Inequality in mortality risk of older adults between rural and urban areas exposure to temperature variations in China

Kun Hou; Liqiang Zhang; Feng Yang; Wei Hu; Xia Xu; Xin Yao; Zhen Wang

doi:10.1186/s12889-025-24088-y

. 2025 Sep 1;25:3000. doi: 10.1186/s12889-025-24088-y

Inequality in mortality risk of older adults between rural and urban areas exposure to temperature variations in China

Kun Hou ^1,^✉, Liqiang Zhang ², Feng Yang ³, Wei Hu ⁴, Xia Xu ⁵, Xin Yao ², Zhen Wang ⁶

PMCID: PMC12400752 PMID: 40890674

Abstract

Background

Non-optimal ambient temperatures have been demonstrated to negatively affect a variety of health outcomes, particularly population mortality. However, the mortality risk of older adults between rural and urban areas exposure to temperature variations remain unclear.

Methods

Here, leveraging the panel data for 27,193 older adults from the largest and most complete Chinese Longitudinal Healthy Longevity Survey during 2005–2018, we explored the impacts of temperature variations like low/high and extreme temperatures on mortality risk of older adults in China. Subgroup analyses were performed by place (urban and rural), sex, disease, education, income and health-risk behaviors (e.g., chronic drinking).

Results

The relative mortality risk of older adults in rural areas increased by 3.57% (95% CI: 0.56-6.79%) for each 1 °C ascent of high temperature above 25 °C, which was much higher than that in urban areas 2.15% (95% CI: 0.27-4.06%), whereas increased by 8.94% (95% CI: 1.03-15.21%) for each 1 °C descent of low temperature below − 15 °C, much higher than that in urban areas 6.37% (95% CI: 1.83-12.92%). The expected future additional cumulative deaths of older adults in mainland China without effective interventions could approach to 473,351 (95% CI: 282,989 − 878,693) attributable to future high temperature rise with the most extreme climate scenario of SSP5-8.5 under the International Coupled Model Comparison Program Phase 6 (CMIP6) from 2020 to 2050.

Conclusion

The effects of high and low temperatures on the mortality risk of older adults in rural areas is much greater than that in urban areas, and future expected temperature increases would lead to a significant increase in excess deaths among older adults in China. These comparable findings provide key scientific evidence for policymakers in the planning of public-tailored interventions to mitigate the risks of abnormal climate, especially in rural areas.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-24088-y.

Introduction

Climate change and anomalies are emerging as major threats to human health and well-being [1–3]. Greenhouse gas concentrations in the atmosphere keep increasing at rates incompatible with limiting warming to 2 °C by 2050 [1]. Correspondingly, the global mean surface temperature is growing at an unprecedented rate, and direct negative health impacts of climate change arise from abnormal high and low temperatures and increased extreme weather events involving heat waves and cold spells [4, 5]. Mortality attributable to climate change account for a significant proportion of the total deaths, in which temperature variations play an important role [6]. High and low ambient temperatures are related to a variety of diseases, including cardiopulmonary, respiratory and infectious diseases, as well as mortality [7–10], mediated through certain physiological mechanisms [11].

Both recent global studies [9, 12, 13] and regional analyses in North America [14, 15], Europe [16] and Western Pacific [17] about the impacts of abnormal temperatures on mortality reveal that large disparities of the temperature–mortality association persist across various regions [9]. Region-specific analyses are urgently needed to identify the vulnerable subpopulations, thereby informing climate adaptation, emergency response management and regional policies. A large proportion of regional studies of temperature and mortality have been conducted in high-income regions or countries, whereas few involved low- and middle-income areas [4]. This limits a deeper understanding of the effect of temperature variations on mortality in low- and middle-income areas and health protection efforts in climate change-sensitive regions [18]. With the most population and the largest developing country, China has been experiencing rapid climate change. Taking the impacts of temperature variations on mortality risk of Chinese populations as a case study is representative for addressing the issue.

China’s average surface temperature is rising at a rate of 0.24 °C per decade from 1951 to 2018, higher than that of 0.13 °C of the global average during the same period [19]. China has experienced more frequent heat waves, typical of which occurred in multiple cities in July, 2022 [20], and eastern urban areas have faced significant increases in average and extreme thermal stress [21]. At the same time, China is experiencing population aging on an unprecedented scale as the world’s largest older population county [22]. Dramatic changes in China’s aged population are increasing the health risks posed by abnormal high or low temperature, especially for mortality [23].

There are massive regional and urban-rural disparities in mortality risk varying with geographical environment, social structure, residential facilities and adaptability [24]. Thus, temperature variations may have different impacts on populations living in rural and urban regions. Unfortunately, studies of temperature-mortality responses in older adults have rarely focused on the urban-rural differences. In recent decades, the standard of living of rural populations has been improving, and food and economic systems have been shifted from small-scale, rainfed agriculture and livestock systems to mechanized systems [25, 26]. Compared with populations in urbans in China, those living in rural areas still have poor socioeconomic status and low demographic and housing characteristics. Moreover, older adults in rural areas have relatively limited affordability to access health care, are less likely to obtain additional insurance, and often face greater public health challenges than those in urban areas [27]. All these may exacerbate health hardship among already vulnerable populations [28]. How climate impacts mortality of rural older adults over time in China as well as the differences between older adults in rural areas and those in urbans in terms of mortality are not well explored.

Here, we explore the relationship between exposure to abnormal temperature and all-cause mortality of Chinese older adults based on the Chinese Longitudinal Healthy Longevity Survey (CLHLS) during 2005 to 2018. Our aim is to quantify the temperature-mortality responses in overall and subgroup. Subgroup analyses are performed by place (urban and rural), sex, employment, education, age, geographical location. Our results provide preliminary evidence that mortality risk of older adults is strongly impacted by temperature anomalies, especially extreme temperature for older men in rural areas.

Materials and methods

Temperature data

Temperature data are derived from the fifth generation ECMWF reanalysis for the global climate and weather (ERA5) dataset, which used field and satellite remote sensing measurements to derive a global dataset of various meteorological variables covering the complete spatio-temporal period of this study (https://www.ecmwf.int/). We utilize the grids to estimate the daily temperature data of mainland China with a resolution of 30 km, whereby the data of monthly average temperature for each district and county is obtained. Temperature exposure data for each district or county were obtained by aggregating grid-scale data to the county level and weighting them by county boundary area. When the county area was smaller than a single grid, the temperature data were derived from the grid within which it was located.

Air pollution data

Air pollution has been shown to be associated with mortality risk in previous studies [29–31]. We use the air pollutants of PM_2.5 and O₃ as the confounding factors to represent the interference effect of air pollution on the model, which are obtained from the China High Air Pollutants (CHAP) dataset [32, 33]. The spatial and temporal resolutions of this dataset are 1 km and 1 day, respectively, and the time span is 2000–2023, which fully covers the entire time span of our study.

Mortality of older adults

Our study utilized data from the Chinese Longitudinal Healthy Longevity Survey (CLHLS), a long-term, population-based cohort study targeting older adults in China. Initiated in 1998, the CLHLS is conducted every 2–3 years and encompasses roughly half of the cities and counties across 23 provinces, including provincial-level municipalities and autonomous regions. These areas collectively represent about 85% of the nation’s total population. The CLHLS survey comprises two distinct questionnaires: one for surviving respondents and another for family members of deceased elderly individuals. For surviving respondents, the survey gathers detailed information on demographics, socioeconomic factors, self-assessed health and quality of life, cognitive and physical functioning, mental health, personality traits, lifestyle, caregiving, disease management, and healthcare expenditures. The questionnaire for family members of deceased participants focuses on the circumstances, time, and place of death.

Our analysis includes data from all five waves of the CLHLS conducted between 2005 and 2018 (Wave 1: 2005; Wave 2: 2008–2009; Wave 3: 2011–2012; Wave 4: 2014; Wave 5: 2017–2018). The dataset comprises 27,193 elderly respondents from 917 county-level administrative units, with 15,815 females (58.16%) and 11,378 males (41.84%). Among these participants, 1,108 took part in all five waves (none of whom passed away), 2,746 participated in four waves (747 of whom died), 3,739 attended three waves (1,940 of whom died), 5,975 joined two waves (3,907 of whom died), and 14,793 participated in only one wave (9,274 of whom died). Ultimately, we established a panel dataset containing 27,193 respondents over a 13-year follow-up period, during which 15,868 participants passed away. The urban and rural areas were defined based on the urban establishment system and administrative divisions in China, which were used by the CLHLS data to classify the elderly into urban and rural areas using the residential addresses of the interviewed elderly.

