Skip to main content
NCBI home page
Sage Choice logoLink to Sage Choice
. 2024 Oct 23;59(1):18–28. doi: 10.1177/00048674241290449

Youth suicidality risk relative to ambient temperature and heatwaves across climate zones: A time series analysis of emergency department presentations in New South Wales, Australia

Cybele Dey 1,2,✉, Jianyun Wu 3, John Uesi 3, Grant Sara 2,3,4, Michael Dudley 2,5, Katherine Knight 1, James G Scott 6,7, Ollie Jay 8, Michael Bowden 9,10, Iain E Perkes 1,2,11
PMCID: PMC11667958  PMID: 39441101

Abstract

Objective:

Youth suicidality prevalence continues to rise alongside hot weather severity. Links between these two variables are underexplored. We examined associations between daily temperature and emergency department suicidality presentations by young people. We assessed these associations for five regions covering New South Wales as determined by ‘climate zone’ and analysed for heatwave effects as well as based on demographic subgroups.

Methods:

Daily emergency department presentations for suicidality by people aged 12–24 years across New South Wales, Australia, during warmer months (November to March) from 2012 to 2019 were examined in relation to daily mean temperature and heatwaves (⩾3 consecutive days ⩾ 95th percentile of long-term daily mean temperature) and by climate zone, using a generalised additive model with negative binomial distribution. Risks for age- and sex-based subgroups were also calculated.

Results:

New South Wales youth suicidality presentation rates were significantly higher on hotter days. For every 1°C rise above average daily mean temperature, youth suicidality presentations to New South Wales emergency departments increased by 1.3%. Heatwaves did not increase presentation rates beyond single-day daily mean temperature effects. These findings were predominantly replicated across climate zones and demographic subgroups, though the association between suicidality and ambient temperature was weaker in coastal regions including Eastern Sydney.

Conclusion:

There is a positive linear association between ambient temperature and youth suicidality presentations to emergency departments. Risks are increased on single hot days, not only during heatwaves. Public health, broader societal approaches to heat and health system planning should consider impacts on youth suicidality of predicted increases in hot weather severity and frequency.

Keywords: Climate change, hot temperature, adolescent, young adult, self-injurious behaviour

Introduction

Youth suicide and suicidality have increased in Australia in recent years. Presentations by people aged 10–24 years with suicidality, i.e. suicidal thoughts or behaviours, to emergency departments (EDs) across the Australian state of New South Wales (NSW) rose by 8.4% annually across the 5 years to 2020 (Sara et al., 2022). A sign of clinically significant psychological distress (RANZCP, 2020), suicidality is also a risk factor for death by suicide among young people (Brahmbhatt et al., 2019; Cha et al., 2018). Indeed, there has been an increase in the Australian suicide mortality rate of girls and women aged 10–24 years (Sara et al., 2022; Stefanac et al., 2019). Suicide is the leading cause of death among people aged 15–24 years in Australia (Australian Institute of Health and Welfare, 2022), and the second leading cause worldwide among people aged 15–19 years (Liu et al., 2022). This critical public health problem is incompletely accounted for by known risk factors such as bullying, mental illness and the COVID-19 pandemic (Holt et al., 2015; Perera et al., 2018; Sara et al., 2022; Swedo et al., 2021), and direct exposure to increased temperatures is one of multiple, potential contributing factors proposed (Sara et al., 2022).

The co-occurring increases in hot weather and youth suicidality warrant examination of links between these two variables (Liu et al., 2021; Nori-Sarma et al., 2022). Associations between hot weather and mental distress in adults, including ED presentations with suicidality, have been replicated by meta-analyses (Florido Ngu et al., 2021; Liu et al., 2021; Nori-Sarma et al., 2022). These studies implicate duration (heatwaves), ambient temperature, humidity, climate zones and demographic characteristics as candidate factors with respect to risk of mental health harms.

Health warning systems for hot weather typically focus on heatwaves (Williams et al., 2022). Prolonged exposure during heatwaves does indeed worsen health outcomes such as cardiovascular deaths (Liu et al., 2022). However, single hot days are also associated with increased psychiatric morbidity and mortality among adults (Liu et al., 2021; Nori-Sarma et al., 2022). The question of what, if any, additional suicidality risk occurs during heatwaves above that attributable to daily ambient temperature is unresolved (Liu et al., 2021; Nori-Sarma et al., 2022).

Risk of suicidality relative to hot weather among children (4–11 years), adolescents (12–17 years) and young adults (18–24 years) is similarly underexamined. Numerous studies have included young adults, though they have been grouped together with adults generally. Only three previous studies of suicidality and hot weather have included children or adolescents, all three found increases in suicidality in this cohort, albeit none of these studies distinguished adolescent from overall child and adolescent groups (Basu et al., 2018; Doganay et al., 2003; Kok and Tsoi, 1993). Among people aged 0–19 years living in Singapore, regression analysis revealed correlation between suicide deaths and a range of weather variables, with maximum temperature accounting for 58% of the variation in suicide deaths (Kok and Tsoi, 1993).

Relationships between weather and health across a large geographical expanse show regional differences that are accounted for in part by local climate (Dimitrova et al., 2022; Liu et al., 2021). Geographical climate zones are classified based on long-term seasonal patterns in temperature and humidity or rainfall which determine natural vegetation patterns, e.g. tropical, arid and temperate (Köppen, 1884). The Köppen–Geiger climate zone classification system is used by the World Meteorological Organisation (2024), adopted by national governments for environmental impact assessments and used to investigate the health impacts of climate change (Australian Government Bureau of Meteorology, 2023; Beck et al., 2023; Liu et al., 2023). For instance, a recent meta-analysis established that suicide risk relative to ambient temperature is greatest in tropical zones (Liu et al., 2021).

