Risk Factors for Reaching Core Body Temperature Thresholds in Florida Agricultural Workers : Journal of Occupational and Environmental Medicine

Secondary Logo

Journal Logo


Risk Factors for Reaching Core Body Temperature Thresholds in Florida Agricultural Workers

Mac, Valerie PhD, APRN, FNP-C; Elon, Lisa MA; Mix, Jacqueline PhD, MPH; Tovar-Aguilar, Antonio PhD; Flocks, Joan MA, JD; Economos, Eugenia; Hertzberg, Vicki PhD; McCauley, Linda PhD, RN

Author Information
Journal of Occupational and Environmental Medicine 63(5):p 395-402, May 2021. | DOI: 10.1097/JOM.0000000000002150


Agricultural workers and other outdoor workers have been identified as a vulnerable occupational group with increased risk of adverse health outcomes from rising global temperatures.1 According to the Centers for Disease Control and Prevention (CDC), over 400 workers in agricultural and nonagricultural settings suffered heat-related deaths during the time period of 1992 to 2006.2 Most were workers between the ages 20 to 54 and a predominant percentage of the deceased were agricultural workers, carrying a 20-fold increased risk of heat-related death as compared with all other US occupational groups.2 With a limited ability to avoid heat exposure in agricultural work environments, it is important to identify key risk factors that place a worker at increased risk for core body temperatures over the recommended limits in order to avoid heat-related morbidity and mortality.

In order to provide guidance to states that have no mandated heat standard, the National Institute for Occupational Safety and Health (NIOSH) updated the Criteria for a Recommended Standard: Occupational Exposure to Heat and Hot Environments report in 2016, including new which included new guidelines for core body temperature thresholds in occupational workers.3 According to the Occupational Safety and Health Association (OSHA), heat exhaustion begins at 38.0 °C.4 The American College for Governmental and Industrial Hygienists (ACGIH), workers without medical clearance should not have core body temperatures that exceed 38.0 °C (100.4 °F) for extended periods of time,3,5 however specific criteria for what constitutes “extended periods of time” is not defined. Also specified by ACGIH and supported by NIOSH, the second tier physiologic limit for core body temperature is 38.5 °C (101.3 °F) a limit specific for workers who are known to be acclimatized and are medically cleared and monitored.3,5

Florida, a state second only to California in the gross value of fresh produce production,6 experiences subtropical conditions with extreme heat and humidity. This environment is not expected to improve given a general warming trend during the last three decades.7 The state of Florida does not mandate specific mandated heat rules or standards for workers. In our previous pilot study of 43 fernery workers in Florida during the summers of 2012 and 2013, participants experienced core body temperatures that approached 38.0 °C on more than half of the workdays studied, and female participants had a five-fold higher risk for this core temperature elevation when adjusting for physical activity.8 Building on these pilot data, we examined 257 male and female fernery, crop, and nursery workers participating in the Girasoles (Sunflower) study (R01OH010657) to describe the occurrence of core body temperatures that exceed recommended Tc thresholds among Florida agricultural workers and to examine risk factors associated with these events. We examined the relationship with environmental heat, physical activity, and dehydration as possible drivers of Tc. Other risk factors considered were workers’ sex, hydration practices, body mass index (BMI), age, the presence of comorbidities, length of the workday, type of agricultural work, and years of experience in agriculture.


Study Design and Population

The Girasoles (Sunflower) study was a descriptive investigation of Florida agricultural workers in Florida aimed to describe physiologic responses to heat stress. Agricultural workers are employed in a variety of agricultural settings including ferneries, nurseries, and field crops, were recruited from five locations in central and south Florida. Each participated agreed to have their physiological status monitored for up to 3 workdays and provided daily bio-specimens collection. This study was a partnership between university researchers at Emory University and the Farmworker Association of Florida, a grassroots, statewide, community-based membership organization with a 35-year history working with agricultural workers in Florida. The Farmworker Association of Florida (FWAF) has more than 10,000 members statewide who work in various agricultural operations including ferneries, nurseries, and field crops.

Recruitment and Eligibility

Trained community workers from the FWAF used outreach strategies to recruit a convenience sample of agricultural workers during the summer months from 2015 to 2017. Approximately 70% of the workers who attended group enrollment discussions and screening for the study were enrolled. Workers were eligible for the study if: (a) they were 18 to 54 years of age at the time of study; (b) spoke English, Spanish, or Haitian Creole speaking; and (c) had worked in agricultural settings for at least 1 month prior to study participation. Workers were excluded if they had a history of Type 1 diabetes or were currently pregnant. Additional exclusions pertained to the ingestion of a core body temperature sensor pill and included: (a) ever diagnosed with any disorders of the esophagus, stomach, or intestines; (b) ever had surgery of the esophagus, stomach, or intestines; (c) have trouble swallowing; (d) have chronic constipation; or (e) have a pacemaker. This study was approved by the Institutional Review Board of Emory University (IRB00075192). Participation of human subjects did not occur until after informed consent was obtained.

