Development and validation of an Outdoor Environmental Heat Index (OETI) based on geographical, environmental and meteorological data Authors: Habibollah Dehghan, Saeid Yazdanirad, Zahra Ordudari Published: 2026 ABSTRACT AbstractAn increasing number of workers are facing various physical factors, such as heat stress due to the inevitable climate change. This study aimed to develop and validate an Outdoor Environmental Heat Index(OETI) without using measurement instruments, relying only on meteorological and geographical data and using structural equation modeling. This cross-sectional study was conducted on 201 men who worked outdoors in both hot and dry, as well as hot and humid environments. Alongside collecting personal information from the participants, tympanic temperature and heart rate were measured both at rest and 90 min after the start of work. Concurrently, the climatic conditions of the workplace were assessed using instruments, along with geographic information and data from meteorological databases. Additionally, the metabolic rate and insulation value of the individuals’ clothing were estimated, and information related to other observed parameters was also recorded. The results revealed that the main factors, with a total coefficient of 0.8 had a significant direct effect on tympanic temperature. According to the results, the indirect effect coefficients of the most important variables were as follows: globe temperature (0.59), dry bulb temperature (0.55), relative humidity (0.53), work place (0.59) and heat source (0.55). The overall index score was divided into four categories based on established optimal cut-off points of 18.5, 20.02, and 21.46. Furthermore, the results showed that the OETI developed using device data and with meteorological data was able to justify 69% and 59% of the changes in tympanic temperature. OETI was developed based on a combination of observational, geographic, and meteorological data, making it a practical tool for assessing heat stress without need to measurement ins As a result, this index justifies 69% of the tympanic temperature (R2 =0.69).Keywords: Meteorological data, Observational data, Heat stress index, Environmental factors, Outdoor workplace, AMOS modelSubject terms: Health care, Health occupations, Risk factors Introduction An increasing number of workers are facing various physical factors, such as heat stress due to the inevitable climate change1, the advancements in industry, technological changes, and the modernization of human life2. People’s responses to heat stress vary, with those most susceptible being significantly impacted3. Working outdoors in hot weather presents significant risks of heat stress and increased physiological strain for those who operate in such environments without proper precautions, primarily due to intense physical activities and prolonged exposure to sunlight3–5. This risk is particularly relevant for individuals employed in agriculture, construction, brick kilns, salt pans, and other labor-intensive outdoor occupations. Such exposure to extreme heat markedly heightens the likelihood of heat-related illnesses (HRIs) and can lead to decreased productivity6. Furthermore, these workers face a unique set of challenges in their daily routines, often undertaking physically demanding tasks in uncontrolled weather conditions. They are disproportionately impacted by climate change and heat waves6. As climate change advances, average and extreme temperatures are on the rise, and heat waves are becoming increasingly frequent, placing outdoor workers at an even greater risk of HRIs7. Prolonged exposure to heat stress can result in occupational diseases and harm to vital organs, as well as diminished productivity, decreased physical and mental performance, an increased incidence of accidents, and lowered safety levels in the workplace8–10. Heat strain is generally accepted to be predicted by six heat stress factors11. Additionally, internal metabolic heat production in the body (related to physical work intensity) and clothing (such as protective clothing) significantly affect the exchange and balance of body heat with the external environment4. Two key clothing properties, thermal insulation and evaporative resistance, are important factors that influence human body heat exchange in hot environments12. The various indices have been developed to evaluate the six mentioned factors. In general, these indices are categorized into three major groups: rational indices, which are based on calculations related to the heat balance equation; empirical indices, which are derived from both objective and subjective assessments; and direct indices, which are based on direct measurements of environmental variables13. The most commonly used heat stress index is the Wet Bulb Globe Temperature (WBGT), which integrates air temperature, humidity, radiant heat, and wind speed into a