Here we investigate relationships between youth presentations to EDs with suicidality and ambient temperatures in NSW, a state with a population of 1.3 million young people living across 801,150 km2. We examine relationships between youth suicidality presentations to EDs for five climate zones covering NSW. We also aim to disambiguate suicidality risk of heatwaves from that attributable to temperature. We then apply this approach to distinct climate zones across the state of NSW (Australian Bureau of Statistics, 2022; Australian Government Geoscience Australia, 2004) before examining temperature–suicidality associations based on age group and sex.

Methods

Outcome measure: daily youth suicidality presentations

We examined presentations by people aged 12–24 years, between 1 January 2012 and 31 December 2019, to all hospital EDs in the state of NSW. This sample was chosen in order to study emergency department suicidality presentations across the whole population of young people aged 12–24 years in NSW. To study associations with exposure to higher daily temperatures and heatwaves, rather than the established seasonal increase in spring relative to winter (Doganay et al., 2003), we limited analyses to the warmer months, i.e. November to March. We used a method validated and implemented by NSW Health to identify suicidality presentations via either ICD-10 diagnosis codes (i.e. self-harm, suicidal ideation, drug overdose or poisoning) or ‘presenting problem’ free text (i.e. indicating self-harm or suicidal ideation) (Sara and Wu, 2023). We analysed ‘suicidality’ as a diagnostic grouping based on associations between ambient temperature and transdiagnostic emergency mental health presentation rates being observed among children and young people (Basu et al., 2018) and adults (Nori-Sarma et al., 2022).

Climate zones

We used the Köppen–Geiger climate zone system (Supplementary Material Table S1), as adapted by the Australian government and used for infrastructure and service planning (Figure 1) (Australian Building Codes Board, 2019). We aggregated ED presentations to the local government area (LGA) of the patient’s residential address, and then converted these and age-specific LGA population estimates to climate zone level estimates (Australian Bureau of Statistics, 2022).

Figure 1.

Figure 1.

Open in a new tab

Climate zones across New South Wales.

The Australian Government Bureau of Meteorology (2023) climate zones in New South Wales shown are: ‘hot dry summer, cool winter’; ‘warm humid summer, mild winter’; ‘warm temperate’; ‘mild temperate’; ‘cool temperate’ (Australian Building Codes Board, 2019). For internal consistency and concision, we removed winter descriptions from the first two zones, referring to ‘hot dry’ and ‘warm humid’ herewith.

Temperature and heatwaves

We used daily Australian Gridded Climate Data, which model daily minimum and maximum temperature at a resolution of 0.05°×0.05° in longitude and latitude (National Computational Infrastructure, 2021). We overlaid these grid coordinates of temperature onto LGA digital shapefiles, which we then aggregated into climate zones (National Computational Infrastructure, 2021). We aggregated daily temperatures at each grid point to produce a daily mean temperature (DMT) for each zone. We defined heatwaves as ⩾3 consecutive days of DMT’s distribution at or above the 95th percentile of DMT at the NSW state level and for each climate zone based on year-round data (Centre for Epidemiology and Evidence, 2021; Coates et al., 2022; Liu et al., 2021).

Statistical analysis

Our primary analysis was applied to all youth (12–24 years) presentations across the state of NSW. For dates between 1 November and 31 March in the study period, using the daily number of ED presentations as the dependent variable, we analysed the effect of DMT using a generalised additive model with a negative binomial distribution (Wood, 2017). The DMT effect was modelled using a cubic regression spline. Variations in ED presentations with adolescent suicidality occur across annual (Radhakrishnan et al., 2023; Sara et al., 2022; Stapelberg et al., 2020) seasonal, e.g. school terms (Lueck et al., 2015; Stapelberg et al., 2020), and within-week timescales (Sara et al., 2022). We used cubic regression splines to control for variations in daily ED presentations across the study period based on (1) annual trend from 2012 to 2019 financial years, due to the warmer months falling within the same financial year and (2) day-of-year (i.e. counting number of days from 1 November to 31 March within each annual cycle) using cubic regression splines and (3) day-of-week (Monday to Sunday) as a categorical variable. We also used these time-based control variables to generate a model which we refer to as the temporal model in order to compare goodness of fit (see Statistical analysis, paragraph 6). The DMT effect was also modelled using a cubic regression spline to capture potential non-linear associations between DMT and daily presentations. To test for additional heatwave effects in context of variation in previously published results (Liu et al., 2021), we added a binary variable to days that satisfied the heatwave definition above (Coates et al., 2022; Liu et al., 2021). In our primary analysis, we did not find significant heatwave effects additional to that of DMT (Table S2 in Supplementary Material). Therefore, we re-estimated the DMT effects without the heatwave variable.

Owing to variable impacts on health of ambient temperature in association with other meteorological characteristics (Liu et al., 2021, 2023) we investigated distinct regions of the state of NSW. In our secondary analysis, we modified the above models to include a second categorical variable representing climate zone level intercepts. The DMT and heatwave effects in each zone were estimated through interaction terms by applying a cubic regression spine to each climate zone, and an interaction between the climate zone and the heatwave variables, separately. The mathematical specifications of the models are provided in Supplementary Material (Model specifications).

Finally, we conducted subgroup analyses to estimate DMT and heatwave effects for each sex (Clemens et al., 2022; Mauvais-Jarvis et al., 2020) and age group (Vergunst and Berry, 2021) (12–17 and 18–24) at statewide and climate zone levels. Our adolescent age definition was consistent with the Australian national survey of child and adolescent mental health (Zubrick et al., 2016). Our young adult age was consistent with Australian government definitions of youth (Health and Social Policy Branch NSW Ministry of Health, 2017).

