Suicide and Ambient Temperature: A Multi-Country Multi-City Study

Background: Previous literature suggests that higher ambient temperature may play a role in increasing the risk of suicide. However, no multi-country study has explored the shape of the association and the role of moderate and extreme heat across different locations. Objectives: We examined the short-term temperature–suicide relationship using daily time-series data collected for 341 locations in 12 countries for periods ranging from 4 to 40 y. Methods: We conducted a two-stage meta-analysis. First, we performed location-specific time-stratified case-crossover analyses to examine the temperature–suicide association for each location. Then, we used a multivariate meta-regression to combine the location-specific lag-cumulative nonlinear associations across all locations and by country. Results: A total of 1,320,148 suicides were included in this study. Higher ambient temperature was associated with an increased risk of suicide in general, and we observed a nonlinear association (inverted J-shaped curve) with the highest risk at 27°C. The relative risk (RR) for the highest risk was 1.33 (95% CI: 1.30, 1.36) compared with the risk at the first percentile. Country-specific results showed that the nonlinear associations were more obvious in northeast Asia (Japan, South Korea, and Taiwan). The temperature with the highest risk of suicide ranged from the 87th to 88th percentiles in the northeast Asian countries, whereas this value was the 99th percentile in Western countries (Canada, Spain, Switzerland, the UK, and the United States) and South Africa, where nearly linear associations were estimated. The country-specific RRs ranged from 1.31 (95% CI: 1.19, 1.44) in the United States to 1.65 (95% CI: 1.40, 1.93) in Taiwan, excluding countries where the results were substantially uncertain. Discussion: Our findings showed that the risk of suicide increased with increasing ambient temperature in many countries, but to varying extents and not necessarily linearly. This temperature–suicide association should be interpreted cautiously, and further evidence of the relationship and modifying factors is needed. https://doi.org/10.1289/EHP4898


Table of Contents
Data collection Table S1. Summary statistics of other weather variables and additional information by country. Table S2. Location-specific total number of suicide, average temperature (°C), maximum suicide temperature (°C) (MaxST) and lag-cumulative relative risk (RR) for the MaxST vs. MinST, estimated by using a conditional Poisson model adjusting for seasonality, long-term time trend, and the day-of-week from the first-stage modeling. Table S3. P-values of the LRT (log-likelihood ratio test) between linear and nonlinear models with different number of internal knots of the distributed lag nonlinear function for temperature. Table S4. Random-effects meta-regression models in the second-stage analysis. Table S5. Country-specific minimum suicide temperature (°C) (MinST) and maximum suicide temperature (°C) (MaxST) and pooled lag-cumulative relative risks (RRs) with 95% confidence intervals (CIs) by country in subgroup analyses by sex and age groups (young, < 65 years of age; elder, ≥ 65 years of age) using a conditional Poisson model adjusting for seasonality, long-term time trend, and the day-of-week. Figure S1. Quasi-Akaike Information Criteria (AIC modified with overdispersion) between linear and nonlinear models using a quadratic B-spline with one to three internal knots for the temperature-suicide association (A) pooled across all locations, locations in western countries (Brazil, Canada, South Africa, Spain, Switzerland, UK, and US), and locations in eastern countries (Japan, South Korea, Philippines, Taiwan, and Vietnam), (B) pooled for each country, and (C) in capital cities/regions where the number of suicides was largest for each country. BR, Brazil; CA, Canada; JP, Japan; KR, South Korea; PH, the Philippines; SA, South Africa; SP, Spain; SW, Switzerland; TW, Taiwan; UK, the UK; US, the United States; VN, Vietnam. Figure S2. Country-specific lag-response associations for maximum suicide temperature (MaxST) vs. minimum suicide temperature (MinST) over the extended lags of up to 6 days (with 95% confidence interval, shaded grey) using a conditional Poisson model adjusting for seasonality, long-term time trend, and the day-of-week. RR, relative risk. Figure S3. Location-specific lag-cumulative temperature-suicide associations as the best linear unbiased prediction (BLUP) (with 95% CIs, shaded grey) and the corresponding temperature distributions. The temperature-suicide associations were estimated by using a conditional Poisson model adjusting for seasonality, long-term time trend, and the day of week. The blue and red vertical dotted lines indicate the minimum suicide temperature and maximum suicide temperature, respectively. CI, confidence interval; RR, relative risk. Figure S4. Country-specific pooled lag-response associations for the maximum suicide temperature (MaxST) vs. the minimum suicide temperature (MinST) with the 95% confidence intervals (vertical bars) using a conditional Poisson model adjusting for seasonality, long-term time trend, and the day of week. RR, relative risk. Figure S5. Distributions of the location-specific (A) minimum suicide temperature (ºC) (MinST) and (B) maximum suicide temperature (ºC) (MaxST) against summer average temperature (ºC) by country. Larger symbols indicate the country-specific median values of the MinST and MaxST against the average of the summer temperature by country. The summer temperature was defined as the average of daily ambient temperature during June to September in the countries in Northern Hemisphere and during December to March in the countries in Southern Hemisphere. Figure S6. Country-specific and location-specific lag-cumulative temperature-suicide associations for maximum suicide temperature (MaxST) vs. minimum suicide temperature

South Africa
We collected data from 52 districts from 1 January 2000 to 31 December 2013 and excluded 13 districts (6 with substantial missing values of temperature above 30% and 7 with a few suicides less than 30 for the entire study period). Therefore, a total of 39 districts were included in our study (see the full list in Table   S2). Daily suicide counts were kindly supplied by Statistics South Africa, who had no role in the data analysis or interpretation. It is likely that the suicide dataset was underreported, compared with the previous publications (Pillay-van Wyk et al. 2016;WHO 2014). Daily mean temperature was obtained from the Agricultural Research Council of South Africa and the National Oceanographic and Atmospheric Association (NOAA) of the United States; the latter was also the source of the daily total of sunshine duration. The missing rates for the daily series of temperature ranged 0.0%-28.9% (0.0%-22.2% after removing stratum without suicide) across the districts.

Spain
We collected data from 52 provincial capital cities from 1 January 1990 to 31 December 2013 and excluded two cities (Palencia with no temperature data since 1991 and Ceuta with no temperature data before 2003).
A total of 50 cities were included in our study (see the full list in Table S2). Daily suicide counts were obtained from Spain National Institute of Statistics. Daily mean temperature and the daily total of sunshine duration were obtained from Spain National Meteorology Agency. The missing rates for the daily series of temperature ranged 0.0%-16.8% (0.0%-15.2% after removing stratum without suicide) across the cities.

Switzerland
We collected data from 8 regions (cantons; see the full list in Table S2) from 1 January 1995 to 31 December 2013. Daily suicide counts were obtained from the Federal Office of Statistics. The suicide data includes in part assisted suicide from 1995 to 2008. Daily mean temperature, relative humidity, and the daily total of sunshine duration were collected from the IDAWEB web database (a service provided by MeteoSwiss, the Zurich/Fluntern). The missing rates for the daily series of temperature ranged 0.0%-0.0% across the regions.

Taiwan
We collected data from 3 cities (see the full list in Table S2) from 1 January 1994 to 31 December 2007.
Daily suicide counts were obtained from the Department of Statistics, Ministry of Health and Welfare in Taiwan. Daily mean temperature, relative humidity, and the daily total of sunshine duration were obtained from the Taiwan Environmental Protection Administration. The missing rates for the daily series of temperature ranged 0.0%-0.0% across the cities.

UK
We collected data from 10 regions in England and Wales (see the full list in Table S2) from 1 January 1990 to 31 December 2011. Daily suicide counts were obtained from the Office of National Statistics. Daily mean temperature and the daily total of sunshine duration were obtained from the British Atmospheric Data Centre. The missing rates for the daily series of temperature ranged 0.0%-0.0% across the regions.

USA
We collected data from 135 cities from 1 January 2001 to 31 December 2005 and excluded Honolulu with no suicide in the study period. A total of 134 cities were included in our study (see the full list in Table S2).
Daily suicide counts were obtained from the National Center for Health Statistics. Daily mean temperature and relative humidity were obtained from the National Climate Data Center of the National Oceanic and Atmospheric Administration. The missing rates for the daily series of temperature ranged 0.0%-11.2% (0.0%-11.7% after removing stratum without suicide) across the cities.

Vietnam
We collected data from Ho Chi Minh City from 1 January 2010 to 31 December 2013. Daily suicide counts were obtained from the official book named A6 provided by Vietnam Ministry of Health. Daily mean temperature and relative humidity were obtained from the National Climate Data Center of the National Oceanic and Atmospheric Administration. There was no missing of the daily series of temperature. a Median values and ranges (minimum and maximum) of the location-specific latitude and longitude.
b Mean values and ranges (minimum and maximum) of the location-specific relative humidity and sunshine duration over the study periods. NA: not applicable.
c Median values and IQRs (25 th -75 th percentiles) of the annual number of hotter days, defined as the daily mean temperature above 25 ºC that is the absolute mean temperature of maximum suicide temperature (MaxST) approximately between Japan and South Korea.   Hemisphere. Table S5. Country-specific minimum suicide temperature (°C) (MinST) and maximum suicide temperature (°C) (MaxST) and pooled lag-cumulative relative risks (RRs) with 95% confidence intervals (CIs) by country in subgroup analyses by sex and age groups (young, < 65 years of age; elder, ≥ 65 years of age) using a conditional Poisson model adjusting for seasonality, long-term time trend, and the day-of-week.