Data Collection


Data collection in each agricultural community occurred at local FWAF offices. The protocol consisted of a baseline assessment followed by up to 3 monitored workdays. At baseline, trained interviewers administered an occupational heat-related illness (HRI) survey in the participant's primary language. This instrument was adapted from a previous survey used with agricultural worker populations9 and included information on socio-demographics (eg, age, sex, nationality, marital status, years of education, health histories), and work characteristics (eg, years worked in agriculture, days worked per week, hours worked per day, hydration habits).


On workdays, participants reported to the FWAF office before and after their work shift. Before work, participants provided a urine samples for measurement of urine specific gravity using a point-of care Osmolality Meter (Osmocheck®) (Vitech Scientific Ltd., West Sussex, UK) and a fingerstick blood sample for the measurement of blood glucose using the iSTAT Handheld Blood Analyzer Point of Care system with a Chem8+ cartridge (Abbott Laboratories, Abbott Park, IL). Participants were fitted with equipment that was concealed by their clothing to monitor core body temperature, heart rate, and physical activity every 30 seconds during the workday. Continuous core body temperatures (Tc) were captured every 30 seconds from an ingestible temperature sensor using a CorTemp® monitoring device (HQInc., Palmetto, FL), which also recorded simultaneous, heart rate from the Polar® T31 HR transmitter strap. Workers ingested the temperature sensor capsule during the evening before they reported for their workday data collection, as pilot data showed that ingesting the capsule the evening before resulted in a high probability of a sensor reading throughout the next workday.10 Each morning, field staff tested for the presence of the ingestible core temperature sensor; if the signal was not found, it was assumed the capsule had been eliminated and the participant was given another sensor to ingest with food and water.

Work activity level was measured using continuous monitoring with a triaxial accelerometer ActiGraph™ GTX3+ (ActiGraph, LLC, Pensacola, FL). The ActiLife 6 (ActiGraph, LLC, Pensacola, FL) software package was used for initializing and calibrating the ActiGraphs as well as downloading raw ActiGraph counts. Participants wore the accelerometer on a belt around their waist placed above the illac crest and along the midaxillary line.

Ambient temperature and relative humidity were obtained each day from the Florida Automated Weather Network (FAWN),11 which collects environmental data every 15 minutes at monitoring stations in each study community. To quantify exposure specific to each participant, weather data summaries were based on their specific work hours, and we report the maximum temperatures for each person-workday. We calculated wet-bulb globe temperature (WBGT)12 and heat index (HI). HI was calculated by using the National Weather Service algorithm.13 Heat index from regional weather data in Florida has been shown to accurately reflect worksite environmental conditions and stratify heat stress risk.8

Statistical Analysis

Primary agricultural work type was classified into three groups: ferneries, nurseries, or field crops according to workers’ self report of current work activities. Body mass index (BMI) was calculated from measured height and weight (kg/m2); workers were defined as obese if their BMI was more than or equal to 30, based on the World Health Organization criteria.14 Blood pressure was collected using an electronic blood pressure cuff at the baseline visit. History of diabetes or hypertension was self-reported from participants by asking if they had been diagnosed by a healthcare provider. Baseline data were summarized by calculating means and standard deviations for continuous variables and frequency counts and percentages for categorical variables.

Work activity level was calculated using raw accelerometer counts from the Actigraph using three planes of motion including vertical, anteroposterior, and mediolateral to quantify the number of total minutes spent in moderate to vigorous levels of activity for each workday. For this calculation, we defined moderate to vigorous activity as 2690 counts per minute (CPM) or greater.15,16 Data processing for our raw accelerometer count has been described in a previous paper describing physical activity of agricultural workers in the Girasoles (Sunflower) Study.17

Core body temperatures (Tc) were only considered to have exceeded the recommended limit (38.0 °C or 38.5 °C) on a study day if two consecutive temperature readings (1 minute of elapsed time) were above these thresholds. Participants were classified as ever exceeding a threshold during the study if there was an occurrence on at least one of their observed workdays. The time spent above a threshold was calculated by summing all instances when a participant's Tc met the criteria. Core temperature and heart rate files were considered useable if the number of missing readings were less than 20% of the total readings for that workday. Data were examined for physiologic plausibility utilizing a core temperature maximum value of 40.5 °C and heart rate range of 30 to 210 beats per minute (bpm) as discussed by Hertzberg, Mac.18 We restricted analyses to participants with at least one monitored workday and at least one useable core body temperature file.