For the primary analysis, if the association between DMT and suicidality presentations appeared linear, we also provide a direct estimate as the relative change in the number of daily presentations, i.e. relative risk, and associated 95% confidence intervals (CIs) for each 1°C increase in DMT. For other analyses with DMT, we present the associations with DMT as smoothed curves showing the estimated relative risk and 95% CIs compared to the average DMT over the range of observed temperatures across the state of NSW and in each climate zone.

To aid interpretation of temperature effects, we report the estimated relative risks and 95% CIs at temperatures 5°C below and above the average DMT. The associations with heatwaves (Table S2 in Supplementary Material) were expressed as the estimated relative risk with 95% CI by comparing between heatwave days and non-heatwave days in NSW and in each zone. For ease of discussion and visualisation, the DMT effects from primary (statewide) and secondary (climate zone) analyses are presented in the same figure.

We conducted an analysis of deviance to compare models (1) with time-based control variables, i.e. ‘temporal’, (2) then adding weather variables, ‘temporal + DMT’ and (3) temporal + DMT + heatwave; for each analysis, we also used deviance explained as a goodness-of-fit statistic and Akaike information criterion (AIC). A model was considered statistically significantly improved when a p-value from the analysis of deviance was < 0.05.

We conducted analyses using R (version 4.1.2, R Core Team, 2021). We used mgcv (v1.8-38) package (Wood, 2017) for generalised additive models.

Results

Ahead of our primary analysis results, we first present data on the exposure and outcome. Statewide NSW temperatures increased across the study period from 2012 to 2019 (Figure 2). Across the warmer months (November to March) included in the analysis, the statewide average DMT was 21.9°C. Average DMTs were lowest in the cool temperate climate zone (18.7°C) and highest in the hot dry zone (24.4°C). The highest single DMT (35.3°C) and overall single point temperature (49.1°C) both occurred in the hot dry zone. Table 1 shows Australian Bureau of Meteorology climate zones and their corresponding LGAs in NSW, with population estimates and the 95th percentile threshold for DMT. Half of the NSW youth population resided in the warm temperate zone.

Figure 2.

Figure 2.

Open in a new tab

New South Wales statewide daily mean temperature throughout the year (2012–2019).

Each concentric ring represents a year, starting with 2012. The grey ‘cut out’ section represents the warmer months from 1 November to 30 March.

Table 1.

Population, suicidality presentations, daily temperatures and heatwaves for New South Wales and by climate zone.

Climate zones Population aged 12–24 (% total) Suicidality presentations (rate per 100,000) Average DMT °C (min–max) Heatwave, threshold Heatwave, total days
Statewide 1,316,302 4221 21.9 (14.0–31.0) 25.2°C 94
Hot dry 131,630 (10%) 4519 24.4 (13.7–35.3) 29.1°C 103
Warm humid 65,815 (5%) 7046 23.1 (17.2–31.5) 25.7 °C 53
Warm temperate 658,151 (50%) 3968 22.1 (15.5–31.8) 25.3°C 45
Mild temperate 408,054 (31%) 4008 21.3 (13.7–30.6) 24.7°C 72
Cool temperate 52,652 (4%) 4779 18.7 (8.1–29.3) 23.1°C 105
Open in a new tab

New South Wales population aged 12–24 years is for 2019 (Centre for Epidemiology and Evidence, 2021). Results are for warmer months (1 November to 30 March) through the study period of January 2012 to December 2019. Emergency department presentation numbers and average daily mean temperatures are for warmer months only.

We identified 132,296 suicidality presentations by youth from 2012 to 2019, of which 55,566 occurred during the five warmer months and were included in the analysis (Table 1). Nearly two-thirds of presentations (64.4%) were by females aged 12–17 years. Approximately 56.9% of total presentations occurred in the afternoon and evening between 12 pm and 10 pm, Table S3 shows the 24-hour distribution of ED presentations (Supplementary Materials).

Analysis of deviance showed that adding DMT to the temporal model improved model performance in primary, i.e. across the state of NSW (Akaike information criterion: 8350 vs 8378, p < 0.0001), secondary, i.e. climate zones (Akaike information criterion: 28517 vs 28574, p < 0.0001) and subgroup (Tables S4 and S5 in Supplementary Materials) analyses. Heatwaves had no significant added effect above that of DMT with p > 0.05 in all analyses of deviance when comparing between models with DMT alone and with both DMT and heatwave variables (full results in Supplementary Material, Tables S4 and S5); therefore, we present results of ‘temporal + DMT’ model in the manuscript, i.e. without the heatwave variable.

Across the state of NSW, there was a linear positive association between DMT and suicidality presentations by youth to EDs (Figure 3). At the NSW average DMT of 21.9°C for warmer months, there was an average of 45.7 youth suicidality presentations per day. For every 1°C increase in DMT, presentations increased by 1.3% (relative risk 1.01, 95% CI = 1.01–1.02). For instance, on days with a DMT of 26.9°C, the relative risk was 1.065 (95% CI = 1.04–1.09), whereas on days with a DMT of 16.9°C, the relative risk was 0.94 (95% CI = 0.91–0.96).

Figure 3.

Figure 3.

Open in a new tab

Risk of youth suicidality relative to temperature across New South Wales and climate zones.

Relative risk is equal to 1.0 at the average DMT across warmer months (November to March) for the study period (January 2012 to December 2019) in the state of New South Wales (21.9°C) and each climate zone, with 95% CIs indicated by shading.