Methods

Bayesian spatio-temporal model

We take the monthly average temperature (°C) to reflect the effect of temperature on mortality risk [7, 34]. Based on the death information from the CLHLS dataset of the interviewed elderly, we calculated the number of deaths per district or county for each month, which was used to represent the number of deaths in the corresponding district or county for the current month, and this number of deaths was logarithmic used to correlate with various variables such as temperature. The correlation of mortality risk to the variable of interest was achieved by logarithmically quantifying the number of deaths, similar to what has been done in other previous studies. The count of elderly deaths per month for each district or county is characterized through the Poisson distribution [9, 34]. We explored the relationship between temperature variations and mortality using a Bayesian spatio-temporal model that accounts for complex interaction effects of time and space, which has been shown to adequately eliminate the interference of confounding factors [34]. The logarithmically transformed count of deaths is modeled as the sum of components of a variety of dependent and relational terms involving the geographical location of death (in which county), time of death (in which month of the year), the overall time, and monthly average temperature [34], as shown in Eq. (1).

where death_{county−time} denotes the number of deaths per county in each month. temperature_{county−time} accounts for the mean monthly temperature of each district or county, whose functional relationship with the dependent variable is denoted by f. γ₀ represents the general intercept, and δ₀ accounts for the variation effect with overall time. γ_county and δ_county represent the intercept and corresponding time slope of a given county or district, interpreted as the spatially structured and unstructured random effects across various adjacent counties, and how these random effects change over time. According to the address information of each elderly person when being interviewed, whose residential area is divided into districts or counties. Districts and counties are referenced as administrative units of the same level based on China’s administrative divisions, which explains the use of district or county descriptions here. We use the Besag, York and Mollie spatial model to simulate the spatial random effects [35], whose spatial structured and unstructured effects can be fully captured and γ_month and δ_month characterized [34]. account for the intercept of a given month and slope over time, interpreted as the difference in random effects among adjacent months and the change over time. We simulate the monthly random intercept and slope using the first-order random walk priors to effectively identify the trend of monthly variation [34]. December is distributed immediately following January of the next year taking into account the cyclic structure of the random walk [36]. λ_{county−month} and φ_{county−month} account for the intercept and slope of the county-month interaction over time, indicating the variations in county-specific mortality levels and trends across various months, respectively. ϴ_time denotes the monthly complex nonlinear change with overall time, which is captured using a first-order random walk [36, 37]. ε_{county−time} is an over-dispersion term, accounting for the potentially possible disturbing factor not covered in the model, which is simulated as N(0, σ²_ε). PM_2.5 and O₃ represent the interference effects of air pollutants PM_2.5 and ozone, respectively, whose coefficients are indicated by β₁ and β₂.

We assess the relationship between temperature variations and deaths in two stages. First, we characterized the nonlinear relationship between mortality risk and temperature variations through various fitting functions f involving 3/5 nodes of natural cubic splines, polynomial, and b-spline [7]. The temperature coefficient obtained by fitting the function f is multiplied by each temperature value across the entire range, and the exponential result of the product reflects the contribution of a particular temperature value to the risk of mortality. Then we select a temperature and use its risk contribution as a reference to calculate the relative deviation of the risk contribution of other temperatures, which is considered as the relative change in risk and is used to reflect the association between different temperatures and the relative risk (RR) of mortality [7]. The trend of the temperature-mortality curve is not associated with the selection of specific reference temperature value. Here 7 °C is set as the reference temperature, whose left and right sides are placed at high and low temperatures for easy interpretation of the results in the research [8, 9]. We use the deviation information criterion (DIC) to select the best fit from different fitting functions with the lowest DIC value [38].

Next, based on the temperature-mortality relative risk relationship curve, we divide the low temperature part into three intervals of −26~−15 °C, −15~−5 °C and − 5 ~ 7 °C, and the high temperature part is divided into 7 ~ 15 °C, 15 ~ 25 °C, 25 ~ 34 °C. The linear effect of temperature variation on mortality risk is independently estimated in each interval, which is shown in Eq. (2):

where β accounts for the coefficient of temperature, indicating the relative risk for every 1 °C ascent or descent in mean monthly temperature after being exponential [34, 39]. We utilize the percentage change in relative risk to estimate the temperature effect for the mortality, which is denoted as (expβ−1)*100% [39].

Displacement effects of temperature variation

The displacement effects of temperature account for the impacts of the lag effect of the previous month and the lead effect of the following month on the mortality risk of the current month, respectively [7, 40]. Based on the distributed lag effect of temperature [8], a further improved Bayesian spatio-temporal model is used to estimate the displacement effects of temperature on mortality risk, as is shown in Eq. (3):

where β_L=0 denotes the temperature effect of the current month, and β_L=1 accounts for the temperature effect of the previous month. The accumulation of β_L=0 + β_L=1 represents the overall lagged effects of temperature on mortality. Similarly, the lead effect represents the effect of following month’s temperature on current mortality risk. Consistent with previous studies, the integrated nested Laplacian approximation (INLA) is used to quantify the accurate effect of temperature on the mortality risk of older adults [34, 39, 41].

Influence of extreme abnormal temperatures on mortality risk

We classify the geographical distribution of older adults according to the five major climatic zones of tropical monsoon, subtropical monsoon, temperate monsoon, temperate continent and alpine plateau in China [42]. Then, the influence of extreme high or low temperature on the mortality risk of older adults is assessed within each climatic zone.

We define the temperatures above the 97.5% percentile in the temperature range within each climate belt as the extreme high temperature and the temperatures below the 2.5% percentile as the extreme low temperature [9]. The results show that the 2.5th percentile of the temperate monsoon climate is −16.6 °C, and the 97.5th percentile is 27.3 °C; the 2.5th percentile of the subtropical monsoon climate is 2.5 °C, and the 97.5th percentile is 28.9 °C; the 2.5th percentile of the tropical monsoon climate is 7.6 °C, and the 97.5th percentile is 29.7 °C. The temperate continental and alpine plateau zones do not involve the older adults interviewed, therefore those two climatic zones are not included in the study. We use Eq. (3) to investigate the cumulative lagged effects of extreme high and low temperature on the mortality risk of older adults in each climate zone (see above).

Heterogeneity analysis of mortality risk in older adults

We assess the heterogeneity of mortality risk affected by temperature variations within the six intervals of 26~−15 °C, −15~−5 °C, −5 ~ 7 °C, 7 ~ 15 °C, 15 ~ 25 °C, and 25 ~ 34 °C separately for rural and urban areas by different socio-economic demographic and physiological characteristics, including smoking, drinking, comorbidities, education, age, gender, and income, respectively. We use Eq. (3) to estimate the effect of temperature rise or drop on the mortality risk of older adults within each subgroup. The significance of differences in temperature effects across different subgroups is assessed using Eq. (4).

where Q₁ and Q₂ represents the standard deviations SE₁ and SE₂ of the estimated temperature effect between any two subgroups. Similar to previous studies [43–45], we consider that the temperature effect factor ≥ 2 is important and noteworthy, suggesting that the corresponding subgroup is a disturbing factor affecting mortality in older adults.

Effects of temperature under future climate scenarios

To predict the effects of future temperature increase on Chinese older adults under different climate scenarios in the future, we use model (5) to quantify the effect of temperature on the mortality rate of the aged population [7]. The temperature effects are referred to as the sum of the previous month’s effect and the current month’s effect. The effect of the high temperature component on mortality is approximately linear, so consistent with Eq. (3). We use a linear function to calculate the additional effect of future temperature rise on death rate.

where death-rate_{county−time} is the relative mortality rate, representing the ratio of the count of deaths per month in each county to the total number of deaths in the sample (covering all older adults in the study period who were alive at baseline) [34, 37], and the meaning of the remaining variables remains unchanged with Eq. (3). We utilize the percentage change in the death rate to assess the magnitude of the temperature effect for older adults, which is denoted as (expβ^L−1)*100% [34].