To account for the nested data structure, generalized linear mixed modeling (GLMM) was used for simple and multivariate inferential analyses. The intraclass correlation coefficient was 0.19 for Tc38 and 0.31 for Tc38.5. Odds ratios and 95% confidence limits were calculated to examine the risk factors for exceeding Tc38 and Tc38.5. For each outcome, the set of a priori predictors included demographics (age, sex), health-related variables (BMI, history of hypertension or elevated BP during study, history of type II diabetes or elevated glucose during study, post-work USG [and separately, pre-work USG]), work-related variables (agricultural work type, years working in US agriculture, workday duration, minutes of moderate-vigorous physical activity during work hours), and environmental heat conditions (maximum HI during work hours).

Given that several predictor variables may be mediators,19 we used Baron and Kenney's20 method to examine if there were mediation effects in the following pathways: HI → urine specific gravity (usg) before and after work → Tc; HI → physical activity → Tc; physical activity → urine specific gravity (usg) before and after work → Tc. Additionally, the assumptions of logit linearity for continuous predictors and absence of multicollinearity were confirmed. We used the R package lqmm21,22 for linear quantile mixed models to examine the relationship between the median amount of time exceeding Tc thresholds and the following predictors: age, sex, BMI, agricultural work type, years working in US agriculture, workday duration, post-work USG, minutes of moderate to vigorous physical activity during work hours, and maximum HI during work hours.

To further characterize core body temperature responses to heat stress, temperatures were plotted by time of day, with data summarized by quantiles (50th, 75th, 90th, and 95th) at each 30-second time point. Analyses were performed with SAS version 9.4 software (Cary, NC) and R.23 Statistical significance was evaluated using alpha = 0.5.


A total of 257 participants were recruited during the summers of 2015 and 2016. After limiting the dataset to those participants who had at least 1 monitored workday as well as at least one usable core body temperature file, 221 participants (465 workdays) remained for analysis. Table 1 shows that the data collection in the five central and south Florida communities occurred when there was substantial exposure to moderate and high-risk heat indices. All participants worked at least two-thirds of their study period with environmental conditions requiring precautions to prevent HRI which begins at a heat index of 80 °F (Table 1). The mean heat index from 6 AM to 6 PM during the study period was 90.1 °F (SD = 5.6).

TABLE 1 - Percent of Study Period Under Various Heat Index Risk Levels, Based on 15-minute FAWN Readings Between 6 AM and 6 PM, by Study Location
Heat Index, °F
<80 80–<91 91–<103 ≥103
Location Primary Work Latitude Study Months Low Risk Moderate Risk High Risk
Pierson Fernery 29.2 N May, Jun, Jul 28 24 33 15
Homestead Nursery 25.7 N Jul, Aug 24 12 52 13
Apopka Nursery 28.7 N Jun, Jul 22 24 49 5
Immokalee Field Crops 26.4 N Aug, Sep, Oct 37 18 44 1
Fellsmere Nursery/Field Crops 27.8 N May 29 41 30 0
These heat indices are calculated based on shade conditions; participants working in the sun or those wearing heavy or impermeable protective clothing would be at higher risk than indicated.

On an average day, 49% of participants exceeded Tc38 and 10% exceeded Tc38.5; there was little change in the rates on subsequent monitoring days (Table 2). Over the course of up to 3 workdays, 67% (147/221) of participants exceeded Tc38 and 16% (35/221) exceeded Tc38.5 on at least 1 day. An additional 14% reached, but did not exceed the Tc38 on at least 1 day. The proportions did not vary significantly among fernery, nursery, and field crop workers (Table 3).

TABLE 2 - Percent of Participants With Core Body Temperatures Exceeding the Recommended Limits on Observed Workdays
NIOSH 2016 OSHA 2015
>38 >38.5 ≥38 ≥38.5
Overall N %
First workday 177 51 8 67 13
Second workday 150 49 12 65 14
Third workday 142 45 11 60 15
 Average percent 221 49 10 64 14
 Percent exceeding on at least 1 day 221 67 16 81 21
By work type
 Average percent
  Fernery 61 48 8 67 12
  Nursery 89 47 8 65 12
  Crop 67 53 15 62 17
 Percent exceeding on at least 1 day
  Fernery 61 71 15 89 21
  Nursery 89 61 10 80 16
  Crop 67 70 24 76 28
Sample sizes are smaller than the full cohort due primarily to loss of core temperature files to technical issues. 221 participants had at least one day's usable core temperature data.
Averaged using generalized linear mixed models.
Four participants could not be classified into one of the three work types.

