitoring such as perceptual strain has been widely used to assess heat strain; however, it is said to do little to contribute to the protection of workers' health 23 . Self-monitoring of workers' conditions using their physiological responses has been determined to be most effective to prevent HRI 23 . Continuous objective monitoring is more effective for construction workers because their working site is outdoor and they often work in small group or alone. To use objective monitoring, early recognition of heat strain and management corresponds to the situation can be enabled even from the remote site. A wearable sensor for monitoring health conditions can assess physiological responses and its technology is continually developing 7,16,23 . Our Table 3. Vital data. Italic font describes significant difference (p < 0.05). Mean ± Standard deviation. p value was calculated by using Student's t test.  [24][25][26] . In addition, insufficient sleep has been shown to be strongly related to fatigue 27 . Heart rate increase is correlated to fatigue during physical exercise 28 . Furthermore, fatigue due to insufficient sleep might cause acute increases in heart rate. Therefore, heart rate immediately after warming up may be a good index for HRI risk assessment. We used Japanese radio calisthenics for warming up. These are simple exercises that can be performed regardless of age and can move the muscles and joints effectively according to the rhythm of the radio. It is considered a therapeutic exercise to promote health in Japan. The required time for calisthenics is approximately 3 min, and it has an exercise intensity of 4-4.5 metabolic equivalents (METs) 29 . In the present study, however, all subjects wore vital sensors after warming up. Wearing sensors before warming up and monitoring heart rate while warming up would have been more effective for evaluating workers' condition. After such evaluation, allocation of work could have been done according to workers' physical condition.
Aging has been reported to be significantly related to increased HRI risk. On the other hand, our previous study indicated that younger workers are at a higher risk of HRI 24 . Generally, exertional HRI occurs in healthy young people, whereas classical HRI is more likely to occur in the elderly 30 . One possible reason is that younger people tend to pay little attention to preventive measures for HRI 31 . Another possible reason is that younger age groups tend to engage in more physically intensive work, whereas older age groups tend to engage in less physically intensive work 25 . The current study demonstrates that energy consumption was significantly low in workers aged 60 years or older, but HRI risk did not decrease in that age group. Physical work capacity of a 65 year old can be reduced by up to ∼50% compared with an average 25 year old 32 . Older worker might feel higher physical strain than young worker. That is because HRI risk in elderly increase even in light workloads.
Regarding construction experience, it was significantly associated with reduced HRI risk. We previously reported that experience of outdoor manual work reduced risk of HRI 24 . Lack of experience is also reported to increase HRI risk 33 , and younger people tend to pay little attention to preventive measures for HRI 31 . Education for workers with little experience is necessary to raise awareness of HRI.
In the present study, maximum skin temperature was significantly associated with HRI risk. Ideally, core body temperature should be used to monitor HRI risk. To assess body temperature from the skin, measurements should be taken from more than four different points of the body 34 . Skin temperature may not increase proportionally to rapidly increasing core body temperature. The ratio of core-skin temperature can vary between 0.9/0.1 (core/skin) on rectal temperature ≤ 36.8 °C and 0.7/0.3 on rectal temperature ≥ 39.0 °C for the vasodilated and vasoconstricted skin, respectively 21 . However, rapidly elevating skin temperature indicates elevating core body temperature so that skin temperature is an effective index for HRI risk evaluation.
Working at Site 2 had a significantly higher OR for HRI risk in the present study. WBGT recorded at Site 2 was higher than that at Site 1, but maximum WBGT did not show a significant relationship to HRI risk. And more, most of the personal data did not show significant difference and the hypertension was recorded only in Site 1. Some preventive measures, such as having a rest space, allocating rest time, and encouraging water intake, were taken in both sites. These measures were thought to be the reason why a significant relationship was not observed between WBGT and HRI risk. The American Conference of Governmental Industrial Hygienists (ACGIH) determined the WBGT threshold by workload 35 . As we could not get the data of workload, we used the energy consumption data instead, which however did not reflect the workload accurately. By clearing the differences in detailed occupational conditions such as work load and management between the two sites, more effective measures for HRI may be employed. Further investigation is required.
This study has several limitations. First, we could not obtain data regarding the workers' clothing. HRI occurs as a result of complex factors, and clothing is strongly related to physiological performance 23 . Second, the wearable sensor used to obtain vital information in every 3 min in the current study had missing data due to mechanical reasons. We could not estimate the rate of change in skin temperature and its effect on heat strain.
In the present study, age, working area, heart rate immediately after warming up, maximum skin temperature and experience of construction were all significantly related to an increased heart rate indicative of HRI risk. Increased heart rate immediately after warming up may indicate morning fatigue due to, for example, insufficient sleep, drinking alcohol the night previously and sickness. Asking workers about their condition each morning may lead to effective preventive measures of HRI during work. For occupational risk management, monitoring environmental and personal conditions is essential. This study demonstrated the usefulness of continuous monitoring for HRI risk. Except for immediately after warming up, there might be pattern that predict HRI risk. We indicated a potential of continuous monitoring for the prevention of HRI. The importance of continuous monitoring should be increased in order to reduce the occurrence of occupational HRI.