We extend the temperature effect on the elderly death rate obtained in Eq. (5) to the entire aged population in China, and estimate the impact of temperature change on the mortality risk of older adults under different scenarios in the future. The expected temperature increment is derived from the temperature estimates of Shared Socioeconomic Pathway (SSP) scenarios, including the SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 under the CMIP6 [46]. The temperature increase under each scenario is estimated using a multi-model ensemble (MME) [47]. The projected monthly temperature change is calculated for each climate grid cell. We compute the expected monthly temperature change for each county in each month based on the county’s administrative divisions, in which the grid cells falling into are weighted [7]. We also obtain the expected annual change of China’s elderly population (aged > 65) from 2020 to 2050 [22], weighting by the population-gridded to acquire the expected number of older adults for each county in each year [48]. Equation (6) is utilized to calculate the additional impact of the expected temperature rise on the mortality of older adults [7]:

where Deaths accounts for the cumulative deaths of older adults in county c caused by the expected temperature rise, and PET_ct represents the expected number of older adults in each year t of county c from 2020 to 2050. β_c indicates the estimated overall lag effect (including the combined effects of the current and lagged temperatures) of each 1℃ temperature increase on the death rate of older adults. ∆T_cmt represents the expected monthly temperature change in county c over 12 months of each year from 2020 to 2050. The excess deaths in all counties are added up to compute the total number of deaths in older adults resulted by future temperature changes under different climate scenarios.

Sensitivity analysis

We adjust and control for different combinations of the time and space terms in Eq. (3), involving the county-specific intercept and its time-slope variables γ_county and δ_county, the month-specific intercept and its slope with time γ_month and δ_month, the county-month intercept and its slope term λ_{county−month} and φ_{county−month}. We use the varied model to explore the quantitative effects of high and low temperatures on the mortality risk of rural older adults, and the results of Supplementary Table S4 confirms the robustness of the findings.

Results

Association of temperature variations with mortality risk

We calculated the count of deaths in older adults per month in each county to assess the effect of monthly mean temperature on the deaths of older adults [7, 34]. We utilized a Bayesian spatio-temporal model to infer and characterize the nonlinear association between temperature and the mortality risk, which fully accounted for the temporal variation of temperature exposure and death date, the spatial variation of geographic location, and the temperature-mortality spatiotemporal interaction effect [34]. Four different fitting functions of f were used to depict the relationship between temperature variations and the relative risk of mortality (Methods). The deviation information criteria (DIC) [38] were used to select the optimum fit from various fitting functions, of which the results exhibit that the polynomial function achieves the optimum fit (in red curve). Fig. 1 illustrates that the relationship between temperature variations and the relative risk of mortality presents a U-shape, with the temperature in the middle having the least effect, while the high and low temperature at both ends of the curve have a larger impact on the relative risk, which is consistent with prior studies on temperature and mortality [8, 9].

Fig. 1 — Temperature-mortality risk relationship curve of older adults using four fitting functions including polynomial (red), B-spline (purple), and natural cubic spline with 3 (blue)/5 nodes (orange). The light blue shading accounts for the 95% confidence interval (CI) of the polynomial function, and the histogram on the horizontal axis represents the range distribution of temperature

Based on the distributed lag effect of temperature [8, 13], we extended the Bayesian spatio-temporal model to assess the displacement effects of temperature on the mortality risk of older adults (Methods). The displacement effects for the six different intervals of low and high temperatures in the rural and urban areas were quantified, respectively. Fig. 2 shows that the lag effect in the previous month is positive and significant, and the temperature effect in the following month (i.e., the lead effect) is negative, indicating that it does not promote the increase of the relative risk of mortality. The temperature of the current month exerts the most significant effect on promoting the mortality risk of older adults. Therefore, in the subsequent study, we refer to the overall lag effect of temperature (i.e., the sum of the previous and current month) to represent the influence of temperature on mortality risk [7].

Fig. 2 — Displacement effects of temperature for different low and high temperature intervals. a, −26–15 °C. b, −15–5 °C. c, −5–7 °C. d, 7–15 °C. e, 15–25 °C. f, 25–34 °C

Quantitative effect of high and low temperature

According to the association curve of Fig. 1, the ranges of low temperature are divided into three intervals of −26–15 °C, −15–5 °C, and − 5–7 °C; the ranges of high temperature include three intervals of 7–15 °C, 15–25 °C, and 25–34 °C. An improved Bayesian spatio-temporal model was used to estimate the quantitative effects of high or low temperature on mortality risk within various temperature intervals for the rural and urban areas, respectively (Methods). Supplementary Fig. S1 presents the monthly mean temperatures in urban and rural areas, where the baseline temperature in urban areas differs from that in rural areas. As shown in Fig. 3, the quantitative effects of different temperature intervals on the mortality risk are consistent with the result curve of Fig. 1, of which the relative risk of mortality of older adults increases with the descending of temperature; as the temperature rises, the relative risk increases. The relative mortality risk of older adults in rural areas increased by 3.57% (95% CI: 0.56-6.79%) for each 1 °C ascent of high temperature above 25 °C, which was much higher than that in urban areas 2.15% (95% CI: 0.27-4.06%), whereas increased by 8.94% (95% CI: 1.03-15.21%) for each 1 °C descent of low temperature below − 15 °C, much higher than that in urban areas 6.37% (95% CI: 1.83-12.92%). Overall, the effects of low temperatures on the mortality risk of older adults are more substantial than that of high temperatures, and older adults in rural areas are more sensitive to temperature variations compared with those in urban areas.

Fig. 3 — Quantified effects of high and low temperature variations on the relative risk of mortality. a, Rural areas. b, Urban areas. The low/high temperature effect refers to the corresponding relative risk change for each 1°C decrease/increase in temperature

Impacts of extreme temperature in climate zone

The area of mainland China is mainly distributed in five major climate zones [42], and we separately counted the deaths of older adults in urban and rural areas within each climate zone. The Bayesian spatio-temporal model was used to estimate the influence of extreme high and low temperatures on mortality risk of older adults in each climate zone (Methods). The temperate continental and alpine plateau zones were not included in the study for these regions did not involve the elderly interviewed. The urban population of tropical monsoon zone, subtropical monsoon zone, and temperate monsoon zone were estimated at 84 million, 351 million, and 163 million, and the rural population was estimated at 36 million, 189 million, and 77 million, respectively, in the seventh national census of China [49]. As shown in Fig. 4 and Supplementary Table S1, in the areas located in the tropical monsoon zone, the relative risk of mortality for older adults achieves the largest and increased by 3.51% (95%CI, 1.81–6.24) for every 1 °C increase in extreme high temperature. For the older adults located in the temperate monsoon zone, the relative risk of mortality for older adults achieves the largest and increased by 8.93% (95%CI, 4.48–13.69) for every 1 °C decrease in extreme low temperature. Overall, we find that the effects of extreme high or low temperatures are larger in the rural areas within different climatic zones than that in the urban areas.

Heterogeneity of mortality risk in older adults

We investigated the heterogeneity of high and low temperatures on the mortality risk of older adults in rural and urban areas, respectively, and explored the differences of the temperature impact on the mortality risk in various subgroups (Methods). The heterogeneity analysis of the impacts of different temperature ranges on the mortality risk of older adults in rural areas are shown in Supplementary Figs. S2-S3. Supplementary Fig. S2 illustrates that older adults in rural areas with older age, lower education, and lower annual income are more likely to be affected by ambient temperature, and women have a greater potential mortality risk than men. Supplementary Fig. S3 shows that ambient temperature of various ranges in rural areas contributes to an increased risk of death in older adults with medical conditions including hypertension, heart disease, chronic drinking habits, arthritis, stroke, Parkinson’s disease, and Alzheimer’s disease relative to their counterparts. The heterogeneity analysis of the impact of different temperature ranges on the mortality risk of older adults in urban areas are shown in Supplementary Figs. S4-S5. Supplementary Fig. S4 exhibits that older adults in urban areas who are older, female, with lower education and annual income, are more affected by the external environment of various temperature ranges, which is consistent with rural areas. Supplementary Fig. S5 illustrates the contribution of ambient temperature in urban areas to an increased risk of death in older adults with medical conditions including hypertension, heart disease, stroke, Parkinson’s disease, and Alzheimer’s disease. The effects of high and low temperature exposure on the mortality risk of older adults are significantly heterogeneous among different subgroups.

Impacts of future temperature variations

Based on the combination of the shared socioeconomic path and the typical concentration path under the CMIP6 [46], we estimated the annual change of temperature in mainland China under five new scenario models, including the scenarios of SSP1-1.9, SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5. Combined with the expected changing trend of the elderly population in mainland China [22], we calculated the count of additional deaths of older adults in mainland China resulted by the expected annual temperature rising from 2020 to 2050 (Methods).