Across climate zones, increases in DMT were predominantly associated with increased risk of suicidality presentations (Figure 3). A linear relationship (as seen at a statewide level) was evident in the hot dry, mild temperate and cool temperate zones. The warm humid zone and warm temperate zone did not show this clear linear association. In the warm temperate zone, the relative risk of youth ED suicidality presentations was 0.93 (95% CI = 0.89–0.97) at 17.1°C and 1.0 (95% CI = 0.96–1.04) at 27.1°C. There was substantial variance in the data for the warm humid zone.

Demographic subgroup analysis

Increased risk of suicidality presentations with higher DMT was replicated among all demographic subgroups across four of the five climate zones (Figure 4).

Figure 4.

Figure 4.

Open in a new tab

Risk of adolescent and young adult suicidality relative to ambient temperature across New South Wales and by climate zone.

Relative risk is equal to 1.0 at the average DMT across warmer months (November to March) for the study period (January 2012 to December 2019) in the state of NSW (21.9°C) and each climate zone, with 95% CIs indicated by shading.

Among adolescents, there was a positive linear relationship between DMT and suicidality presentations across the state of NSW (Figure 4). This pattern was also seen among adolescents in the mild temperate zone where a DMT of 16.3°C was associated with a relative risk of 0.91 (95% CI = 0.86–0.95) which increased to 1.105 (95% CI = 1.06–1.16) in association with a DMT of 26.3°C. Results for adolescents had wide confidence intervals and non-significance in climate zones other than the mild temperate.

Among young adults, statewide results showed increasing DMT was associated with ED suicidality presentations, although with wide confidence intervals at higher DMTs; this association was observed in all climate zones, apart from the warm temperate (Figure 4).

For female youth, statewide and warm humid zone findings replicated the overall youth result of a positive, linear relationship between DMT and suicidality (Figure 5). For instance, in warm humid zone with a DMT of 18.1°C the relative risk was 0.82 (95% CI = 0.75–0.90), whereas on days with a DMT of 28.1°C, the relative risk was 1.22 (95% CI = 1.12–1.33). In contrast, a non-linear relationship between DMT and female youth suicidality risk was seen in the hot dry zone albeit with greater variance in the data.

Figure 5.

Figure 5.

Open in a new tab

Risk of female and male youth suicidality relative to ambient temperature across New South Wales and by climate zone.

Relative risk compared to the average DMT across warmer months (November to March) for the study period in New South Wales and in each climate zone, with 95% CIs indicated by shading.

Among male youth, statewide results showed increases in suicidality ED presentations above average DMT and then plateaued, although CIs were wide at higher DMTs (Figure 5). In the hot dry zone, male youth suicidality risk increased linearly with DMT. For instance, a DMT of 19.4 °C was associated with a relative risk of 0.91 (95% CI = 0.86–0.96), whereas this increased to 1.10 (95% CI = 1.04–1.17) in association with a DMT of 29.4°C.

In the warm temperate zone, no demographic subgroup showed statistically significantly increased risk of suicidality presentations to ED with higher DMT (Figure 5).

Discussion

We revealed an association between ambient temperature and emergency presentation by youth with suicidality, replicated across most climate zones, age groups and sexes. Across the state of NSW, for each 1°C rise above the average DMT of 21.9°C, emergency presentations rose by 1.3%. For instance, on days with a 24-hour mean temperature of 30°C, presentation rates were 11% greater than rates on days with a mean temperature of 21.9°C. This linear relationship was replicated across hot dry, mild temperate and cool temperate zones, whereas the warm humid and warm temperate zones showed variance. Heatwaves showed no significant effect above that of ambient temperature.

To our knowledge, ours is the first published study to report associations between ambient temperature and suicidality ED presentations specifically in adolescents, youth or young adults. In addition, this is the first study of these age groups using local climate zones. We also show that these relationships between in youth ED suicidality presentation and temperature are not accounted for by known associations with day-of-week or school term (Sara et al., 2022).

We established that the positive, linear association between ambient temperature and ED suicidality presentations among adults (Basu et al., 2018) also occurs among young people with suicidality. This ‘dose-response’ type of relationship between health and local environment is established in other fields of medicine. Altitude sickness, for instance, is a life-threatening illness caused by rapid change in oxygen pressure (McKenna and Martin, 2016). High ambient temperature is similarly linked to serious medical illness and death (Ebi et al., 2021).

As with altitude acclimatisation, gradual exposure to higher ambient temperatures does lead to physiological change with improved heat stress tolerance (Ebi et al., 2021). The extent to which heat acclimatisation has a practical impact in everyday life, as opposed to controlled lab settings, is not yet established and, as with altitude, there are finite physiological limits to heat acclimatisation (Ebi et al., 2021; McKenna and Martin, 2016). Where acclimatisation is insufficient, behavioural adaptation, such as accessing cool shelter, becomes crucial. Indeed, capacity to reduce exposure to heat is linked to morbidity and mortality during heat extremes (Ebi et al., 2021).

In the current study, adolescent girls and young women showed linear association between DMT and suicidality in all zones except the hot dry. In contrast, among males, the linear association was clear in the hot dry zone with greater uncertainty in other zones. We did not compare rates between males and females. Nonetheless, emergency suicidality presentations by youth are predominantly accounted for by female adolescents, future research could examine risk during hot weather relative to risk and exposure factors which may vary between sexes. For instance, differences between sexes could reflect biological dispositions (Giersch et al., 2022; Mauvais-Jarvis et al., 2020), occupational exposure, e.g. outdoor military training (Périard et al., 2022) or socio-cultural factors known to increase with hot weather such as unpaid carer responsibilities and gendered violence (Clemens et al., 2022; Mauvais-Jarvis et al., 2020).