TABLE 3 - Characteristics of Florida Agricultural Workers by Ever Exceeding Core Body Temperature Thresholds (n = 221); Girasoles Study 2015–2017
Core Body Temperature
>38.0 °C >38.5 °C
Characteristics No n = 74 Yes n = 147 No n = 186 Yes n = 35
Mean (SD), Median [Q1, q3], or % (n)
  Female 55% (41) 67% (98) 65% (121) 54% (19)
  Male 45% (33) 33% (49) 35% (65) 45% (16)
  Mexico 59% (44) 71% (105) 68% (127) 63% (22)
  Central America 14% (10) 12% (18) 11% (12) 20% (7)
  Caribbean Islands 24% (18) 15% (22) 19% (35) 14% (5)
  United States 3% (2) 1% (2) 2% (3) 3% (1)
 Age, yrs 38 (9) 38 (9) 38 (9) 39 (8)
 Years of education, yrs 7 (3) 7 (4) 7 (3) 7 (4)
 History of hypertension (Hx HTN) 8% (6) 8% (12) 9% (16) 6% (2)
 Hx HTN or elevated blood pressure reading 22% (16) 31%(44) 28% (51) 26% (9)
 History of diabetes (Hx DMII) 9% (7) 7% (10) 8% (14) 9% (3)
 Hx DMII or elevated glucose reading 22% (16) 19% (28) 19% (34) 29% (10)
 Body mass index 27.4 (4.3) 29.2 (4.9) 28.5 (4.9) 29.2 (4.4)
 Body fat percentage
  For male participants 22.3 (5.5) 22.6 (7.5) 22.0 (6.4) 24.5 (7.7)
  For female participants 34.2 (5.1) 36.0 (5.6) 35.3 (5.5) 36.0 (5.2)
Work environment
 Wet bulb globe temperature, °F, maximum 80 [78, 81] 80 [79, 81] 80 [78, 81] 81 [79,81]
 Heat index, °F, maximum 101 [98, 103] 101 [98, 102] 101 [98, 103] 103 [97, 102]
 Basic characteristics
  Years worked in agriculture, yrs 12.3 (8.5) 11.8 (8.0) 12.6 (8.2) 8.9 (7.1)
  Mean hours worked per day, hrs 8.1 (1.7) 7.4 (1.9) 7.6 (1.8) 7.6 (1.9)
  Self-reports drinking more water at work 96% (71) 97% (143) 97% (180) 97% (34)
  Self-reports drinking sport drinks at work 66% (48) 69% (102) 67% (124) 74% (26)
  Self-reports drinking energy drinks at work 14% (10) 18% (27) 16% (29) 23% (8)
  Self-reports drinking more soda at work 45% (33) 52%, (76) 50% (93) 46% (16)
  Self-reports drinking more juice at work 41% (30) 37% (55) 38% (71) 40% (14)
  Urine SG mean§, before work 1.020 (0.006) 1.020 (0.005) 1.019 (0.005) 1.022 (0.006)
  Urine SG mean§, after work 1.022 (0.008) 1.024 (0.006) 1.023 (0.007) 1.026 (0.006)
Over 1 to 3 days of observation, the participant had at least 1 day on which the Tc threshold was exceeded. Values in bold are significantly different (P < 0.05), tested using generalized linear mixed models.
Based on average maximums experienced per participants during their work hours.
Drink questions queried on baseline survey: “When it is hot out, do you drink more….”
§Participant mean value over their workdays.

In multivariable models adjusting for demographic, health, work, and environmental characteristics, we found that having worked longer in US agriculture was protective against experiencing a Tc over 38.0 °C (per 5 years of experience: OR 0.79, 95% CI 0.66, 0.95) (Fig. 1). Higher BMI increased the odds of exceeding Tc38 (per five units: OR 1.34, 0.997, 1.81), as did more time spent in moderate to vigorous activity (per hour: OR 1.41, 95% CI 1.20, 1.64). The environmental heat index was associated with 23% increased odds of exceeding Tc38 per 5 °F increase (OR 1.22, 95% CI 0.99, 1.50). Compared with fernery workers, field crop workers were more likely to exceed Tc38 (OR 2.15, 95% CI 1.00, 4.61). Sex, age, histories of hypertension and diabetes, work duration, and post-workday USG were not associated with an increased risk of experiencing a Tc > 38. In a separate model, pre-workday USG was also not associated with Tc38 (per.01 unit: OR 0.80, 95% CI 0.57, 1.14). No evidence of mediation effects, based on lack of significant relationships and lack of beta coefficient changes, were found in the suspected causal pathways we examined: HI → urine specific gravity (usg) before and after work → Tc; HI → physical activity → Tc; physical activity → urine specific gravity (usg) before and after work → Tc.