As shown in Fig. 5, the extra deaths in older adults caused by temperature changes are different with different climate scenarios. Under SSP5-8.5, the global temperature rises significantly, resulting in a rapid increase in the extra deaths of older adults. With the strengthening of energy conservation and emission reduction measures, the temperature rising trend under SSP3-7.0, SSP2-4.5, SSP1-2.6 and SSP1-1.9 is weakened, respectively, and the corresponding extra deaths of the aged population also decreases. The additional cumulative deaths in older adults caused by the expected temperature rise under the five climate scenarios were calculated for various periods in the future from 2020 to 2050, and the results are shown in Supplementary Table S2. The expected future cumulative deaths of older adults in mainland China due to temperature rise under the most extreme climate scenario of SSP5-8.5 could approach to 121,093 (95% CI: 73,702 − 198,994), 243,987 (95% CI: 131,374–454,125) and 473,351 (95% CI: 282,989 − 878,693) from 2020 to 2030, 2020 to 2040, 2020 to 2050, respectively.

Fig. 4 — The impacts of extreme high or low temperature in different climate zones for rural and urban area. a, Extreme high temperature for rural area. b, Extreme low temperature for rural area. c, Extreme high temperature for urban area. d, Extreme low temperature for urban area. The three red or blue regions represent the increase in mortality risk of older adults in each climatic zone for every 1°C rise/fall under extreme high or low temperature conditions, respectively, and darker colored areas correspond to a greater increased risk of mortality. The height of the three-dimensional prism accounts for the total number of deaths of older adults in each province, and the gray area represents the temperate continental climate zone and the plateau alpine climate zone, where older adults in this study are not involved.

Sensitivity analysis

We varied the spatio-temporally structured or unstructured terms in model (3) to verify the reliability and robustness of the model (see Methods). We used the adjusted model to investigate the quantitative effects of high and low temperatures on the mortality risk of rural older adults. The results of Supplementary Table S3 exhibit that various combinations of different time and space terms in the model have less interference on the quantified relationship between high and low temperatures and mortality risk among rural older adults, which further confirms and reflects the reliability of this research.

Discussion

Exposure to abnormal high and low temperature variations have been demonstrated to exert profound and long-lasting effects on various health outcomes and mortality [9, 12]. Our study included a relatively large sample of deaths in older adults in China from 2005 to 2018, covering 80% range of mainland China. Compared with previous studies on temperature and mortality risk [9, 10, 34], this research controlled for additional socioeconomic demographic characteristics (residential environment, annual income, education) health status, and health-risk behaviors of elderly individuals (medical comorbidities, smoking, and drinking) with a long-term follow-up. We have improved the Bayesian spatio-temporal models suitable for characterizing and quantifying the nonlinear relationship between temperature variations and the mortality risk of older adults, and have explored the heterogeneity of the mortality risk affected by temperature with different socioeconomic, geographical location, demographic characteristics, health status and health-risk behaviors.

High and low temperature were found to promote the mortality in older adults, which was consistent with several prior researches on temperature and mortality [8, 9]. The effects of high or low temperature on the mortality risks of older adults were greater in rural areas than that of urban areas. Specifically, low temperatures were found to exert a greater impact on mortality risk than high temperatures. This might be explained by the large temperature difference between indoor and outdoor temperatures in winter, which increases the death of the older adults from cardiovascular and cerebrovascular diseases [50]. The displacement effect of temperature indicates that the temperature in the current month has the greatest effect on mortality risk, the temperature in the previous month has an effect on mortality risk, while the temperature in the next month has no significant or negative effect on mortality risk, which is similar to previous studies [7]. Mainland China mainly includes five different types of climatic zones [42], three of which are examined by our research. The population distribution within these climatic zones contributes a high proportion of the Chinese population [51], which are representative for assessing the relationship between the mortality risk and temperature variations. We find that extreme high temperature has the greatest impact on the tropical monsoon zone and the least impact on the temperate monsoon zone, while extreme low temperature has the greatest impact on the temperate monsoon zone and the least impact on the tropical monsoon zone, which is caused by the difference between the upper and lower boundaries of extreme high and low temperatures, as shown in Supplementary Table S4. Higher or lower temperature extremes exert greater influence on the mortality risk [8, 9].

The results illustrate that rural older adults who were female, with low education, medical conditions (e.g., hypertension, heart disease, arthritis, stroke, Parkinson’s disease, or Alzheimer), and chronic drinking, and without spouses and children, have a higher heat-related mortality risk compared with their counterparts. Older adults with higher annual income and education in urban were less affected by the environmental temperature. This might be explained by that older adults in urban areas with higher education and annual income afford better economic capacity and conditions to mitigate the impact of external uncomfortable temperature, including the penetration rate of air conditioning or the application of winter heating facilities [4, 18]. Moreover, we find that women were more sensitive to the effects of temperature than men. This is mainly related to the greater tolerance of men in higher or lower temperature environments compared to women, to some extent reducing the event of heat or cold stress-related death [43, 52]. The additional deaths of older adults in China caused by future temperature rise should not be underestimated, but our research has also confirmed that reasonable energy conservation and emission reduction can effectively reduce the excess deaths of older adults in mainland China due to the expected rise of temperature. The mortality data of the elderly in this study are distributed throughout China, involving a wider range of temperature and climate zone differences compared with previous similar studies [53]. We use the Bayesian geographic spatiotemporal model to better fully control for and capture the impact of complex changes over time and space on temperature-mortality risk [34, 54]. The mortality risk of the elderly is affected by more risk factors [55], and we provide a more detailed stratified analysis compared to several prior studies [10, 11, 53], especially the multiple comorbidities and different socio-demographic characteristics of the elderly. Our study estimates the expected mortality of older Chinese adults from future temperature increases [7], which is crucial for understanding the impact of future climate change on the mortality risk of the elderly. These findings provide an insight into the association of temperature-mortality risk among older adults, and raise concerns of the vulnerability of older adults to extreme temperature exposure.

There are some limitations in the research. The temperate continental and alpine plateau zone were not examined in this study. More adequate survey data would help to fully identify and understand the impact of temperature variations within these climatic zones on mortality risk of older adults in the future. We used the total number of older adults in the sample as the baseline data to calculate the monthly death rate to estimate the additional deaths caused by future temperature rise, which might be underestimated to a certain extent of deaths in older adults [7]. Our projections focused specifically on the additional mortality burden associated with rising temperatures under future climate change scenarios, with an emphasis on heat-related health impacts. We did not account for the potential decline in cold-related mortality that may result from milder winters in a warming climate. Higher winter temperatures could reduce cold-related deaths due to less frequent or less severe cold events [56], and such reductions may partly offset the overall temperature-related mortality burden. The primary aim of the study was to quantify the additional mortality burden attributable to rising temperatures under climate change scenarios, with a focus on heat-related impacts. This approach reflects the growing concern over the negative and urgent consequences of rising temperatures, which are a dominant trend under climate change and pose increasing risks to public health [1, 7]. Our modeling framework did not include this aspect due to the lack of reliable estimates or long-term projections for future temperature drops in cold-related mortality under climate scenarios. Therefore, the potential protective effects of rising temperatures for cold-related mortality were not incorporated in this study, and this remains an important area for future research. In the heterogeneity analysis, we assessed the effect of temperature on mortality risk in older adults by stratification according to urban/rural, individual characteristics, and disease, and examined the significant differences between different subgroups to highlight the key grouping variables affecting mortality risk in older adults. We noticed that the confidence intervals of the stratified analysis partially overlapped, which might cause the results of the temperature effect to appear exaggerated. The meteorological data we used had a resolution of 30 km, indicating that the impact of the “urban heat island” was not taken into account, which affects the accurate assessment of temperatures in urban areas or within cities, leading to the biased quantification of the temperature effect on mortality risk. Air pollutants such as NO₂, S0₂, and CO were not included in the model, mainly because the time span of the data satisfying the accuracy requirements did not match that of this study. Other air pollutants with longer time spans that satisfy the accuracy requirements can be controlled in the model in subsequent studies, which are further used to detect the impact of the interference effect of confounding factors on the research results. Due to the limitations of data observation and evaluation, we are unable to obtain the specific values of household air pollution, which would lead to certain deviations in the assessment of the exposure level of the elderly to air pollution. More accurate data on household air pollution could improve the accuracy of the assessment of the elderly’s exposure to air pollutants. We cannot derive the geographic information at the township level for the limitation of data information acquisition. Clustering at the township level is one of the factors influencing the research results, which can be addressed based on richer geographic information data in subsequent research.