Across the whole youth cohort, risk of suicidality relative to ambient temperature was maintained across climate zones except for the warm temperate zone. The warm temperate area in the state of NSW runs along the coastal region and includes Eastern Sydney and is predominantly socio-economically advantaged (Australian Building Codes Board, 2019; Australian Bureau of Statistics, 2021). Socio-economic advantage is associated with access to better-quality housing and air-conditioning (Jay et al., 2021).

In contrast to the warm temperate zone (which did not show increases in suicidality with ambient temperature), risk for youth living in the adjacent mild temperate zone did increase linearly with DMT. The mild temperate zone in NSW is an outer metropolitan, urban region including large areas of housing with poor thermal performance and little green infrastructure; both of these features are associated with an urban heat island effect and higher daily temperatures (Australian Building Codes Board, 2019; Jay et al., 2021; Santamouris et al., 2017). This part of Sydney is home to over half a million people living with the greatest relative socio-economic disadvantage, a factor which is associated with lower air conditioner ownership and usage (Australian Bureau of Statistics, 2021; Jay et al., 2021).

Access to air-conditioning and better-quality housing enables behavioural adaptation to reduce exposure to heat stress (Jay et al., 2021). Thus, although links were not evident between suicidal behaviour and socio-demographic factors for adolescents in Australia in the National Mental Health survey (Zubrick et al., 2016), youth suicidal behaviour in association with hot weather may be linked with socio-demographic factors that are, in turn, linked with higher exposure to heat. Mechanisms of increased heat exposure related to socio-demographic factors include urban heat island effect and reduced access to behavioural adaptation to heat stress (Ebi et al., 2021).

While demographic subgroups showed generally consistent increases in risk with higher ambient temperature, males in the hot dry zone were the only group to clearly mirror the linear risk seen in the hot dry zone overall youth result. It may be that male youth are at increased risk in the hot dry zone through greater exposure to physical outdoor work or sport despite hot weather. Supporting this interpretation, the young adult cohort in our study showed elevated risk at higher temperatures, whereas the adolescent group did not. Indeed, male workers and younger workers are known to have increased mental health risk relative to ambient temperature (Périard et al., 2022).

Implications

With ongoing increases in hot weather predicted for coming decades (CSIRO and Australian Bureau of Meteorology, 2022), health services must prepare for potential increasing youth suicidality presentations. Current public health heat planning and warnings typically relate to heatwaves rather than single hot days, they also typically omit a mental health component (Williams et al., 2022). Our findings that youth suicidality increases on single hot days extend previous research showing increases in ED presentations across diagnostic and presentation categories for adult and all-age groups (Basu et al., 2018; Nori-Sarma et al., 2022). In addition, youth suicidality ED presentations are linked to subsequent increased risk of young people dying by suicide (Cha et al., 2018). We think it appropriate that health warnings and secondary prevention efforts related to hot weather are reviewed considering these NSW-specific findings.

The implications of increasing frequency and severity of hot weather on youth suicidality are substantial (Florido Ngu et al., 2021). Globally, most youth live in low-income and middle-income countries (LMIC), often with insufficient public health data infrastructure to examine interactions between hot weather and young people’s mental health at a population level to inform policy (Clemens et al., 2022). Our access to population-wide public health data across multiple climate zones enabled us to explore this on a scale that may assist public mental health planning in countries where resources to conduct local research are not available. The generalisability of our findings is aided by the representation of four of the six major international Köppen–Geiger zones across the region we studied, i.e. the Australian state of NSW. While the observed link between youth suicidality and exposure to higher daily temperature may be relevant in the many parts of the world with these Köppen–Geiger zones, the policy responses would be expected to vary, including between LMICs and high-income countries (Arpin et al., 2021; Kruk et al., 2022). The availability of evidence-based guidance on ways to stay cool, even in resource-poor settings (Jay et al., 2021), could be relevant globally. Nonetheless, quality housing, infrastructure and green space that mitigate exposure to heat are often unavailable in the countries (Ebi et al., 2021) in which most young people live (Clemens et al., 2022).

Limitations

We tested the relationship between two daily averages (ED presentations, daily temperatures). These daily averages necessarily simplify more complex within-day variations in the timing of ED presentations or temperature changes. Most ED presentations will not occur at the daily average temperature, but at higher or lower temperatures during the same day. ED presentations may follow hours or days of distress or self-harm. Therefore, for some presentations in our data, putative heat-related distress or self-harm may occur on the day before ED presentation. For other presentations, distress and ED presentation may occur early in the day, prior to a later onset of high temperatures. These issues may lead to imprecision in our estimates, potentially obscuring or weakening any possible association between self-harm presentations and temperature. We believe that it is unlikely that these issue lead to systematic bias or overestimation because (1) variations and imprecision appear equally likely in both directions, (2) our analysis of heat wave effects considered multiple hot days, where these possible lead and lag effects are reduced, with no significant difference in findings and (3) most ED presentations in this age group occur in the afternoon and evening (Table S3 in Supplementary Materials shows the distribution of ED presentations over the 24-hour cycle).

Although our main finding is consistent with a biological contribution, our study was not designed to investigate mediating pathways between ambient temperature and suicidality. Nonetheless, a biological factor could increase vulnerability in those already at higher risk.

While we did not explicitly control for school days, associated variance was controlled for via day-of-week and day-of-year, given that school days predominantly fall on the same days of the week and days of the year across different weeks and years within the study period. Although youth suicidality could also be affected by heat exposure on days prior to the presentation, we did not investigate lagged effects of DMT and heatwaves as previous research suggested risk increases are greatest on the current day, approaching no increase within 48 hours (Kim et al., 2019). We do not know whether the same pattern is seen in other presentation types, e.g. primary care.