single metric14. The index can be easily utilized by laypersons to manage heat stress in an industrial environment due to its straightforward measurement process. Despite its limitations such as underestimating the effect of wind speed and its inability to account for two other significant heat stress factors, metabolic rate and clothing effect WBGT remains a valid thermal index for managing occupational heat stress under appropriate conditions, also, this index necessitates the use of measuring devices and expertise in measurement15. To overcome the limitations of environmental index such as the WBGT, researchers developed indices based on meteorological data. However, each of these meteorological-based index also comes with its own set of limitations, which can affect their accuracy and applicability in various contexts. Among these indices are the Predicted Heat Stress (PHS) and the Universal Thermal Climate Index (UTCI), both of which are prominently discussed in the literature alongside the WBGT8. Among the limitations of the UTCI index is that it solely considers changes in environmental parameters. Therefore, it can be said that it is primarily applicable for assessing environmental conditions without incorporating individual factors such as metabolic rate or worker effort and clothing effects8,16. Although the PHS (Predicted Heat Strain) index takes into account all factors involved in heat transfer between the body and the environment but the insulation of clothing is limited to a maximum of 1.0 clo and In addition, it also exhibits computational complexity17. Other commonly used heat indices include the Humidex18, the Heat Stress Index19, and the heat Index20, which are only require temperature and humidity data, neglecting the effects of solar radiation and wind speed, both of which play a significant role in the heat stress experienced by outdoor workers21. By reviewing the literature, we have determined that existing heat stress indices each come with their own set of advantages and disadvantages. This highlights the necessity for an index that fulfills essential criteria, including simplicity, accessibility, ease of measurement, accuracy, validity, reliability, repeatability, and comprehensiveness. Given the intensification of global warming and the ongoing need for outdoor work and the emergence of heat-related illnesses, there is a strong demand for an index designed to estimate heat stress in outdoor environments. Consequently, this study aimed to develop and validate the Outdoor Heat Index (OETI) without the use of measurement instruments or measurement experts, relying solely on meteorological and geographical data, empirical relationships, and employing structural equation modeling (SEM) to create a comprehensive, valid, and user-friendly tool. Methods Participants This cross-sectional study was conducted in the cities of Assaluyeh and Isfahan, representing hot and humid and hot and dry environments, respectively. The study involved 201 male workers, with 101 employed in a petrochemical company and the remaining 100 from a refinery holding company. This study complies with the Declaration of Helsinki in terms of ethics approval and consent for participation. Considering the objective of this study, which focuses on a new index, the lowest expected correlation between the novel index and body temperature was assumed to be 0.2. Based on this assumption, and with a confidence level of 95% and a statistical power of 80%, the required sample size was calculated as follows22: In this equation, the values of , , and W were 1.96 based on a confidence level of 95%, 0.84 based on a test power of 80%, and 0.203 based on a minimum correlation coefficient of 0.2, respectively22. For selecting research place also, two locations were selected to provide a variety of climatic conditions for the development of the index. One location was a factory in central Iran with a hot and dry climate, and the other location was a factory in southern Iran with a hot and humid climate. Within each industry, a variety of climatic conditions from moderate to hot temperatures and from low to high humidity were selected for the study. Due to the nature of their work, employees in Assaluyeh were required to labor under intense sunlight, with maximum temperatures reaching 44 °C and maximum humidity levels at 63%. In contrast, the working conditions in Isfahan were somewhat similar, but the maximum temperature was 42 °C with maximum humidity 43%. Participants were selected to cover a wide range of parameters involved in the study. Various departments of both industries were visited to ensure a comprehensive assessment. Workstations were selected based on the tasks assigned to each station, allowing for a broad representation of personal and environmental parameters in the study. Initially, 600 employees were invited to participate, but ultimately, 201 individuals