Our study reveals and demonstrates the difference in the adverse effects of high and low temperatures on the aged population between rural and urban areas in China. Effective energy conservation and emission reduction measures are proven to be effective in reducing the additional elderly deaths caused by temperature rise based on our findings, and necessary heating or cooling measures are recommended for the elderly population exposed to specific ambient temperatures simultaneously. The quantitative assessment of this study advances a more nuanced clarification of the epidemiology of the temperature-mortality association, particularly regarding the negative impacts of extreme high or low temperature, and also contributes to the implementation of targeted public health interventions in the areas sensitive to climate change.

Conclusion

The results of this study provide evidence of regional inequalities in the mortality risk of older Chinese population exposed to abnormally high and low temperatures, especially in rural areas where the risk of mortality is significantly greater than that of urban areas. This quantitative assessment contributes to a more accurate understanding of the epidemiology of mortality associated with ambient temperature exposure, and also offers an insight into the health-related effects of climate change.

Supplementary Information

Supplementary Material 1.^{(564.7KB, docx)}

Acknowledgements

We appreciate the contributions of all the data collectors involved and the cooperation of all the participants involved in the study.

Abbreviations

CMIP6: The International Coupled Model Comparison Program Phase 6
CLHLS: Chinese Longitudinal Healthy Longevity Survey
ERA5: The fifth generation ECMWF reanalysis for the global climate and weather
CHA: China High Air Pollutants
DIC: Deviation information criterion
INLA: Integrated nested Laplacian approximation
SSP: Shared Socioeconomic Pathway
MME: Multi-model ensemble

Authors’ contributions

K.H. contributed to the conception of the study, performed the experiment and the data analyses, and wrote the manuscript. L.Z., F.Y., and W.H. contributed significantly to analysis and manuscript preparation. K.H. and X.X. performed the data analyses and figures. X.X., X.Y., and Z.W. helped perform the analysis with constructive discussions.

Funding

This research was funded by the National Natural Science Foundation of China (Grant No. 42401519), the Natural Science Foundation of Jiangsu Province (Grant No. BK20230430), and the Startup Foundation for Introducing Talent of NUIST (Grant No. 2022r041).

Data availability

The CLHLS dataset is accessible to researchers upon reasonable request through its official public repository. Comprehensive information regarding the data access application process can be found at: https://opendata.pku.edu.cn/dataverse/CHADS.