Green space in the local neighbourhood has been found to be protective against externalising behaviours linked to higher long-term ambient temperatures (Younan et al., 2018) and childhood residential green space is associated with lower risk of psychiatric disorders in adolescence (Engemann et al., 2019). However, we did not have access to green space and housing quality data of sufficient quality to account for these variables in analysis. Substance abuse is associated with mental health emergency presentations during hot weather and with worse outcomes (Liu et al., 2021); however, we did not have the data to examine rates of substance abuse, nor the substance/s used in the current presentation among youth presenting to ED with suicidality.

Future research directions

Climate zone analysis could further clarify connections between risks to mental health and physical health relative to heat exposure. The lack of additional heatwave effect suggests a different mechanism for increases in ED presentations for suicidality compared with other heat-related illnesses which show significant heatwave effects such as heat exhaustion and cardiovascular collapse (Joe et al., 2016).

Examining whether relative socio-economic disadvantage is associated with greater risk of youth suicidal behaviour during hot weather is an important question, given the potential for targeted public health measures to improve access for this group to cool shelter during hot weather.

Further desiderata include considering all mental health presentations, i.e. in addition to suicidality, given the recent, large North American study linking heat to higher mental health ED presentations (Nori-Sarma et al., 2022). Further exploration of biological mechanisms underpinning the role of hot weather in generating mental distress and the role of substance abuse during hot weather could provide valuable insights to inform effective interventions (Rony and Alamgir, 2023).

Conclusion

Ambient temperature is associated with increased risk of emergency presentations by youth with suicidality. Hot days, whether single event or within a heatwave, predict risk of youth suicidality. Heat effects were prominent in outer metropolitan urban and regional or rural areas, many of which have greater exposure to higher ambient temperature in our study, and reduced access to better-quality housing, green and blue infrastructure (Ebi et al., 2021). Use of climate zones alongside data on housing quality and green space may identify regions with different levels of risk and opportunities for prevention of and service planning for youth suicidality during hot weather.

Behavioural adaptation enabled by better-quality housing, air-conditioning and green space may mitigate increased risk of youth suicidality during hot weather. Consideration of potential mental health benefits for youth and single hot days may enhance public health approaches to hot weather by extending their focus beyond heatwaves, older populations and physical health.

Supplemental Material

sj-docx-1-anp-10.1177_00048674241290449 – Supplemental material for Youth suicidality risk relative to ambient temperature and heatwaves across climate zones: A time series analysis of emergency department presentations in New South Wales, Australia
sj-docx-1-anp-10.1177_00048674241290449.docx (55.1KB, docx)

Supplemental material, sj-docx-1-anp-10.1177_00048674241290449 for Youth suicidality risk relative to ambient temperature and heatwaves across climate zones: A time series analysis of emergency department presentations in New South Wales, Australia by Cybele Dey, Jianyun Wu, John Uesi, Grant Sara, Michael Dudley, Katherine Knight, James G Scott, Ollie Jay, Michael Bowden and Iain E Perkes in Australian & New Zealand Journal of Psychiatry

Footnotes

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article: this research was conducted within existing resources of the authors’ affiliated institutions..

Funding: The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This research was conducted within existing resources of the authors’ affiliated institutions.

Data Availability Statement: Health data in this study were extracted by the NSW Ministry of Health, the department of state government responsible for the planning, funding, management and oversight of public hospitals, EDs and community mental health services. These were routinely collected and aggregated administrative data, used with permission from the data custodian. Data on NSW ED presentations are available at https://www.healthstats.nsw.gov.au/#/r/103071. Climate data were sourced from publicly available Australian Bureau of Meteorology datasets https://geonetwork.nci.org.au/.

Supplemental Material: Supplemental material for this article is available online.

References

  1. Arpin E, Gauffin K, Kerr M, et al. (2021) Climate change and child health inequality a review of reviews. International Journal of Environmental Research and Public Health 18: 10896. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Australian Building Codes Board (2019) Climate zone map: New South Wales and Australian Capital Territory. In: Board NCCABC (ed.) Climate Zone Map (VC00031.3). Canberra, ACT, Australia: National Construction Code Australian Building Codes Board, pp. 1–6. [Google Scholar]
  3. Australian Bureau of Statistics (2021) Index of Relative Socio-Economic Disadvantage’ Census of Population and Housing, SEIFA. Canberra, ACT, Australia: Australian Bureau of Statistics. [Google Scholar]
  4. Australian Bureau of Statistics (2022) Estimated Resident Population by Local Government Area (2022), Age and Sex, 2001 to 2022. Canberra, ACT, Australia: Australian Bureau of Statistics. [Google Scholar]
  5. Australian Government Bureau of Meteorology (2023) Climate classification maps. Available at: www.bom.gov.au/climate/maps/averages/climate-classification/ (accessed 15 December 2023).
  6. Australian Government Geoscience Australia (2004) Area of Australia – States and Territories. Canberra, ACT, Australia: Australian Government Geoscience Australia. [Google Scholar]
  7. Australian Institute of Health and Welfare (2022) Deaths in Australia. Canberra, ACT, Australia: Australian Institute of Health and Welfare. Available at: https://www.aihw.gov.au/reports/life-expectancy-deaths/deaths-in-australia/contents/leading-causes-of-death (accessed 16 October 2024). [Google Scholar]
  8. Basu R, Gavin L, Pearson D, et al. (2018) Examining the association between apparent temperature and mental health-related emergency room visits in California. American Journal of Epidemiology 187: 726–735. [DOI] [PubMed] [Google Scholar]
  9. Beck HE, McVicar TR, Vergopolan N, et al. (2023) High-resolution (1 km) Köppen-Geiger maps for 1901–2099 based on constrained CMIP6 projections. Scientific Data 10: 724–724. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Brahmbhatt K, Kurtz BP, Afzal KI, et al. (2019) Suicide risk screening in pediatric hospitals: Clinical pathways to address a global health crisis. Psychosomatics 60: 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Centre for Epidemiology and Evidence (2021) Population estimates NSW. Available at: www.healthstats.nsw.gov.au/#/r/107309 (accessed 21 December 2022).
  12. Cha CB, Franz PJ, MGuzmán E, et al. (2018) Annual research review: Suicide among youth – epidemiology, (potential) etiology, and treatment. Journal of Child Psychology and Psychiatry and Allied Disciplines 59: 460–482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Clemens V, Von Hirschhausen E, Fegert JM. (2022) Report of the intergovernmental panel on climate change: Implications for the mental health policy of children and adolescents in Europe-a scoping review. European Child & Adolescent Psychiatry 31: 701–713. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Coates L, Van Leeuwen J, Browning S, et al. (2022) Heatwave fatalities in Australia, 2001–2018: An analysis of coronial records. International Journal of Disaster Risk Reduction 67: 102671. [Google Scholar]
  15. CSIRO and Australian Bureau of Meteorology (2022) State of the Climate 2022. Canberra, ACT, Australia: CSIRO and Australian Bureau of Meteorology. [Google Scholar]
  16. Dimitrova A, McElroy S, Levy M, et al. (2022) Precipitation variability and risk of infectious disease in children under 5 years for 32 countries: A global analysis using demographic and health survey data. The Lancet Planetary Health 6: e147–e155. [DOI] [PubMed] [Google Scholar]
  17. Doganay Z, Sunter AT, Guz H, et al. (2003) Climate and diurnal variation in suicide attempts in the ED. American Journal of Emergency Medicine 21: 271–275. [DOI] [PubMed] [Google Scholar]
  18. Ebi KL, Capon A, Berry P, et al. (2021) Hot weather and heat extremes: Health risks. The Lancet 398: 698–708. [DOI] [PubMed] [Google Scholar]
  19. Engemann K, Pedersen CB, Arge L, et al. (2019) Residential green space in childhood is associated with lower risk of psychiatric disorders from adolescence into adulthood. Proceedings of the National Academy of Sciences 116: 5188–5193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Florido Ngu F, Kelman I, Chambers J, et al. (2021) Correlating heatwaves and relative humidity with suicide (fatal intentional self-harm). Scientific Reports 11: 22175. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Giersch GEW, Garcia CK, Stachenfeld NS, et al. (2022) Are there sex differences in risk for exertional heat stroke? A translational approach. Experimental Physiology 107: 1136–1143. [DOI] [PubMed] [Google Scholar]
  22. Health and Social Policy Branch NSW Ministry of Health (2017) NSW Youth Health Framework 2017–2024. St Leonards, NSW, Australia: Health and Social Policy Branch NSW Ministry of Health. [Google Scholar]
  23. Holt MK, Vivolo-Kantor AM, Polanin JR, et al. (2015) Bullying and suicidal ideation and behaviors: A meta-analysis. Pediatrics 135: e496–e509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jay O, Capon A, Berry P, et al. (2021) Reducing the health effects of hot weather and heat extremes: From personal cooling strategies to green cities. The Lancet 398: 709–724. [DOI] [PubMed] [Google Scholar]
  25. Joe L, Hoshiko S, Dobraca D, et al. (2016) Mortality during a large-scale heat wave by place, demographic group, internal and external causes of death, and building climate zone. International Journal of Environmental Research and Public Health 13: 299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Kim Y, Kim H, Gasparrini A, et al. (2019) Suicide and ambient temperature: A multi-country multi-city study. Environmental Health Perspectives 127: 117007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kok LP, Tsoi WF. (1993) Season, climate and suicide in Singapore. Medicine, Science and the Law 33: 247–252. [DOI] [PubMed] [Google Scholar]
  28. Köppen W. (1884) Die Wärmezonen der Erde, nach der Dauer der heissen, gemässigten und kalten Zeit und nach der Wirkung der Wärme auf die organische Welt betrachtet. Meteorologische Zeitschrift 1: 5–226. [Google Scholar]
  29. Kruk ME, Lewis TP, Arsenault C, et al. (2022) Improving health and social systems for all children in LMICs: Structural innovations to deliver high-quality services. The Lancet 399: 1830–1844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Liu J, Hansen A, Varghese BM, et al. (2023) Estimating the burden of disease attributable to high ambient temperature across climate zones: Methodological framework with a case study. International Journal of Epidemiology 52: 783–795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Liu J, Varghese BM, Hansen A, et al. (2021) Is there an association between hot weather and poor mental health outcomes? A systematic review and meta-analysis. Environment International 153: 106533. [DOI] [PubMed] [Google Scholar]