were selected as participants according to the established inclusion and exclusion criteria. Conditions that could influence heat stress as intervening variables, based on scientific literature, were established as inclusion criteria for the study. These criteria required participants to have more than one year of experience working in a hot environment (to ensure heat adaptation), be between the ages of 18 and 70, and not have diabetes, hyperthyroidism, or any cardiovascular, pulmonary, psychiatric, infectious, renal, gastrointestinal, or musculoskeletal diseases. Additionally, participants were instructed to avoid medications that could affect heart rate and blood pressure, and favorable conditions, such as the absence of excessive ear canal mass, were vital for individual selection. They were also advised to refrain from consuming coffee, caffeine, or alcohol for 12 h prior to the study23–25. Exclusion criteria included a tympanic temperature above 39 °C, noncooperation during the measurement of physiological parameters, and a heart rate exceeding the value calculated using the Eq. 225,26. Where: HRmax: maximum heart rate. Table 1 displays the statistical distribution of demographic original data. Table 1Statistical distribution of demographics data.Demographics factorsAge(year)206136.09 ± 8.13Height(m)160198175.58 ± 6.9Weight(kg)5511679.11 ± 12.29Work experience(year)13011.62 ± 7.01Physical activity (hours per week)0202.39 ± 3.37 Statistical distribution of demographics data. Data collection This study has been approved and registered with the ethics code(IR.MUI.DHMT.REC.1402.028) by Isfahan University of Medical Sciences and all methods were performed in accordance with the relevant guidelines and regulations. Participants rested in a cool room for 30 min prior to the experiment. During this rest period, they completed an informed consent form for participation in the study. Subsequently, their demographic information, including height and weight, was measured using a portable stadiometer (Seca206,Seca, Basel, Switzerland) with a precision of1cm and a calibrated electronic weighing scale(Seca 899,Seca) with a precision of 0.1 kg, respectively27. Additionally, their work history and daily physical activity levels were recorded through medical records. Then, heart rate and tympanic temperature were measured according to ISO 9886 standard at 20,25, and 30 min and the average values ​​were recorded. Then the worker returned to his workplace and began performing his work duties. Physiological parameters were measured 90 min after the start of work, in accordance with the ISO9886 standard. At the same time, environmental climatic parameters, including dry bulb temperature(Ta), wet bulb temperature(Tw), ground temperature (Tg), and wind speed (Ws), were measured according to the ISO 7243 and ISO 7726 standards using the calibrated WBGT meter (TES1369B WBGT meter). The WBGT device was positioned at the designated location, and after allowing a minimum of 20 min for environmental adaptation, the required measurements were recorded28. On the other hand, the aforementioned parameters from the Microsoft Network)MSN(weather forecast site and the geographic information of the region were also noted to calculate the radiant temperature29. Additionally, to determine the thermal insulation of clothing, the ISO9920standard was utilized30, while the average metabolism was assessed according to the ISO8996standard. The metabolic rate was further corrected using the BS 7243 standard31. Tympanic temperature was used as the gold standard, measured with a tympanic thermometer(Braun Thermoscan3,Braun, South Boston, MA, USA), which has an accuracy of ± 0.2°C32. The heart rate was also measured using a heart rate monitor(Beurer, PM70,Germany) with an accuracy of ± 1 beat per minute)33. Environmental parameters were assessed using a TES1369B WBGT meter, which has an accuracy of0.1°C34. wend speed was also measured using TES-1340 (TES Electrical Electronic Corp, Taipei, Taiwan) with a precision of ± 3%35. In this study, since factors such as exposure time, workplace, and heat source can influence an individual’s heat stress in an outdoor environment, this information was documented by observing the working conditions and recording the times of work. In this study, the workplace factor was defined as the average duration of time that individuals were exposed to sunlight that was quantified by the 5-point scale 0(Under an hour),1(1–2 h),2(2–4 h),3(4–6 h),4(6–8 h),5(more than 8 h). The heat source parameter also pertains to the amount of heat received from sources other than the sun that was quantified by the 5-point scale (0: Only sunlight, 1: very low,2: low,3: medium,4: high,5: very high). Exposure time also refers to the time which environmental and individual factors are measured sunlight that was quantified by the 5-point scale 0(8–10 AM),1(10–11 AM),2(11–12 AM),3(16–17 PM),4(14–16 PM),5(12–14 PM). To gather the necessary objective and subjective information, a form was designed to record the data provided by the participants (Table 2). Table 2A form designed to collect relevant information from people involved in the study.VariableDemographics factorsAge(year)Height(m)Weight(kg)Work experience(year)Physical activity (hours per week)Meteorological parametersDry temperature (°C)Wind speed (km/h)Relative humidity(%)Parameters measure by devicesDry temperature(°C)Wet temperature(°C)Wind speed(km/h)Total metabolism (watts)Clothing thermal resistance (clo)Resting tympanic temperature(°C)Resting heart rate (beat/min)Working tympanic temperature(°C)In 20 minIn 25 minIn 30 minWorking heart rate (beat/min)Average duration of time that individuals were exposed to sunlight0(Under an hour)1(1–2 h)2(2–4 h)3(4–6 h)4(6–8 h)5(more than 8 h)The amount of heat received from sources other than the sun(0: Only sunlight1: very low2: low3:medium4: high5: very high)The time which environmental and individual factors are measured sunlight0(8–10 AM)1(10–11 AM)2(11–12 AM)3(16–17 PM)4(14–16 PM)5(12–14 PM) A form designed to collect relevant information from people involved in the study. In this study, to estimate solar radiant energy, geographic and meteorological information was utilized in conjunction with the following empirical formula developed by Hargreaves and Samani, as well as Allen which is applicable to both hot and dry, as well as hot and humid regions36,37, Where: : Extraterrestrial solar radiation (MJ/m2 day) : Measured global solar irradiation (MJ/m2 day) a: 0.16 for interior regions, 0.17 for arid and semiarid climates and 0.19 for coastal regions. : Maximum Temperature (°C). : Minimum temperature (°C). solar constant (0.082)38. n: is the Number of the day in the year between 1(1 January) and 365 or 366 (31 December) : Latitude of the site (°). : Sun declination angle (°). : Sunrise hour angle of the day (°). After estimating solar radiation using the formula mentioned in the previous section and recording the temperature and airflow speed from weather forecast sites, we calculated the globe temperature using the following formula. Where: : globe temperature (°C). : air temperature (°C). : solar radiation (). : wind speed) Simultaneously, the globe temperature was measured using the WBGT meter. Subsequently, the correlation between the globe temperature measured with the WBGT meter and the globe temperature calculated using the following formula was assessed using the Pearson correlation coefficient. Index development The model was designed using AMOS software, and the indirect effect values of the 9 parameters on the change in tympanic temperature were utilized as variable coefficients in the new index. Before determining the final formula, each of the variables was normalized between 0 and 1 based on the values measured in the present study. The calculation of the new index is based on Eq. 8. In the above formula, C1 to C9 represent the effect coefficients of the basic variables. Here, denotes the normalized dry bulb temperature, represents the normalized wet bulb temperature, indicates the normalized globe temperature, signifies the normalized wind speed, refers to the normalized total metabolism, and corresponds to the normalized thermal insulation of the clothing. indicates the normalized exposure time of work, signifies the normalized work place, refers to the normalized heat source. Statistical analyses The collected data were entered into the Statistical Package for Social Sciences (SPSS) version26. Skewness and kurtosis analyses indicated that the variables had a normal distribution. Consequently, Pearson’s correlation test was employed to examine the relationships between the variables. An empirical model was depicted based on the relationships between variables using AMOS software. To evaluate the model fit, absolute, comparative, and normative fit indices were employed. In this model, the factor loading values of critical variables were calculated with respect to thermal strain as a latent variable and tympanic temperature as the gold standard variable. Based on the results, a new index was developed. Subsequently, the total index score was classified into four risk levels, ranging from low to very high, using receiver operating characteristic (ROC) analysis. The boundaries between the risk levels in the ROC analysis were determined at tympanic temperatures of 37.5 °C,38.0 °C, and 38.5°C39. The closest point to the ideal state in each of the ROC curves was used to determine the optimal cut-off points for the index. The validity of the index was assessed through linear regression analysis between the OETI and tympanic temperature. Results Table 3 displays the statistical distribution of original data such as main, and physiological factors. The distribution of original demographic data is also shown in the Methods section. As indicated in Table 3, the collected data encompass a wide range of values necessary for the model, which can result in tympanic temperature variations spanning from 36.7 to 39.1 °C. Table 3Statistical distribution of demographics, fundamental, and physiological factors.VariableMinimumMaximumMean ± SDMain factorsMeteorological parameters Dry temperature (°C) 334437.01 ± 2.75Wind speed (km/h) 274.43 ± 1.16Relative humidity (%) 106329.03 ± 12.43Solar radiation(°C) 31.1740.9635.72 ± 2.56Parameters measure by devices Dry temperature (°C) 32.144.836.94 ± 2.88Wet temperature (°C) 18.629.822.29 ± 2.96Wind speed (km/h) 184.61 ± 1.4Relative humidity (%) 116530.25 ± 12.47Globe temperature (°C) 32.5943.1037.52 ± 2.72Total metabolism (watts)148416296.48 ± 93.74Clothing thermal resistance (clo)0.51.30.78 ± 0.27Physiological factorsResting tympanic temperature (°C) 36.237.836.87 ± 0.13Working tympanic temperature (°C)36.538.937.84 ± 0.77Resting heart rate (beat/min) 589072.23 ± 8.52Working heart rate (beat/min)6312092.57 ± 11.81 Statistical distribution of demographics, fundamental, and physiological factors. Meteorological parameters Dry temperature (°C) Wind speed (km/h) Relative humidity (%) Solar radiation(°C) Parameters measure by devices Dry temperature (°C) Wet temperature (°C) Wind speed (km/h) Relative humidity (%) Globe temperature (°C) Resting tympanic temperature (°C) Resting heart rate (beat/min) Table 4 also presents the Pearson correlation between the variables. A significant relationship was observed between tympanic temperature and all the basic factors. Details of the correlation coefficients for the other variables can be found in Table 4. Table 4Pearson correlation among the study variables.Variable123456789101112131415TdM-TdD0.745**-Tw0.708**0.673**-WsM0.263**0.267**0.198**-WsD0.154*0.210**0.0240.723**-RHM0.539**0.407**0.591**0.0920.121-RHD0.541**0.416**0.592**0.1190.140*0.995**-Met0.406**0.495**0.503**0.174*−0.0160.309**0.316**-Clo0.287**0.223**0.1000.221**0.274**0.0710.0970.151*-Tg0.870**0.770**0.678**0.292**0.223**0.442**0.4550.484**0.403**-SR0.685**0.676**0.529**0.172*0.1230.570**0.574**0.517**0.203**0.718**-ET0.304**0.393**0.334**0.219**0.0460.263**0.257**0.483**0.1090.368**0.346**-WP0.485**0.569**0.412**0.0570.0300.452**0.461**0.398**0.202**0.562**0.525**0.401**-HS0.354**0.444**0.343**0.164*0.234**0.444**0.463**0.258**0.195**0.465**0.462**0.282**0.586**-Tt0.488**0.494**0.426**0.144*0.171*0.606**0.614*0.307**0.294**0.600**0.463**0.265**0.577**0.617**-(1) TdM: Dry Temperature(Metrological), (2) TdD: Dry Temperature(Device), (3) Tw: Wet Temperature, (4) WsM: Wind Speed (Metrological), (5) WsD: Wind Speed (Device), (6) RHM: Relative Humidity (Metrological), (7) RHD: Relative Humidity (Device), (8) Met Total Metabolism, (9) Clo: Clothing Thermal Insulation, (10) Tg: Globe Temperature, (11) SR: Solar Radiation, (12) ET: Exposure Time, (13) WP: Work Place, (14) HS: Heat Source, (15) Tt: Tympanic Temperature.** P < 0.01** P < 0.01 Pearson correlation among the study variables. (1) TdM: Dry Temperature(Metrological), (2) TdD: Dry Temperature(Device), (3) Tw: Wet Temperature, (4) WsM: Wind Speed (Metrological), (5) WsD: Wind Speed (Device), (6) RHM: Relative Humidity (Metrological), (7) RHD: Relative Humidity (Device), (8) Met Total Metabolism, (9) Clo: Clothing Thermal Insulation, (10) Tg: Globe Temperature, (11) SR: Solar Radiation, (12) ET: Exposure Time, (13) WP: Work Place, (14) HS: Heat Source, (15) Tt: Tympanic Temperature. ** P < 0.01 ** P < 0.01 Figure 1 illustrates the theoretical model analyzed using SEM. The results revealed that the main factors, with a total coefficient of 0.8 had a significant direct effect on tympanic temperature. Table 5 reports the effect coefficients of the variables in producing thermal strain. Among main factors, Tg, Td and Relative humidity had the highest indirect effects on tympanic temperature, with coefficients of 0.599, 0.555 and 0.530, respectively. Table 6 presents the appropriate fit indices of the theoretical model used in the present study. The results indicate that all the values obtained from the fit indices fall within the optimal range, confirming the adequacy of the model fit. The theoretical model analyzed by SEM. Table 5Effect coefficients of the variables in producing thermal strain.VariableDirect effectIndirect effect P value Dry temperature0.6970.555P < 0.001Relative humidity0.6660.530P < 0.001Globe temperature0.7530.599P < 0.001Wind speed0.2400.191P = 0.006Total metabolism0.4950.394P < 0.001Clothing thermal insulation0.2780.221P < 0.001Exposure Time0.4300.342P = 0.001Work Place0.7440.592P < 0.001Heat Source0.6950.553P < 0.001Thermal strain0.796-P < 0.001 Effect coefficients of the variables in producing thermal strain. Table 6Goodness-of-fit indices of the analyzed model.IndexNameThreshold of FitnessObtained valueAbsolute fitness indicesGoodness-of-fit index(GFI)> 0.90.959Adjusted goodness-of-fit index(AGFI)> 0.90.950Comparative fitness indicesNormed fit index(NFI)> 0.90.925Comparative fit index(CFI)> 0.90.956Incremental fit index(IFI)0–10.957Normed fit indexRoot mean squared error of approximation(RMSEA)< 0.10.079Normed Chi-square(X2/df)1–32.240 Goodness-of-fit indices of the analyzed model. The OETI index was created using normalized variables and indirect effect coefficients, represented by the following formula: Where: Td: Dry Temperature (°C), RH: Relative Humidity (%), Tg: Globe Temperature (°C), WS: Wind Speed (Km/h), Met: Total Metabolism (Watt), Clo: Clothing Thermal Insulation (clo), ET: Exposure Time, WP: Work Place, HS: Heat Source. In this equation, it is important to note that the signs of the coefficients for wind speed change from positive to negative when dry temperature is below 35 °C(normal skin temperature), as heat strain decreases under these conditions. Figure 2; Table 7 displays ROC curves related to different risk categories and various risk zones of OETI index, respectively. In this figure, the points with highest sensitivity and specificity were considered as best cut-off points of OETI index and their values were extracted from coordinate tables related to the ROC curves. The results indicated that the optimal cut-off points between low and moderate risk, moderate and high risk, and high and very high risk were equal to 18.50 (sensitivity = 0.889 and specificity = 0.880), 20.02 (sensitivity = 0.939 and specificity = 0.893), and 21.46 (sensitivity = 0.904 and specificity = 0.685), respectively. The area under of ROC curves (AUC) was also calculated as 0.954 (95% CI: 0.929, 0.980) (p < 0.001), 0.952 (95%CI: 0.921, 0.983) (p < 0.001), and 0.799 (95%CI: 0.741,0.858) (p < 0.001), respectively. Figure 3 also reports linear regression curve between tympanic temperature and OETI index. As a result, this index justifies 69% of the tympanic temperature (R2 =0.69). Receiver operating characteristic (ROC) curves of a low and moderate, b moderate and high, and c high and very high risk zones. Table 7The risk levels of OETI index.Risk levelEquivalent values of OETI indexLowLess than 18.50Moderate18.50 to 20.02High20.02 to 21.46Very highMore than 21.46 The risk levels of OETI index. Linear regression curve between core temperature and a OETI computed by device data and b OETI calculated by metrological data. Similarly, Fig. 3 includes a linear regression curve depicting this relationship based on meteorological data, where the OETI index explains 59% of the variation in tympanic temperature (R²=0.59). Discussion In this study, a new index was developed based on meteorological information, eliminating the need for specialized tools and expertise. The results indicate that this index demonstrates the necessary validity, and access to the information required for its estimation is readily available. Since the results indicate that the thermal strain estimated using this index has a direct effect with a factor loading of 0.79 on tympanic temperature, it can be concluded that tympanic temperature can be reliably predicted by the OETI. In this study, we will first examine the indirect effect coefficients of the variables, followed by an assessment of the validity and fitness of the index. We will then evaluate its effectiveness in assessing heat stress risk by analyzing the index’s ability to account for changes in ear temperature, and we will compare these findings with those from other studies. According to the results, the indirect effect coefficients of the variables were as follows: globe temperature (0.59), dry bulb temperature (0.55), relative humidity (0.53), wind speed (0.19), metabolism (0.39), clothing thermal insulation (0.22), exposure time (0.34), work place (0.59) and heat source (0.55). In the study by Yazdanirad et al., The indirect effect coefficients of the variables were as follows: Tg (0.77), Td (0.75), wind speed (0.24), metabolism (0.49), and clothing thermal insulation (0.39)40,41. In this study, similar to the findings of Yazdanirad41 and suman42, the highest coefficient was associated with Tg. Conversely, in the studies conducted by Moran43 Golbabaei44, Dehghan24 and Epstein45, dry temperature exhibited the highest correlation coefficient among the measured variables. In the WBGT index46, which is recognized as a valid heat stress index, unlike this study and the studies conducted by Epstein45, suman42, Dehghan24, Golbabaei44 and Yazdanirad41 the wet bulb temperature had the highest weight in the developed index. The reason the wet bulb temperature coefficient is higher than those for globe and dry bulb temperature is that the range of changes in wet bulb temperature is smaller than that of globe and dry bulb temperature. If the ranges of changes in these temperatures were equal and their values were normalized, the coefficients for dry bulb and ground temperatures would increase and the results were consistent with those of this study as well as findings from other studies. On the other hand, elevated relative humidity can contribute to heat stress primarily when dry air temperatures and globe temperatures are high, as it hinders the evaporation of sweat from the skin surface. This reduction in evaporation impairs the body’s ability to cool itself, increasing the risk of heat stress47. Additionally, in this study, as observed in the research by Golbabaei44 and Dehghan24, relative humidity ranked next in terms of correlation. Also, in this study, in accordance with the observations of Yazdanirad41 and Golbabaei44 and contrary to the results of Dehghan et al.24, wind speed was considered to have the least effect among the main variables. Since wind speed is not a constant value and its effect is influenced by both dry air temperature and skin temperature, it is logical that its impact on thermal strain is less pronounced compared to the other parameters48. In this study, similar to the findings of Yazdanirad41, metabolism had a higher coefficient than thermal insulation of clothing, although the coefficients varied between studies and the coefficients obtained from the study of Dehghan24 were very different from this study. However, in general, the coefficients of influence for these two variables are lower than those of the environmental variables. Since the effect of these two parameters on producing thermal strain is dependent on the environmental conditions, the lower coefficients appear to be justified48. In general, the coefficients obtained in this index differ from those of other indices. These discrepancies may be attributed to variations in the types of studies conducted. For instance, the study by Dehghan et al.24 was observational-perceptual in nature, relying on data collected through instrumental measurements. In contrast, this study was categorized as environmental, with data collection based on meteorological data. These methodological differences can significantly influence the resulting coefficients and their interpretations. The calculated scores of the developed OETI index were classified into four risk levels using ROC analysis. According to previous studies, if the area under the curve (AUC) values are greater than 0.90, the diagnostic accuracy of the curves is considered excellent49. In this study, the results indicated that the developed index demonstrated good and acceptable validity and fitness, supporting its effectiveness in assessing heat stress risk. Furthermore, the results showed that the OETI index developed using device data and with meteorological data was able to justify 69% and 59% of the changes in tympanic temperature. For instance, Yazdanirad et al. found that the EHSRA index explained 75% of the variation in tympanic temperature41. Also, In the study by Monazzam et al., the R² values between the WBGT and PHS index with tympanic temperature were found to be 71% and 76%, respectively50. Similarly, Dehghan et al. reported a correlation of 0.73 between HSSI scores and tympanic temperature24. According to the results, the accuracy of the index developed in this study was lower than that of indices developed using instruments. This discrepancy in accuracy might stem from the fact that instruments generally capture data with greater precision than weather stations. Further, weather stations are typically positioned away from any local heat sources or obstructions, and air velocity measurements are commonly taken at a standard elevation of 10 meters4. Also, which may be attributed to the measurements being taken under diverse climatic conditions and the differences in study design. The index developed by Epstein and Moran, which was based on meteorological data, demonstrated a strong correlation coefficient with the WBGT index(R²=0.95). However, this index had some limitations, as it only considered dry and wet bulb temperatures without accounting for radiant temperature. Additionally, the clothing factor was restricted to light summer clothing, and since wet bulb temperature data were not directly available from meteorological stations, a complex formula was employed to estimate it. Furthermore, this index was designed for a temperature range of 15–33°C45. The Environmental Stress Index (ESI) was one of the indices developed based on meteorological data, incorporating only the parameters of air temperature, humidity, and solar radiation, with solar radiation measured using a radiometer. The relationship between the ESI and WBGT demonstrated very strong correlations, with R² values exceeding 0.98143. The Universal Thermal Climate Index(UTCI) is an index that directly utilizes meteorological data to assess thermal risk in outdoor environments. UTCI takes into account factors such as radiant heat, air temperature, wind speed, and humidity. The radiant temperature is calculated based on cloud cover, global radiation, and meteorological models51. When cloud cover forecasts are inaccurate, estimates of solar radiation can vary by ± 2 °C to ± 6°C52. Furthermore, the UTCI is designed for conditions involving an average activity level of 135 W/m², an exposure duration of two hours, and clothing insulation of 0.3 clo51. This is particularly relevant since workers often utilize specialized protective equipment with insulation values typically exceeding 0.5 clo53. On the other hand, in the study conducted by Zare et al., no significant association was found between UTCI and the physiological parameters of workers2. This lack of correlation suggests that the UTCI may not fully capture the thermal stress experienced by workers. These issues have introduced limitations to the UTCI index, affecting its accuracy and applicability in certain conditions. However, the accuracy and the number of parameters involved in the index developed in this study were greater than those of other indices based on meteorological data. This index incorporates critical parameters such as heat source, working environment conditions, and duration of heat exposure. Additionally, it provides an accurate and comprehensive estimate of radiant temperature, enhancing its ability to assess heat stress effectively in various outdoor environments. Among the limitations of this index are the following: due to ethical considerations, it was not feasible to use rectal temperature as the gold standard. Additionally, since there were no female workers in the study environments, the statistical population of this study consisted exclusively of men. Therefore, it is recommended that this index be evaluated across different genders, as well as in various occupations and work environments with diverse climatic conditions. Additionally, since the primary objective of this study was to develop an index for classifying environmental indicators, we intentionally excluded subjective factors. This decision paves the way for the future development of a new subjective index that could integrate these elements. Furthermore, as a thermal stress assessment index was developed in this study, factors influencing thermal comfort such as the presence of trees54, sky55, river56, and the effects of light and acoustics57 we’re not addressed. It is recommended that future research consider incorporating these factors. Another limitation of this study is that only internal validation was conducted. It is suggested that this indicator is validated in more diverse environmental conditions and other industries. Conclusion This study involved 201 Iranian men working in the petrochemical industry across hot and dry, as well as hot and humid regions. The aim was to develop and validate a new, simple, valid, and reliable environmental heat stress assessment index for accurately evaluating outdoor heat stress without the use of measuring devices. In this study, factors such as dry temperature, wet bulb temperature, wind speed, relative humidity, and solar radiation were recorded using the device and meteorological stations. Additionally, variables including metabolism, clothing insulation, heat sources, exposure duration, and the type of workplace were also considered. The results showed that the OETI index developed using device data and with meteorological data was able to justify 69% and 59% of the changes in tympanic temperature. This index utilizes only environmental, observational, geographic, and meteorological data across different climatic regions without any device. The new index can be employed to predict the risk of heat strain, facilitate preventive measures, and ultimately reduce the adverse effects of heat exposure. It is suggested that the validity of OETI index be assessed in various climates and countries, and that the gender factor be considered in future studies to enhance the applicability and accuracy of the index across different area and populations. Acknowledgements This research has been sponsored by Isfahan University of Medical Sciences, National Petrochemical Industries Company (Zagros Petrochemical Company) and Isfahan Petro refining Holding. The authors would like to highly appreciate the organizations mentioned. Author contributions H.D: Reviewing and Editing, SupervisionS. Y: Conceptualization, Methodology, Software, Formal analysisZ. O: Resources, Investigation, Data Curation, Writing- Original draft preparation, Visualization. Funding This research has been sponsored by PhD theses of Isfahan University of Medical Sciences, Isfahan Petro refining Holding and National Petrochemical Industries Company. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Ethical approval This study has been approved and registered with the ethics code(IR.MUI.DHMT.REC.1402.028) by Isfahan University of Medical Sciences. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.