Declarations

Ethics approval and consent to participate

The authors declare that this study adhered to the Declaration of Helsinki. The Ethical Committee of Peking University gave its approval before the study began (Registration number IRB00001052-13074). Written informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1.Watts N, Adger WN, Agnolucci P, Blackstock J, Byass P, Cai W, Chaytor S, Colbourn T, Collins M, Cooper A, et al. Health and climate change: policy responses to protect public health. Lancet. 2015;386(10006):1861–914. [DOI] [PubMed] [Google Scholar]
2.Costello A, Abbas M, Allen A, Ball S, Bell S, Bellamy R, Friel S, Groce N, Johnson A, Kett M, et al. Managing the health effects of climate change: lancet and university college London Institute for global health commission. Lancet. 2009;373(9676):1693–733. [DOI] [PubMed] [Google Scholar]
3.Godlee F. Climate change. BMJ. 2014;349: g5945. [DOI] [PubMed] [Google Scholar]
4.Green H, Bailey J, Schwarz L, Vanos J, Ebi K, Benmarhnia T. Impact of heat on mortality and morbidity in low and middle income countries: a review of the epidemiological evidence and considerations for future research. Environ Res. 2019;171:80–91. [DOI] [PubMed] [Google Scholar]
5.Watts N, Amann M, Arnell N, Ayeb-Karlsson S, Belesova K, Berry H, Bouley T, Boykoff M, Byass P, Cai W, et al. The 2018 report of the lancet countdown on health and climate change: shaping the health of nations for centuries to come. Lancet. 2018;392(10163):2479–514. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Watts N, Adger WN, Ayeb-Karlsson S, Bai Y, Byass P, Campbell-Lendrum D, Colbourn T, Cox P, Davies M, Depledge M, et al. The lancet countdown: tracking progress on health and climate change. Lancet. 2017;389(10074):1151–64. [DOI] [PubMed] [Google Scholar]
7.Burke M, González F, Baylis P, Heft-Neal S, Baysan C, Basu S, Hsiang S. Higher temperatures increase suicide rates in the united States and Mexico. NAT CLIM CHANGE. 2018;8(8):723–9. [Google Scholar]
8.Chen R, Yin P, Wang L, Liu C, Niu Y, Wang W, Jiang Y, Liu Y, Liu J, Qi J, et al. Association between ambient temperature and mortality risk and burden: time series study in 272 main Chinese cities. BMJ-BRIT MED J. 2018;363:k4306. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Gasparrini A, Guo Y, Hashizume M, Lavigne E, Zanobetti A, Schwartz J, Tobias A, Tong S, Rocklöv J, Forsberg B, et al. Mortality risk attributable to high and low ambient temperature: a multicountry observational study. Lancet. 2015;386(9991):369–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Ma W, Chen R, Kan H. Temperature-related mortality in 17 large Chinese cities: how heat and cold affect mortality in China. Environ Res. 2014;134:127–33. [DOI] [PubMed] [Google Scholar]
11.Kephart JL, Sánchez BN, Moore J, Schinasi LH, Bakhtsiyarava M, Ju Y, Gouveia N, Caiaffa WT, Dronova I, Arunachalam S, et al. City-level impact of extreme temperatures and mortality in Latin America. Nat Med. 2022;28(8):1700–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Guo Y, Gasparrini A, Armstrong B, Li S, Tawatsupa B, Tobias A, Lavigne E, de Coelho SZS, Leone M, Pan M et al. X: Global Variation in the Effects of Ambient Temperature on Mortality: A Systematic Evaluation. EPIDEMIOLOGY 2014;25(6):781-89. [DOI] [PMC free article] [PubMed]
13.Zhao Q, Guo Y, Ye T, Gasparrini A, Tong S, Overcenco A, Urban A, Schneider A, Entezari A, Vicedo-Cabrera AM, et al. Global, regional, and National burden of mortality associated with non-optimal ambient temperatures from 2000 to 2019: a three-stage modelling study. Lancet Planet Health. 2021;5(7):e415–25. [DOI] [PubMed] [Google Scholar]
14.Lay CR, Sarofim MC, Vodonos Zilberg A, Mills DM, Jones RW, Schwartz J, Kinney PL. City-level vulnerability to temperature-related mortality in the USA and future projections: a geographically clustered meta-regression. Lancet Planet Health. 2021;5(6):e338-46. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Weinberger KR, Harris D, Spangler KR, Zanobetti A, Wellenius GA. Estimating the number of excess deaths attributable to heat in 297 united States counties. Environ Epidemiol 2020;4(3):e096. [DOI] [PMC free article] [PubMed]
16.Martínez-Solanas È, Quijal-Zamorano M, Achebak H, Petrova D, Robine J, Herrmann FR, Rodó X, Ballester J. Projections of temperature-attributable mortality in Europe: a time series analysis of 147 contiguous regions in 16 countries. Lancet Planet Health. 2021;5(7):e446-54. [DOI] [PubMed] [Google Scholar]
17.Zeng W, Yu M, Mai W, Zhou M, Zhou C, Xiao Y, Hou Z, Xu Y, Liu T, Hu J, et al. Age-specific disparity in life loss per death attributable to ambient temperature: A nationwide time-series study in China. ENVIRON RES. 2022;203:111834. [DOI] [PubMed] [Google Scholar]
18.Scheelbeek PFD, Dangour AD, Jarmul S, Turner G, Sietsma AJ, Minx JC, Callaghan M, Ajibade I, Austin SE, Biesbroek R, et al. The effects on public health of climate change adaptation responses: a systematic review of evidence from low- and middle-income countries. Environ Res Lett. 2021;16(7): 73001. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Kang S, Eltahir EAB. North China Plain threatened by deadly heatwaves due to climate change and irrigation. Nat Commun. 2018;9(1): 2894. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Witze A. Extreme heatwaves: surprising lessons from the record warmth. Nature. 2022;608:464–65. [DOI] [PubMed]
21.Luo M, Lau N. Increasing heat stress in urban areas of Eastern china: acceleration by urbanization. GEOPHYS RES LETT. 2018;45(23):13. [Google Scholar]
22.Tatum M. China’s population peak. Lancet. 2022;399(10324):509. [DOI] [PubMed] [Google Scholar]
23.Cai W, Zhang C, Zhang S, Ai S, Bai Y, Bao J, Chen B, Chang N, Chen H, Cheng L, et al. The 2021 China report of the lancet countdown on health and climate change: seizing the window of opportunity. Lancet Public Health. 2021;6(12):e932–47. [DOI] [PubMed] [Google Scholar]
24.Arbuthnott K, Hajat S, Heaviside C, Vardoulakis S. Changes in population susceptibility to heat and cold over time: assessing adaptation to climate change. ENVIRON HEALTH-GLOB. 2016;15(1):S33. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Brown ME, Funk CC. Food security under climate change. Science. 2008;319(5863):580–1. [DOI] [PubMed] [Google Scholar]
26.Morton JF. The impact of climate change on smallholder and subsistence agriculture. Proc Natl Acad Sci U S A. 2007;104(50):19680–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Leider JP, Meit M, McCullough JM, Resnick B, Dekker D, Alfonso YN, Bishai D. The state of rural public health: enduring needs in a new decade. Am J Public Health. 2020;110(9):1283–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Grace K. Considering climate in studies of fertility and reproductive health in poor countries. NAT CLIM CHANGE. 2017;7(7):479–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Liu C, Chen R, Sera F, Vicedo-Cabrera AM, Guo Y, Tong S, Coelho MSZS, Saldiva PHN, Lavigne E, Matus P, et al. Ambient particulate air pollution and daily mortality in 652 cities. NEW ENGL J MED. 2019;381(8):705–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Di Q, Dai L, Wang Y, Zanobetti A, Choirat C, Schwartz JD, Dominici F. Association of short-term exposure to air pollution with mortality in older adults. JAMA. 2017;318(24):2446–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Malashock DA, Delang MN, Becker JS, Serre ML, West JJ, Chang K, Cooper OR, Anenberg SC. Global trends in ozone concentration and attributable mortality for urban, peri-urban, and rural areas between 2000 and 2019: a modelling study. Lancet Planet Health. 2022;6(12):e958-67. [DOI] [PubMed] [Google Scholar]
32.Wei J, Li Z, Li K, Dickerson RR, Pinker RT, Wang J, Liu X, Sun L, Xue W, Cribb M. Full-coverage mapping and Spatiotemporal variations of ground-level Ozone (O3) pollution from 2013 to 2020 across China. REMOTE SENS ENVIRON. 2022;270:112775. [Google Scholar]
33.Wei J, Li Z, Lyapustin A, Sun L, Peng Y, Xue W, Su T, Cribb M. Reconstructing 1-km-resolution high-quality PM2.5 data records from 2000 to 2018 in china: Spatiotemporal variations and policy implications. REMOTE SENS ENVIRON. 2021;252:112136. [Google Scholar]
34.Parks RM, Bennett JE, Tamura-Wicks H, Kontis V, Toumi R, Danaei G, Ezzati M. Anomalously warm temperatures are associated with increased injury deaths. Nat Med. 2020;26(1):65–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Morris M, Wheeler-Martin K, Simpson D, Mooney SJ, Gelman A, DiMaggio C. Bayesian hierarchical spatial models: implementing the Besag York Mollié model in Stan. Spatial and Spatio-temporal Epidemiology. 2019;31: 100301. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Nothdurft A. Climate sensitive single tree growth modeling using a hierarchical Bayes approach and integrated nested Laplace approximations (INLA) for a distributed lag model. For ECOL MANAG. 2020;478:118497. [Google Scholar]
37.Kontis V, Bennett JE, Rashid T, Parks RM, Pearson-Stuttard J, Guillot M, Asaria P, Zhou B, Battaglini M, Corsetti G, et al. Magnitude, demographics and dynamics of the effect of the first wave of the COVID-19 pandemic on all-cause mortality in 21 industrialized countries. Nat Med. 2020;26(12):1919–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Shriner D, Yi N. Deviance information criterion (DIC) in bayesian multiple QTL mapping. COMPUT STAT DATA AN. 2009;53(5):1850–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Blangiardo M, Cameletti M, Baio G, Rue H. Spatial and spatio-temporal models with R-INLA. SPAT SPATIO-TEMPORAL. 2013;4:33–49. [DOI] [PubMed] [Google Scholar]
40.Hou K, Xu X. Ambient temperatures associated with reduced cognitive function in older adults in China. SCI REP-UK. 2023;13(1):17414. [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Gómez-Rubio V, Rue H. Markov chain Monte Carlo with the integrated nested Laplace approximation. Stat Comput. 2018;28(5):1033–51. [Google Scholar]
42.Geng X, Zhang D, Li C, Yuan Y, Yu Z, Wang X. Impacts of Climatic zones on urban heat island: Spatiotemporal variations, trends, and drivers in China from 2001–2020. SUSTAIN CITIES SOC. 2023;89:104303. [Google Scholar]
43.Kan H, London SJ, Chen G, Zhang Y, Song G, Zhao N, Jiang L, Chen B. Season, sex, age, and education as modifiers of the effects of outdoor air pollution on daily mortality in shanghai, china: the public health and air pollution in Asia (PAPA) study. ENVIRON HEALTH PERSP. 2008;116(9):1183–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Lei J, Peng L, Yang T, Huang S, Zhu Y, Gao Y, Zhou L, Shi S, Liu C, Kan H, et al. Non-optimum ambient temperature May decrease pulmonary function: A longitudinal study with intensively repeated measurements among asthmatic adult patients in 25 Chinese cities. ENVIRON INT. 2022;164:107283. [DOI] [PubMed] [Google Scholar]
45.Fan P, Xue X, Hu J, Qiao Q, Yin T, Yang X, Chen X, Hou Y, Chen R. Ambient temperature and ambulatory blood pressure: an hourly–level, longitudinal panel study. SCI TOTAL ENVIRON. 2023;864:160854. [DOI] [PubMed] [Google Scholar]
46.Yang X, Zhou B, Xu Y, Han Z. CMIP6 evaluation and projection of temperature and precipitation over China. Adv Atmos Sci. 2021;38(5):817–30. [Google Scholar]
47.Ahmed K, Sachindra DA, Shahid S, Iqbal Z, Nawaz N, Khan N. Multi-model ensemble predictions of precipitation and temperature using machine learning algorithms. ATMOS RES. 2020;236:104806. [Google Scholar]
48.Chen Y, Guo F, Wang J, Cai W, Wang C, Wang K. Provincial and gridded population projection for China under shared socioeconomic pathways from 2010 to 2100. SCI DATA. 2020;7(1): 83. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Chen Y, Xu C, Ge Y, Zhang X, Zhou Y. A 100 m gridded population dataset of China’s seventh census using ensemble learning and big geospatial data. Earth Syst Sci Data. 2024;16(8):3705–18. [Google Scholar]
50.Barnett AG. Temperature and cardiovascular deaths in the US elderly: changes over time. Epidemiology. 2007;18(3):369. [DOI] [PubMed] [Google Scholar]
51.Chen M, Xian Y, Huang Y, Zhang X, Hu M, Guo S, Chen L, Liang L. Fine-scale population spatialization data of China in 2018 based on real location-based big data. SCI DATA. 2022;9(1): 624. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Renberg J, Lignier MJ, Wiggen ØN, Færevik H, Helgerud J, Sandsund M. Heat tolerance during uncompensable heat stress in men and women wearing firefighter personal protective equipment. APPL ERGON. 2022;101:103702. [DOI] [PubMed] [Google Scholar]
53.Hu K, Guo Y, Hochrainer-Stigler S, Liu W, See L, Yang X, Zhong J, Fei F, Chen F, Zhang Y, et al. Evidence for Urban–Rural disparity in Temperature–Mortality relationships in Zhejiang province, China. ENVIRON HEALTH PERSP. 2019;127(3):37001. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Bilancia M, Demarinis G. Bayesian scanning of spatial disease rates with integrated nested Laplace approximation (INLA). Stat Methods Appl. 2014;23(1):71–94. [Google Scholar]
55.Xi D, Liu L, Zhang M, Huang C, Burkart KG, Ebi K, Zeng Y, Ji JS. Risk factors associated with heatwave mortality in Chinese adults over 65 years. Nat Med. 2024;30(5):1489–98. [DOI] [PubMed] [Google Scholar]
56.Wang S, Zhan W, Zhou B, Tong S, Chakraborty TC, Wang Z, Huang K, Du H, Middel A, Li J, et al. Dual impact of global urban overheating on mortality. NAT CLIM CHANGE. 2025;15(5):497–504. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1.^{(564.7KB, docx)}