  32. Liu L, Villavicencio F, Yeung D, et al. (2022) National, regional, and global causes of mortality in 5-19-year-olds from 2000 to 2019: A systematic analysis. The Lancet Global Health 10: e337–e347. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lueck C, Kearl L, Lam CN, et al. (2015) Do emergency pediatric psychiatric visits for danger to self or others correspond to times of school attendance? The American Journal of Emergency Medicine 33: 682–684. [DOI] [PubMed] [Google Scholar]
  34. Mauvais-Jarvis F, Bairey Merz N, Barnes PJ, et al. (2020) Sex and gender: Modifiers of health, disease, and medicine. The Lancet 396: 565–582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. McKenna H, Martin D. (2016) Surviving physiological stress: Can insights into human adaptation to austere environments be applied to the critical care unit? Trends in Anaesthesia and Critical Care 11: 6–13. [Google Scholar]
  36. National Computational Infrastructure DC (2021) NCI Data Catalogue. Acton, ACT, Australia: NCI Data Collections and Publishing. Available at: https://geonetwork.nci.org.au/ (accessed 1 December 2021). [Google Scholar]
  37. Nori-Sarma A, Sun S, Sun Y, et al. (2022) Association between ambient heat and risk of emergency department visits for mental health among US adults, 2010 to 2019. JAMA Psychiatry 79: 341–349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Perera J, Wand T, Bein KJ, et al. (2018) Presentations to NSW emergency departments with self-harm, suicidal ideation, or intentional poisoning, 2010–2014. Medical Journal of Australia 208: 348–353. [DOI] [PubMed] [Google Scholar]
  39. Périard JD, DeGroot D, Jay O. (2022) Exertional heat stroke in sport and the military: Epidemiology and mitigation. Experimental Physiology 107: 1111–1121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. R Core Team (2021) R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. https://www.R-project.org/ (accessed 22 January 2024). [Google Scholar]
  41. Radhakrishnan L, Carey K, Pell D, et al. (2023) Seasonal trends in emergency department visits for mental and behavioral health conditions among children and adolescents aged 5–17 Years – United States, January 2018–June 2023. MMWR. Morbidity and Mortality Weekly Report 72: 1032–1040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Rony MKK, Alamgir HM. (2023) High temperatures on mental health: Recognizing the association and the need for proactive strategies – A perspective. Health Science Reports 6: e1729. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Royal Australian and New Zealand College of Psychiatrists (RANZCP) (2020) Suicide prevention – the role of psychiatry. Position Statement #101. Available at: https://www.ranzcp.org/clinical-guidelines-publications/clinical-guidelines-publications-library/suicide-prevention-the-role-of-psychiatry (accessed 15 January 2024).
  44. Santamouris M, Haddad S, Fiorito F, et al. (2017) Urban heat island and overheating characteristics in Sydney, Australia. An Analysis of multiyear measurements. Sustainability 9: 712. [Google Scholar]
  45. Sara G, Wu J, Uesi J, et al. (2022) Growth in emergency department self-harm or suicidal ideation presentations in young people: Comparing trends before and since the COVID-19 first wave in New South Wales, Australia. Australian and New Zealand Journal of Psychiatry 57: 58–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Sara GE, Wu J. (2023) Enhanced self-harm presentation reporting using additional ICD-10 codes and free text in NSW emergency departments. Public Health Research and Practice 33: 33012303. [DOI] [PubMed] [Google Scholar]
  47. Stapelberg NJC, Sveticic J, Hughes I, et al. (2020) Suicidal presentations to emergency departments in a large Australian public health service over 10 years. International Journal of Environmental Research and Public Health 17: 5920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Stefanac N, Hetrick S, Hulbert C, et al. (2019) Are young female suicides increasing? A comparison of sex-specific rates and characteristics of youth suicides in Australia over 2004–2014. BMC Public Health 19: 1389. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Swedo EA, Beauregard JL, de Fijter S, et al. (2021) Associations between social media and suicidal behaviors during a youth suicide cluster in Ohio. The Journal of Adolescent Health 68: 308–316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Vergunst F, Berry H. (2021) Mental health and climate change – A developmental life course perspective. European Psychiatry 64: S231–S232. [Google Scholar]
  51. Williams S, Nitschke M, Wondmagegn BY, et al. (2022) Evaluating cost benefits from a heat health warning system in Adelaide, South Australia. Australian and New Zealand Journal of Public Health 46: 149–154. [DOI] [PubMed] [Google Scholar]
  52. Wood SN. (2017) Generalized Additive Models. New York: Chapman and Hall/CRC. [Google Scholar]
  53. World Meteorological Organisation (2024) International Meteorological Vocabulary 182. Available at: https://unterm.un.org/unterm2/en/view/b8f1f401-463d-425d-baa7-7e3b7b913ebd (accessed 24 January 2024).
  54. Younan D, Li L, Tuvblad C, et al. (2018) Long-term ambient temperature and externalizing behaviors in adolescents. American Journal of Epidemiology 187: 1931–1941. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Zubrick SR, Hafekost J, Johnson SE, et al. (2016) Suicidal behaviours: Prevalence estimates from the second Australian child and adolescent survey of mental health and wellbeing. Australian and New Zealand Journal of Psychiatry 50: 899–910. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

sj-docx-1-anp-10.1177_00048674241290449 – Supplemental material for Youth suicidality risk relative to ambient temperature and heatwaves across climate zones: A time series analysis of emergency department presentations in New South Wales, Australia
sj-docx-1-anp-10.1177_00048674241290449.docx (55.1KB, docx)

Supplemental material, sj-docx-1-anp-10.1177_00048674241290449 for Youth suicidality risk relative to ambient temperature and heatwaves across climate zones: A time series analysis of emergency department presentations in New South Wales, Australia by Cybele Dey, Jianyun Wu, John Uesi, Grant Sara, Michael Dudley, Katherine Knight, James G Scott, Ollie Jay, Michael Bowden and Iain E Perkes in Australian & New Zealand Journal of Psychiatry

Articles from The Australian and New Zealand Journal of Psychiatry are provided here courtesy of SAGE Publications

RESOURCES