Methods
Study design. This study was conducted as a cross-sectional study. We conducted an exploratory research about factors associated with heart rate of HRI risk among construction workers.
Study subjects. Study subjects were construction workers, who were employed by a subcontractor, and a prime contractor directed all working procedures and managed the workers' data. The prime contractor managed two sites where the workers were engaged in work during July 1st to September 30th, 2019, which was determined "a period to strengthen the prevention of occupational injury" in Japan. Prime contractor recruited to join this study and 84 (53 in Site 1 and 31 in Site 2) agreed to participate in the study. However, 23 were excluded due to monitoring failure and later refusal. Finally, we analysed vital data of 61 ( www.nature.com/scientificreports/ Data collection. We obtained personal, vital, and environmental data of 61 workers from the prime contractor. The personal data were age, sex, past medical histories (hypertension and diabetes), duration of career in construction (years) and days worked at the site being studied. The vital data were collected using a wearable sensor, LW-360HR (GISupply, Inc. Japan), worn on the left wrist of all workers. The sensor measured and recorded skin temperature of the ankle (°C), heart rate (BPM), and energy consumption (kcal) every three minutes, starting at the beginning of the working day immediately after warming up, and finishing at the end of the working day.
The environmental data were collected by using C-BB-15 cm (CLIMATEC, Inc. Japan) and HMP60 Humidity and Temperature Probe (Vaisala Corporation. Finland) at Site 1 and Davis Vantage Pro 2 (Keisoku Net Service Co., Ltd. Japan) at Site 2. The measurement method was based on ISO 7243: 1989 19 . WBGT was recorded every 10 min.
In the present study, heart rate with high-risk HRI was determined daily maximum heart rate using the following formula: 180 − 0.65 × age, based on the limit of heart rate in workplace determined by ISO 9886: 2004 34 . We determined high-risk group for the days scored high-risk HRI low-risk group was that heart rate did not archive to high-risk HRI.
Statistical analysis. We used R 3.6.1 for all statistical analyses. All data are described as mean (SD: standard deviation), median (25-75 percentile), or n (%). Personal data and environmental data, and heart rate with HRI risk (high risk vs. low risk) were compared between the two sites. To investigate the relationship between workload and age, we divided the subjects into four age groups (< 40, 40-49, 50-59, and ≥ 60 years) and compared daily energy consumption using analysis of variance (ANOVA). A logistic regression model was conducted to determine the factors that were related to heart rate with high-risk HRI. Age (continuous variable), working days, experience of construction, hypertension, diabetes, working site, employer (subcontractor or sub-subcontractor), maximum skin temperature, energy consumption, heart rate immediately after warming up, maximum WBGT (< 25, 25-28, 28-31, > 31 °C) and WBGT difference in a day were selected as dependent variables. WBGT was categorized according to the guidelines for prevention of heat disorders for sporting activities 36 . The adjusted ORs for heart rate with high-risk HRI, and their 95% confidence intervals (95% CI), were calculated.

Ethics.
We confirmed that all research was performed in accordance with Ethical Guidelines for Medical and Health Research Involving Human Subjects 37 . This study was approved by the Ethics Committees of Fukushima Medical University (Application No. 2019-185). The prime contractor explained the study to their workers and obtained written informed consent for participation in the study and data collection. All data was anonymized by the prime contractor before being passed on to us.

Data availability
No additional data are available for this study. However, inquiries concerning the data may be made to the corresponding author.