Data Availability Statement

[CR1] 1.Watts N, Adger WN, Agnolucci P, Blackstock J, Byass P, Cai W, Chaytor S, Colbourn T, Collins M, Cooper A, et al. Health and climate change: policy responses to protect public health. Lancet. 2015;386(10006):1861–914. [DOI] [PubMed] [Google Scholar]

[CR2] 2.Costello A, Abbas M, Allen A, Ball S, Bell S, Bellamy R, Friel S, Groce N, Johnson A, Kett M, et al. Managing the health effects of climate change: lancet and university college London Institute for global health commission. Lancet. 2009;373(9676):1693–733. [DOI] [PubMed] [Google Scholar]

[CR3] 3.Godlee F. Climate change. BMJ. 2014;349: g5945. [DOI] [PubMed] [Google Scholar]

[CR4] 4.Green H, Bailey J, Schwarz L, Vanos J, Ebi K, Benmarhnia T. Impact of heat on mortality and morbidity in low and middle income countries: a review of the epidemiological evidence and considerations for future research. Environ Res. 2019;171:80–91. [DOI] [PubMed] [Google Scholar]

[CR5] 5.Watts N, Amann M, Arnell N, Ayeb-Karlsson S, Belesova K, Berry H, Bouley T, Boykoff M, Byass P, Cai W, et al. The 2018 report of the lancet countdown on health and climate change: shaping the health of nations for centuries to come. Lancet. 2018;392(10163):2479–514. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Watts N, Adger WN, Ayeb-Karlsson S, Bai Y, Byass P, Campbell-Lendrum D, Colbourn T, Cox P, Davies M, Depledge M, et al. The lancet countdown: tracking progress on health and climate change. Lancet. 2017;389(10074):1151–64. [DOI] [PubMed] [Google Scholar]

[CR7] 7.Burke M, González F, Baylis P, Heft-Neal S, Baysan C, Basu S, Hsiang S. Higher temperatures increase suicide rates in the united States and Mexico. NAT CLIM CHANGE. 2018;8(8):723–9. [Google Scholar]

[CR8] 8.Chen R, Yin P, Wang L, Liu C, Niu Y, Wang W, Jiang Y, Liu Y, Liu J, Qi J, et al. Association between ambient temperature and mortality risk and burden: time series study in 272 main Chinese cities. BMJ-BRIT MED J. 2018;363:k4306. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Gasparrini A, Guo Y, Hashizume M, Lavigne E, Zanobetti A, Schwartz J, Tobias A, Tong S, Rocklöv J, Forsberg B, et al. Mortality risk attributable to high and low ambient temperature: a multicountry observational study. Lancet. 2015;386(9991):369–75. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Ma W, Chen R, Kan H. Temperature-related mortality in 17 large Chinese cities: how heat and cold affect mortality in China. Environ Res. 2014;134:127–33. [DOI] [PubMed] [Google Scholar]

[CR11] 11.Kephart JL, Sánchez BN, Moore J, Schinasi LH, Bakhtsiyarava M, Ju Y, Gouveia N, Caiaffa WT, Dronova I, Arunachalam S, et al. City-level impact of extreme temperatures and mortality in Latin America. Nat Med. 2022;28(8):1700–5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Guo Y, Gasparrini A, Armstrong B, Li S, Tawatsupa B, Tobias A, Lavigne E, de Coelho SZS, Leone M, Pan M et al. X: Global Variation in the Effects of Ambient Temperature on Mortality: A Systematic Evaluation. EPIDEMIOLOGY 2014;25(6):781-89. [DOI] [PMC free article] [PubMed]

[CR13] 13.Zhao Q, Guo Y, Ye T, Gasparrini A, Tong S, Overcenco A, Urban A, Schneider A, Entezari A, Vicedo-Cabrera AM, et al. Global, regional, and National burden of mortality associated with non-optimal ambient temperatures from 2000 to 2019: a three-stage modelling study. Lancet Planet Health. 2021;5(7):e415–25. [DOI] [PubMed] [Google Scholar]

[CR14] 14.Lay CR, Sarofim MC, Vodonos Zilberg A, Mills DM, Jones RW, Schwartz J, Kinney PL. City-level vulnerability to temperature-related mortality in the USA and future projections: a geographically clustered meta-regression. Lancet Planet Health. 2021;5(6):e338-46. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Weinberger KR, Harris D, Spangler KR, Zanobetti A, Wellenius GA. Estimating the number of excess deaths attributable to heat in 297 united States counties. Environ Epidemiol 2020;4(3):e096. [DOI] [PMC free article] [PubMed]

[CR16] 16.Martínez-Solanas È, Quijal-Zamorano M, Achebak H, Petrova D, Robine J, Herrmann FR, Rodó X, Ballester J. Projections of temperature-attributable mortality in Europe: a time series analysis of 147 contiguous regions in 16 countries. Lancet Planet Health. 2021;5(7):e446-54. [DOI] [PubMed] [Google Scholar]

[CR17] 17.Zeng W, Yu M, Mai W, Zhou M, Zhou C, Xiao Y, Hou Z, Xu Y, Liu T, Hu J, et al. Age-specific disparity in life loss per death attributable to ambient temperature: A nationwide time-series study in China. ENVIRON RES. 2022;203:111834. [DOI] [PubMed] [Google Scholar]

[CR18] 18.Scheelbeek PFD, Dangour AD, Jarmul S, Turner G, Sietsma AJ, Minx JC, Callaghan M, Ajibade I, Austin SE, Biesbroek R, et al. The effects on public health of climate change adaptation responses: a systematic review of evidence from low- and middle-income countries. Environ Res Lett. 2021;16(7): 73001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Kang S, Eltahir EAB. North China Plain threatened by deadly heatwaves due to climate change and irrigation. Nat Commun. 2018;9(1): 2894. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Witze A. Extreme heatwaves: surprising lessons from the record warmth. Nature. 2022;608:464–65. [DOI] [PubMed]

[CR21] 21.Luo M, Lau N. Increasing heat stress in urban areas of Eastern china: acceleration by urbanization. GEOPHYS RES LETT. 2018;45(23):13. [Google Scholar]

[CR22] 22.Tatum M. China’s population peak. Lancet. 2022;399(10324):509. [DOI] [PubMed] [Google Scholar]

[CR23] 23.Cai W, Zhang C, Zhang S, Ai S, Bai Y, Bao J, Chen B, Chang N, Chen H, Cheng L, et al. The 2021 China report of the lancet countdown on health and climate change: seizing the window of opportunity. Lancet Public Health. 2021;6(12):e932–47. [DOI] [PubMed] [Google Scholar]

[CR24] 24.Arbuthnott K, Hajat S, Heaviside C, Vardoulakis S. Changes in population susceptibility to heat and cold over time: assessing adaptation to climate change. ENVIRON HEALTH-GLOB. 2016;15(1):S33. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Brown ME, Funk CC. Food security under climate change. Science. 2008;319(5863):580–1. [DOI] [PubMed] [Google Scholar]

[CR26] 26.Morton JF. The impact of climate change on smallholder and subsistence agriculture. Proc Natl Acad Sci U S A. 2007;104(50):19680–5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Leider JP, Meit M, McCullough JM, Resnick B, Dekker D, Alfonso YN, Bishai D. The state of rural public health: enduring needs in a new decade. Am J Public Health. 2020;110(9):1283–90. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Grace K. Considering climate in studies of fertility and reproductive health in poor countries. NAT CLIM CHANGE. 2017;7(7):479–85. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Liu C, Chen R, Sera F, Vicedo-Cabrera AM, Guo Y, Tong S, Coelho MSZS, Saldiva PHN, Lavigne E, Matus P, et al. Ambient particulate air pollution and daily mortality in 652 cities. NEW ENGL J MED. 2019;381(8):705–15. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Di Q, Dai L, Wang Y, Zanobetti A, Choirat C, Schwartz JD, Dominici F. Association of short-term exposure to air pollution with mortality in older adults. JAMA. 2017;318(24):2446–56. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Malashock DA, Delang MN, Becker JS, Serre ML, West JJ, Chang K, Cooper OR, Anenberg SC. Global trends in ozone concentration and attributable mortality for urban, peri-urban, and rural areas between 2000 and 2019: a modelling study. Lancet Planet Health. 2022;6(12):e958-67. [DOI] [PubMed] [Google Scholar]