Odds ratios for exceeding Tc thresholds, estimated using multivariate mixed effects logistic regression models containing demographic, health, and work variables. Tc, core temperature.

Different factors predicting risk for Tc38.5 or greater were found. Adjusting for covariates, men had 2.25 greater odds of exceeding Tc 38.5 than women. Odds of exceeding Tc38.5 increased 2.31 times as BMI increased five units (95% CI 1.42 3.76). Working longer in US agriculture was protective against Tc exceeding 38.5 °C (OR 0.56, 95% CI 0.40, 0.78). Increasing time spent in moderate to vigorous activity was marginally associated with increased the odds of exceeding Tc38.5 (per hour: OR 1.26, 95% CI [0.996, 1.60]). Participants with a history of hypertension or observed high blood pressure reading had lower odds of exceeding Tc38.5 (OR 0.37, 95% CI 0.14, 0.96). Environmental heat, age, length of workday, and primary type of work were not significant predictors of Tc38.5. Post-workday dehydration was not significantly associated with exceeding Tc 38.5 (OR 1.33, 95% CI 0.77, 2.37) (Supplemental Table 1, https://links.lww.com/JOM/A876).

The median duration of time workers temperatures exceeded 38.0 °C was 59 minutes [Q1 19 to Q3 131], which occurred on 228 workdays (Supplemental Figure 1, https://links.lww.com/JOM/A877). No significant associations were found between predictor variables and median minutes of Tc > 38. However, when examining the 48 days with workers exceeding Tc > 38.5, the median duration of minutes exceeding Tc38.5 was 13 minutes and the amount of time over the threshold of 38.5 was associated with work type (93 minutes longer for crop workers than fernery workers [95% CI 23, 163]), and higher maximum heat index (increase of 21 minutes per 5 °F increase [95% CI 2, 40]). Furthermore, among those exceeding the Tc thresholds, on average they first exceeded the Tc38 threshold mid-morning (10:38, IQR 9:09 to 12:09), and Tc38.5 about a half hour later (11:10, IQR 9:34 to 13:16). Figure 2 displays the pattern of Tc over the course of the workday, summarized by four quantiles at each reading (every 30 seconds). At any time point after about 8 AM, 5% of the total sample was exceeding the Tc38 threshold; after about 1:30 PM, 25%.

Tc over the course of the workday. The indicated percentile was calculated for each 30 seconds reading. The arrow indicates when the percentile exceeded 38 °C; the median value never exceeded that threshold. At any given reading after about 9:15 AM, 10% of the sample was exceeding the Tc38 threshold. Tc, core temperature.


The results of this study provide evidence of the extent of elevations of core body temperature over the recommended limits in a substantial proportion of Florida agricultural workers. This study also identifies that limits are exceeded early in the day, and provides evidence for risk factors associated with surpassing these thresholds for agricultural workers in the Southeastern United States.

Workers in our study spent the greatest amount of time in conditions classified as moderate risk according to the heat index on observation days. On an average workday, half of participants experienced a Tc over T38 one out of 10 workers exceeding Tc38.5. Alarmingly, the Tc38 limit was surpassed as early as 8 AM by 5% of the sample. Of those that surpassed the Tc38.5 limit, half of those participants had reached that limit by 11:10 AM, substantially before the hottest time of the day. These results have occupational health significance because morning hours are considered to be the safest for workers and supervisors have been observed to cease work in afternoons of very hot days. Currently, employers in Florida are not required to provide more water, rest, or shade on days that are classified as moderate to high risk. OSHA recommends that employers take several actions to prevent heat illness at moderate heat index risk levels. These include frequent, scheduled breaks in cool, shaded areas; reminders to rehydrate; review of key heat-related illness information with workers; and a policy instructing supervisors or coworkers to use a buddy system to look for signs of HRI.24 For those working in the direct sun, OSHA recommends that employers monitor workers very closely for signs and symptoms of HRI and to design schedules for work and rest intervals. In this study our research participants did not offer any examples of changes employers made in work patterns, water, rest and/shade even when temperature levels were putting workers at moderate risk for HRI.

With nearly half of participants exceeding Tc38 in our study, this highlights the importance of mandated HRI prevention strategies in Florida and across the United States. The California Heat Illness Prevention Study (CHIPS) examined core temperature in a sample of California agriculture workers utilizing a similar, but shorter 1-day (vs our 3-day) biomonitoring25 protocol and found that nearly half of their participants had core temperatures that exceeded Tc38. They also found that more than 8% of the workers studied experienced a core body temperature of Tc38.5 or more,25 which is very similar to our finding that 10% of workers exceeded Tc38.5. Our study of Florida workers differed from the Moyce et al25 study in that we sampled predominantly female agricultural workers (63%) and participants self-identified as originating from not only Mexico, but also Central American, Caribbean, and US born workers. The mean HI during the California study was 84.7 °C,26 whereas it was 90.1 °F in our Florida study. The age of workers in both samples was similar.

The median amount of time that worker core temperatures exceeded T38 was 56 minutes per workday with 131 and 19 minutes at the 75th and 25th percentiles, respectively (Fig. 2). These findings are similar to those reported by Yeoman et al,27 in a sample of mine workers who had a median duration of nearly 1 hour of cumulative minutes over Tc38 found in mine workers by NIOSH researchers in a recent study. Both studies found wide variation in the number of cumulative minutes that workers exceeded Tc38. While a much smaller proportion of workers exceeded 38.5 °C as compared with 38.0 °C in our study (10% compared with 49% on an average workday), the frequency still reflects that one in 10 workers spent a median of 13 minutes at the level during which work should have ceased. These data indicate that policymakers and employers should be implementing interventions to alleviate heat-related morbidity and mortality.

The continuous monitoring in our study permitted examination of the pattern of heat strain occurring in workers during the workday. Heat strain gradually increased during the morning hours, peaking in early afternoon and remaining elevated throughout the afternoon. This pattern may be due to the higher levels of physical activity exerted by the participants in the morning, and then higher levels of environmental heat in the afternoon.28 Noting the rise in core temperature beginning by midmorning, employers and employees should be informed that the risk of heat strain and HRI may occur earlier in the day than generally expected; the afternoon hours should not be the only time of day for rest and shade interventions.

Heat protection policies in the workplace should include content that engages workers who are likely at higher risk. Piece-rate pay structures may also play a role in keeping physical activity high even as temperatures soar because workers may chose to not slow down or take an afternoon break to maximize their earnings. Workers are exceeding the recommended core temperature thresholds early in the day, indicating that consistent application of heat interventions should begin at the beginning of the workday. Our study also found that different types of agricultural work might increase the potential for HRI. Exposure to direct sun should also be considered.

Our findings suggest that less experience in US agriculture is a risk factor for Tc385. Lack of acclimatization can place an individual at increased risk,3 so new agricultural workers should be monitored closely. The strenuous nature of agricultural worker may result in workers self-selecting out of agricultural over time, or worse, being unable to continue to work. Conversely, older workers might be at an advantage due to acclimatization over time as well as knowledge of the work and risk-reduction strategies. In addition, the duration of time spent in moderate to vigorous activity was identified as a modest predictor of Tc38.5, so it is possible that the increased risk of Tc38.5 for less experienced workers can be partially attributed to a lack of regular breaks and self-pacing. A recent worker death in Georgia29 that occurred in a young male who had been working in agriculture for less than 1 month highlights a dire and avoidable scenario that must be prevented in the future through deliberate interventions. This worker was working in direct sun while picking tomatoes.

In the current study, a higher BMI was a risk factor for exceeding Tc38 and Tc38.5. Obesity is a widely recognized as a risk factor for HRI.30–32 The mechanism for this increased risk is a diminished ability to dissipate heat generated from environmental heat exposure and from physical activity. Due to the involvement of gut integrity in the pathogenesis of heat stroke,33 the potential interactions between obesity, inflammation, and gut health via the gut microbiome may provide further information to guide interventions to address this risk factor in conjunction with improving BMI.

Surprisingly, we found no indication that previously diagnosed comorbid conditions of diabetes mellitus type II or hypertension were risk factors for temperature elevation over Tc38, and even were protective against Tc38.5. Increased risk of heat illness for those with comorbid conditions may be more related to age (over 60) and lifestyle factors in populations that do not engage in physically active work,34 which is very different from the population in the current study. Nevertheless, health care providers should ask their patients about their work and make recommendations regarding precautions to take in the heat based on the patient's prescribed medications that could alter body heat regulation.35 Self-guided limitation of physical activity intensity could be a possible explanation for the protective effect found for comorbid conditions and Tc38.5.

In 2019, members of the US House of representatives presented a bill to require OSHA to develop a mandatory federal standard regarding heat protections for both indoor and outdoor workers.36 If this bill is passed, OSHA would at the least, have to propose an interim rule within the next 2 years. A heat illness prevention bill, Florida HB 1285,37 which would institute requirements for certain employers such as providing annual heat prevention training was under consideration in the state of Florida until being indefinitely postponed and withdrawn in the Summer of 2019. With growing interest at the federal level and ideological chasms in some states, it is crucial to examine evidence of heat illness risk generated from large studies of agricultural workers from multiple regions.

Strengths and Limitations

This study contributes to the current state of the science surrounding HRI in agricultural workers though a large agricultural worker sample in the state of Florida. Our sample included workers of varying ethnicities, in multiples locations and performing different types of agricultural work. The protocol built upon lessons learned from our previous studies, to successfully examine associations between regional weather data, core temperature, heart rate-based physical activity measures, and clinical measurements of participants’ anthropometrics and dehydration.

While our study design provided the opportunity for multiple monitored workdays for participants, a limitation of our study was that some data were considered unusable due to equipment failure resulting in more than 20% of the data points missing. With multiple workdays monitored, we were still able to collect biomonitioring data for nearly all participants. With the study limited to one state and based on a convenience sample, we can only ascertain the presence and patterns of dangerous core body temperature experienced by Florida agricultural workers.

With the use of regional FAWN readings it is possible that the actual HI and WBGT of the microenvironment at the worksite may have been different, higher or even lower in some environments, than those calculated from the FAWN, but meteorological data readings have been found to be reliable for indicators of workplace climate.38 Finally, we were unable to collect medication history and so we could not ascertain the use of medications for diabetes mellitus type II or high blood pressure; limiting our ability to fully assess the presence of these comorbid conditions versus the treatment for these conditions as risk factors for core temperature rise which would be a meaningful addition to future studies.


Policies to support basic heat-preventative actions are needed to curtail the persistent public health threat of physically active work in the heat. NIOSH and OSHA recommend that employers institute comprehensive HRI prevention programs that include cooling interventions, risk factor assessment and health monitoring of employees, and HRI first aid. However, these programs are not mandatory, giving little incentive for employers to implement such programs. Moreover, those employers that do or consider program implementation may perceive financial disadvantage from the investment in heat interventions. Research that incorporates productivity measures into studies to verify the economic impact of worker safety policies and interventions may increase perceived support from employers. Mandated heat protections with specific, comprehensive HRI prevention programs for workers are needed quickly to avoid preventable death and health consequences in this underserved, yet crucial worker group.


A special thank you to the Farmworker Association of Florida and its members for making this study possible.


1. Gubernot DM, Anderson GB, Hunting KL. The epidemiology of occupational heat exposure in the United States: a review of the literature and assessment of research needs in a changing climate. Int J Biometeorol 2014; 58:1779–1788.
2. Centers for Disease Control and Prevention (CDC). Heat-related deaths among crop workers--United States, 1992–2006. MMWR Morb Mortal Wkly Rep 2008; 57:649–653.
3. National Institute for Occupational Safety and Health (NIOSH). NIOSH Criteria for a Recommended Standard: Occupational Exposure to Heat and Hot Environments. Cincinnati, OH: US Department of Health and Human Services, Center for Disease Control and Prevention, National Institute for Occupational Safety and Health; 2016.
4. Occupational Safety and Health Administration. OSHA Technical Manual (OTM) Section III–Chapter 4: Heat Stress (OSHA). OSHA Technical Manual: Heat Stress; 2008. Available at: https://www.osha.gov/dts/osta/otm/otm_iii/otm_iii_4.html. Accessed February 1, 2020.
5. American Congress of Government and Industrial Hygienists. Threshold Limit Values for Chemical Substances and Physical Agents and Biological Exposure Indicies. Cincinnati, OH: American Congress of Governmental and Industrial Hygienists; 2014.
6. Florida Department of Agriculture and Consumer Services. Florida Agriculture by the Numbers 2017. Maitland, FL: USDA-NASS Florida Field Office; 2018. Available at: https://www.nass.usda.gov/Statistics_by_State/Florida/Publications/Annual_Statistical_Bulletin/FL_Agriculture_Book/2017/Florida_Agriculture_by_the_Numbers2017.pdf. Accessed February 1, 2020.
7. Florida Department of Health. Health Effects of Summer Heat in Florida. Tallahassee, FL: Florida Department of Health; 2015. Available at: http://www.floridahealth.gov/environmental-health/climate-and-health/_documents/heat-profile.pdf. Accessed February 1, 2020.
8. Mac VV, Tovar-Aguilar JA, Elon L, Hertzberg V, Economos E, McCauley LA. Elevated core temperature in Florida Fernery workers: results of a pilot study. Workplace Health Saf 2019; 67:470–480.
9. Fleischer NL, Tiesman HM, Sumitani J, et al. Public health impact of heat-related illness among migrant farmworkers. Am J Prev Med 2013; 44:199–206.
10. Mac VV, Tovar-Aguilar JA, Flocks J, Economos E, Hertzberg VS, McCauley LA. Heat exposure in Central Florida Fernery workers: results of a feasibility study. J Agromedicine 2017; 22:89–99.
11. Florida Automated Weather Network. Archived weather data; 2017. Available at: https://fawn.ifas.ufl.edu/. Accessed August 1, 2018.
12. Bernard TE, Barrow CA. Empirical approach to outdoor WBGT from meteorological data and performance of two different instrument designs. Ind Health 2013; 51:79–85.
13. Rothfusz LP. The Heat Index “Equation” (or, More Than You Ever Wanted to Know About Heat Index). NWS Southern Region Headquarters: National Weather Service Scientific Services Division; 1990. Contract No.: SR 90-23.
14. World Health Organization. Body mass index—BMI. Available at: http://www.euro.who.int/en/health-topics/disease-prevention/nutrition/a-healthy-lifestyle/body-mass-index-bmi. Accessed August 7, 2017.
15. Sasaki JE, John D, Freedson PS. Validation and comparison of ActiGraph activity monitors. J Sci Med Sport 2011; 14:411–416.
16. Aguilar-Farias N, Brown WJ, Peeters GM. ActiGraph GT3X+ cut-points for identifying sedentary behaviour in older adults in free-living environments. J Sci Med Sport 2014; 17:293–299.
17. Mix JM, Elon L, Thein Mac VV, et al. Physical activity and work activities in Florida agricultural workers. Am J Ind Med 2019; 62:1058–1067.
18. Hertzberg V, Mac V, Elon L, et al. Novel analytic methods needed for real-time continuous core body temperature data. West J Nurs Res 2017; 39:95–111.
19. MacKinnon DP, Krull JL, Lockwood CM. Equivalence of the mediation, confounding and suppression effect. Prev Sci 2000; 1:173–181.
20. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986; 51:1173–1182.
21. Geraci M. Linear quantile mixed models: the lqmm package for Laplace Quantile regression. J Stat Softw 2014; 57:1–29.
22. Geraci M, Bottai M. Linear quantile mixed models. Stat Comput 2014; 24:461–479.
23. RCoreTeam. R: A Language And Environment For Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2016.
24. Asuncion Valdivia Heat Illness and Fatality Prevention Act of 2019; 2019.
25. Moyce S, Mitchell D, Armitage T, Tancredi D, Joseph J, Schenker M. Heat strain, volume depletion and kidney function in California agricultural workers. Occup Environ Med 2017; 74:402–409.
26. Vega-Arroyo AJ, Mitchell DC, Castro JR, et al. Impacts of weather, work rate, hydration, and clothing in heat-related illness in California farmworkers. Am J Ind Med 2019; 62:1038–1046.
27. Yeoman K, DuBose W, Bauerle T, Victoroff T, Finley S, Poplin G. Patterns of heat strain among a sample of US underground miners. J Occup Environ Med 2019; 61:212–218.
28. Mix J, Elon L, Vi Thien Mac V, et al. Hydration status, kidney function, and kidney injury in Florida agricultural workers. J Occup Environ Med 2018; 60:e253–e260.
29. Mauldin A. OSHA fines company after farm worker dies. The Moultrie Observer 2019.
30. Nelson DA, Deuster PA, O’Connor FG, Kurina LM. Timing and predictors of mild and severe heat illness among new military enlistees. Med Sci Sports Exerc 2018; 50:1603–1612.
31. Kenny GP, Webb P, Ducharme MB, Reardon FD, Jay O. Calorimetric measurement of postexercise net heat loss and residual body heat storage. Med Sci Sports Exerc 2008; 40:1629–1636.
32. Chung NK, Pin CH. Obesity and the occurrence of heat disorders. Mil Med 1996; 161:739–742.
33. Armstrong LE, Anderson JM, Casa DJ, Johnson EC. Exertional heat stroke and the intestinal microbiome. Scand J Med Sci Sports 2012; 22:581–582.
34. Kenny GP, Yardley J, Brown C, Sigal RJ, Jay O. Heat stress in older individuals and patients with common chronic diseases. CMAJ 2010; 182:1053–1060.
35. Becker JA, Stewart LK. Heat-related illness. Am Fam Physician 2011; 83:1325–1330.
36. Heat Illness Prevention (HB 513), Florida House of Representatives; 2020.
37. U.S. Department of Labor, OSHA. Protective Measures to Take at Each Risk Level. Available at: https://www.osha.gov/SLTC/heatillness/heat_index/protective_measures.html. Accessed February 1, 2020.
38. Mac VV, Hertzberg V, McCauley LA. Examining agricultural workplace micro and macroclimate data using decision tree analysis to determine heat illness risk. J Occup Environ Med 2019; 61:107–114.

agriculture; core body temperature; heat illness; heat index; heat-adaptive interventions; outdoor workers; physical activity

Supplemental Digital Content

Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American College of Occupational and Environmental Medicine.