[CR32] 32.Wei J, Li Z, Li K, Dickerson RR, Pinker RT, Wang J, Liu X, Sun L, Xue W, Cribb M. Full-coverage mapping and Spatiotemporal variations of ground-level Ozone (O3) pollution from 2013 to 2020 across China. REMOTE SENS ENVIRON. 2022;270:112775. [Google Scholar]

[CR33] 33.Wei J, Li Z, Lyapustin A, Sun L, Peng Y, Xue W, Su T, Cribb M. Reconstructing 1-km-resolution high-quality PM2.5 data records from 2000 to 2018 in china: Spatiotemporal variations and policy implications. REMOTE SENS ENVIRON. 2021;252:112136. [Google Scholar]

[CR34] 34.Parks RM, Bennett JE, Tamura-Wicks H, Kontis V, Toumi R, Danaei G, Ezzati M. Anomalously warm temperatures are associated with increased injury deaths. Nat Med. 2020;26(1):65–70. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Morris M, Wheeler-Martin K, Simpson D, Mooney SJ, Gelman A, DiMaggio C. Bayesian hierarchical spatial models: implementing the Besag York Mollié model in Stan. Spatial and Spatio-temporal Epidemiology. 2019;31: 100301. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Nothdurft A. Climate sensitive single tree growth modeling using a hierarchical Bayes approach and integrated nested Laplace approximations (INLA) for a distributed lag model. For ECOL MANAG. 2020;478:118497. [Google Scholar]

[CR37] 37.Kontis V, Bennett JE, Rashid T, Parks RM, Pearson-Stuttard J, Guillot M, Asaria P, Zhou B, Battaglini M, Corsetti G, et al. Magnitude, demographics and dynamics of the effect of the first wave of the COVID-19 pandemic on all-cause mortality in 21 industrialized countries. Nat Med. 2020;26(12):1919–28. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Shriner D, Yi N. Deviance information criterion (DIC) in bayesian multiple QTL mapping. COMPUT STAT DATA AN. 2009;53(5):1850–60. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Blangiardo M, Cameletti M, Baio G, Rue H. Spatial and spatio-temporal models with R-INLA. SPAT SPATIO-TEMPORAL. 2013;4:33–49. [DOI] [PubMed] [Google Scholar]

[CR40] 40.Hou K, Xu X. Ambient temperatures associated with reduced cognitive function in older adults in China. SCI REP-UK. 2023;13(1):17414. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR41] 41.Gómez-Rubio V, Rue H. Markov chain Monte Carlo with the integrated nested Laplace approximation. Stat Comput. 2018;28(5):1033–51. [Google Scholar]

[CR42] 42.Geng X, Zhang D, Li C, Yuan Y, Yu Z, Wang X. Impacts of Climatic zones on urban heat island: Spatiotemporal variations, trends, and drivers in China from 2001–2020. SUSTAIN CITIES SOC. 2023;89:104303. [Google Scholar]

[CR43] 43.Kan H, London SJ, Chen G, Zhang Y, Song G, Zhao N, Jiang L, Chen B. Season, sex, age, and education as modifiers of the effects of outdoor air pollution on daily mortality in shanghai, china: the public health and air pollution in Asia (PAPA) study. ENVIRON HEALTH PERSP. 2008;116(9):1183–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Lei J, Peng L, Yang T, Huang S, Zhu Y, Gao Y, Zhou L, Shi S, Liu C, Kan H, et al. Non-optimum ambient temperature May decrease pulmonary function: A longitudinal study with intensively repeated measurements among asthmatic adult patients in 25 Chinese cities. ENVIRON INT. 2022;164:107283. [DOI] [PubMed] [Google Scholar]

[CR45] 45.Fan P, Xue X, Hu J, Qiao Q, Yin T, Yang X, Chen X, Hou Y, Chen R. Ambient temperature and ambulatory blood pressure: an hourly–level, longitudinal panel study. SCI TOTAL ENVIRON. 2023;864:160854. [DOI] [PubMed] [Google Scholar]

[CR46] 46.Yang X, Zhou B, Xu Y, Han Z. CMIP6 evaluation and projection of temperature and precipitation over China. Adv Atmos Sci. 2021;38(5):817–30. [Google Scholar]

[CR47] 47.Ahmed K, Sachindra DA, Shahid S, Iqbal Z, Nawaz N, Khan N. Multi-model ensemble predictions of precipitation and temperature using machine learning algorithms. ATMOS RES. 2020;236:104806. [Google Scholar]

[CR48] 48.Chen Y, Guo F, Wang J, Cai W, Wang C, Wang K. Provincial and gridded population projection for China under shared socioeconomic pathways from 2010 to 2100. SCI DATA. 2020;7(1): 83. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Chen Y, Xu C, Ge Y, Zhang X, Zhou Y. A 100 m gridded population dataset of China’s seventh census using ensemble learning and big geospatial data. Earth Syst Sci Data. 2024;16(8):3705–18. [Google Scholar]

[CR50] 50.Barnett AG. Temperature and cardiovascular deaths in the US elderly: changes over time. Epidemiology. 2007;18(3):369. [DOI] [PubMed] [Google Scholar]

[CR51] 51.Chen M, Xian Y, Huang Y, Zhang X, Hu M, Guo S, Chen L, Liang L. Fine-scale population spatialization data of China in 2018 based on real location-based big data. SCI DATA. 2022;9(1): 624. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Renberg J, Lignier MJ, Wiggen ØN, Færevik H, Helgerud J, Sandsund M. Heat tolerance during uncompensable heat stress in men and women wearing firefighter personal protective equipment. APPL ERGON. 2022;101:103702. [DOI] [PubMed] [Google Scholar]

[CR53] 53.Hu K, Guo Y, Hochrainer-Stigler S, Liu W, See L, Yang X, Zhong J, Fei F, Chen F, Zhang Y, et al. Evidence for Urban–Rural disparity in Temperature–Mortality relationships in Zhejiang province, China. ENVIRON HEALTH PERSP. 2019;127(3):37001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Bilancia M, Demarinis G. Bayesian scanning of spatial disease rates with integrated nested Laplace approximation (INLA). Stat Methods Appl. 2014;23(1):71–94. [Google Scholar]

[CR55] 55.Xi D, Liu L, Zhang M, Huang C, Burkart KG, Ebi K, Zeng Y, Ji JS. Risk factors associated with heatwave mortality in Chinese adults over 65 years. Nat Med. 2024;30(5):1489–98. [DOI] [PubMed] [Google Scholar]

[CR56] 56.Wang S, Zhan W, Zhou B, Tong S, Chakraborty TC, Wang Z, Huang K, Du H, Middel A, Li J, et al. Dual impact of global urban overheating on mortality. NAT CLIM CHANGE. 2025;15(5):497–504. [Google Scholar]

PERMALINK

Inequality in mortality risk of older adults between rural and urban areas exposure to temperature variations in China

Kun Hou

Liqiang Zhang

Feng Yang

Wei Hu

Xia Xu

Xin Yao

Zhen Wang

Abstract

Background

Methods

Results

Conclusion

Supplementary Information

Introduction

Materials and methods

Temperature data

Air pollution data

Mortality of older adults

Methods

Bayesian spatio-temporal model

Displacement effects of temperature variation

Influence of extreme abnormal temperatures on mortality risk

Heterogeneity analysis of mortality risk in older adults

Effects of temperature under future climate scenarios

Sensitivity analysis

Results

Association of temperature variations with mortality risk

Fig. 1.

Fig. 2.

Quantitative effect of high and low temperature

Fig. 3.

Impacts of extreme temperature in climate zone

Heterogeneity of mortality risk in older adults

Impacts of future temperature variations

Fig. 4.

Fig. 5.

Sensitivity analysis

Discussion

Conclusion

Supplementary Information

Acknowledgements

Abbreviations

Authors’ contributions

Funding

Data availability

Declarations

Ethics approval and consent to participate

Consent for publication

Competing interests

Footnotes

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases