Abstract
Background
Pesticide self-poisoning accounts for 14%–20% of global suicides. Paraquat, a highly lethal pesticide, is involved in a considerable number of suicides in some countries. This study investigated the burden, trends, geographic variations, and demographic characteristics of pesticide and paraquat suicides in Taiwan.
Methods
Suicide data (1987, 1992, 1997, and 2002–2017) were obtained from Taiwan’s cause-of-death mortality data files. Trends and geographic variations in pesticide and paraquat suicides were examined using graphs and maps. Logistic regression was used to investigate the differences in the characteristics of pesticide versus non-pesticide suicides and paraquat versus non-paraquat pesticide suicides.
Results
Age-standardised pesticide suicide rates decreased by 74%, from 7.7 to 2.0 per 100,000, in 1987–2017. Paraquat was the leading agent involved in pesticide suicides, accounting for 59.9% (2,268/3,786) of pesticide suicides where the pesticide agents involved were recorded on the death certificate and at least 4.0% (2,268/56,361) of overall suicides in Taiwan. The percentage of paraquat suicides in overall suicides varied markedly across cities/counties (0.4%–15.6%). Pesticide suicides were more likely to be male, older people aged 65 + years, and from rural areas than non-pesticide suicides. By contrast, paraquat suicides tended to be younger than non-paraquat pesticide suicides, with no difference in sex distributions in the two groups.
Conclusions
The ban on paraquat implemented in 2018–2020 would have a significant impact on pesticide suicides and overall suicides, particularly in high-risk areas and demographic groups.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12889-026-29134-x.
Keywords: Paraquat, Pesticide suicide, Pesticide ban, Suicide prevention, Taiwan
Background
Pesticide self-poisoning is a significant public health issue worldwide, accounting for 14%–20% of global suicides, or 110,000–168,000 deaths, annually [1]. Past research suggested marked variations in pesticide suicide rates over time [2, 3], across different regions within a country [4–6], and across individuals of various demographic and socioeconomic characteristics [7]. Globally, the majority of pesticide suicides occur in the rural areas of countries in South East Asia and the Western Pacific region [1, 8]. Pesticide self-poisoning is also among the most commonly used methods for suicide in Taiwan, with pesticides being the leading agents involved in poisoning-related suicides [9].
Pesticides commonly used in suicide vary markedly in case fatality [10]. Paraquat, a commonly used herbicide, is among the most lethal pesticides involved in self-poisonings [11], with a case-fatality ratio exceeding 70% in some studies [12, 13]. Paraquat was found to be the most frequently ingested pesticide in South Korea and Taiwan, accounting for 35.5% and 34.7% of pesticide poisonings in previous hospital-based studies in South Korea (2008) and Taiwan (2012–2019), respectively [4, 14]. However, few previous studies investigated the epidemiology of paraquat suicides based on population-based nationwide samples.
Product-specific analyses of the characteristics of people who died by pesticide self-poisoning could inform prevention strategies targeted at regulating certain highly hazardous pesticides (HHPs), such as paraquat, and regulating or banning HHPs was recommended by the World Health Organization and Food and Agriculture Organization as among the most effective suicide prevention strategies [15, 16]. In South Korea, for instance, the age-standardised rate of pesticide suicide fell annually by 28.2% between 2011 and 2014 following the ban on paraquat in 2011–2012 [2]. A systematic review of 27 studies from 16 countries showed that national bans on HHPs commonly used in self-poisonings appeared to be effective in reducing pesticide-specific and overall suicide rates, whilst the evidence for an effect of restricting HHPs sales to authorised users was less consistent [17].
In Taiwan, previous studies showed that suicide by pesticide poisoning decreased markedly (67%) between 1987 and 2010 [18], and pesticide suicide rates were greater in rural areas with a higher percentage of agricultural workers [19] and in males and older people [18] than their counterparts. The reduction in pesticide suicide rates in 1987–2010 was thought to be linked to the decrease in the agricultural population, but not to the bans of a number of pesticides during this period [18], as these pesticide bans did not include major pesticides found to be involved in pesticide suicides, such as paraquat [20]. Paraquat (24% soluble concentrate) was later banned in Taiwan in 2018–2020 in two stages—the import and production of paraquat were first banned in February 2018, and the sale and use of paraquat were completely banned in February 2020. However, there have been no previous national studies on the epidemiology of paraquat suicide, including its trends and geographic variations and the characteristics of people who died by paraquat self-poisoning, prior to the nationwide paraquat ban in Taiwan.
Before implementing a pesticide ban aimed at reducing suicides, it is essential to identify the leading agent contributing to pesticide suicides and its burden. Additionally, the pre-ban burden of paraquat suicides across different sociodemographic groups and geographical areas is the basis for estimating the impact of a paraquat ban on suicide prevention. This study aimed to examine the burden of suicides by pesticide and paraquat poisoning before the paraquat ban (2018–2020) in Taiwan. Specifically, we investigated the contribution of paraquat self-poisoning to pesticide suicides in Taiwan, as well as the trends, geographic variations, and characteristics of people who died by pesticide self-poisoning, in particular paraquat self-poisoning, in 1987–2017.
Methods
Data
We extracted suicide data for people aged 15 years or over from Taiwan’s cause-of-death data files, coded using the International Classification of Diseases, Tenth Revision (ICD-10), for years 1987, 1992, 1997, and 2002–2017. The ICD-10 includes a specific 3-digit code (i.e., ICD-10 X68) for suicide by pesticide poisoning. By contrast, in the International Classification of Diseases, Ninth Revision (ICD-9), suicides by pesticide poisoning were grouped with suicides using other solid or liquid substances under the 3-digit ICD-9 code E950. Taiwan officially started to use the ICD-10 in 2009, allowing the identification of pesticide suicides. To identify suicide by pesticide poisoning before 2009, we obtained ICD-10 coded cause-of-death data files for 1987, 1992, 1997, and 2002–2008 from the Ministry of Health and Welfare; these data were specially created for the preparation of the transition from the ICD-9 to the ICD-10; a previous study showed that the change in ICD revisions had little impact on the total number of suicide [21].
In this study, suicides were identified using the following ICD-10 codes: X60–X84 (suicide), Y10–Y34 (undetermined death), W75–W76 and W83–W84 (accident by suffocation), and X48 (accident by pesticide poisoning). A previous study from Taiwan showed that many of the deaths in the above alternative cause-of-death categories (i.e., undetermined death, accident by suffocation, and accident by pesticide poisoning) were likely to be misclassified suicides [22]. For the rest of the paper, the term ‘suicide’ referred to both certified suicides and possible suicides (undetermined deaths, accidental suffocation, or accidental pesticide poisoning) combined. Suicides by pesticide poisoning (‘pesticide suicides’) were identified using ICD-10 68 (suicide), Y18 (undetermined death), and X48 (accident); all other suicides were classified as ‘non-pesticide suicides’.
The following variables were extracted from the cause-of-death data files: sex, age, date of death, and the city/county of registered residence. Age was grouped into four categories (15–24, 25–44, 45–64, and 65 + years). Individuals aged 15–24 years generally represent those who have recently entered the labour market, those aged 25–44 and 45–64 years correspond to individuals in the early and late stages of career development, respectively, and those aged 65 years and over are mostly retired. Based on the date of death, the season variable was created with the following categories–spring (March through May), summer (June through August), autumn (September through November), and winter (December through February). Geographic analyses were conducted at the city/county level, which corresponds to conventional administrative boundaries in Taiwan. More detailed geographic information was not available for pesticide suicide data. Taiwan’s 23 cities/counties were grouped according to the tertiles of the percentage of agricultural workers in the working population—as an indicator of rurality and ease of access to pesticides—resulting in three categories: urban (low percentage; 0.19%–1.16%), semi-rural (medium percentage; 1.21%–7.66%), and rural (high percentage; 9.31%–26.94%). Data for the percentage of agricultural workers were from the 2000 census.
Among pesticide suicides, we extracted information on the pesticides involved from the original free text in the cause-of-death field on the death certificate. We grouped the pesticides according to the classification used by Taiwan’s Council of Agriculture, which used a matrix classifying each pesticide component based on the purpose of use (e.g., herbicide, insecticide, fungicide, rodenticide, etc.) and the chemical structure (e.g., organophosphate, carbamate, etc.) of the pesticides. An 11-category variable was created for the classification of pesticides commonly involved in suicides in Taiwan (Table S1 and Table S2); they included four groups of herbicides (paraquat, glyphosate, other herbicides, unspecified herbicides), four groups of insecticides (organophosphate, carbamate, other insecticides, unspecified insecticides), rodenticides, other pesticides, and ‘unspecified pesticides’ (i.e., the specific pesticides involved were not recorded on the death certificate although pesticide poisoning was certified as the cause of death). The ‘unspecified pesticides’ group was not included in the main analysis concerning the number and percentage of suicides involving different categories of pesticides. Detailed data for the cause-of-death free text on the death certificate were only available for the years 1987, 1992, 1997, and 2002–2012. By contrast, in 2013–2017, detailed data on the pesticide agents involved in suicide were not provided by the Ministry of Health and Welfare, although the number of paraquat pesticides and their percentage in all suicides could be calculated based on data provided by the Ministry of Health and Welfare, which compiled the data for the number of suicides by paraquat poisoning by year, sex, age (in 5-year bands), and city/county as part of its surveillance system to support suicide prevention strategies and monitor trends in paraquat suicide.
Population data were extracted from the Demographic Fact Books (1987–2017) published by the Ministry of the Interior, Taiwan.
Statistical analysis
We calculated the age-standardised rates of pesticide suicide in each year using the World Health Organization (WHO) 2000 world standard population. The percentage of paraquat suicides (i.e., pesticide suicides by paraquat poisoning) in all pesticide suicides with known pesticide agents was calculated by year in 1987, 1992, 1997, 2002–2012, and its trend was examined graphically. The annual number and percentage of pesticide suicides using other categories of pesticides were also calculated.
We calculated the number and percentage of pesticide suicides and paraquat suicides, overall and by sex, age group, and city/county with different levels of the percentage of agricultural workers in 2002–2017. Maps generated with ArcGIS Version 10.4 were used to examine the geographic variations in the rate and percentage (in all suicides) of pesticide suicides and paraquat suicides across the 23 cities/counties in Taiwan. We used Spearman’s correlation coefficients to investigate the correlation between the rates of pesticide and paraquat suicides and the percentage of agricultural workers in the working population across cities/counties.
We used chi-square tests to compare the characteristics between pesticide and non-pesticide suicides, as well as between pesticide suicides involving paraquat (‘paraquat suicides’) and those involving other non-paraquat pesticides (‘non-paraquat pesticide suicides’). We did not include suicides by self-poisoning using unspecified pesticides (n = 4927) and unspecified herbicides (n = 192) in the categories of either paraquat suicides or non-paraquat pesticide suicides due to a lack of information on the specific pesticide agents involved to make such a distinction. Logistic regression analyses were used to investigate the association of demographic characteristics, season, and geographic area (grouped according to the percentage of agricultural workers) with pesticide suicide (versus non-pesticide suicide [reference group]) or paraquat suicide (versus non-paraquat pesticide suicide [reference group]). In subsequent multivariable analyses, we simultaneously adjusted for sex, age group, season, and geographic area to investigate whether these factors were independently associated with pesticide suicide or paraquat suicide.
We conducted sensitivity analyses to examine the potential impact of suicide misclassification on the results by including data for certified suicides only (ICD-10 X60–X84) and excluding deaths of undetermined intent, accidental suffocation, or accidental pesticide poisoning.
The comparison between paraquat and non-paraquat pesticide suicides was restricted to the years (1987, 1992, 1997, and 2002–2012) when detailed information about the pesticide involved in suicides was available; suicides using unspecified pesticides or unspecified herbicides were excluded from the analysis.
A sensitivity analysis was conducted by excluding two cities/counties (i.e., Taichung City and Taichung County) from the comparison due to the high percentage of unspecified pesticide suicides (85.0% and 76.7%, respectively), indicating relatively low data quality in the two cities/counties.
Because suicides using unspecified pesticides accounted for a substantial proportion (56.5%) of overall pesticide suicides, we conducted supplementary analyses to examine whether the city’s or county’s percentage of suicides using unspecified pesticides was associated with the proportion of paraquat suicides among specified pesticide suicides and the rate of paraquat suicides in the city/country, indicating differential certification practices in recording paraquat as the causative agent across cities/counties. Scatter plots and Spearman’s correlation coefficients were used to examine the association between the percentage of unspecified pesticide suicides and (1) the proportion of paraquat suicides among specified pesticide suicides and (2) the age-standardised paraquat suicide rate across cities/counties.
All analyses were conducted using SPSS software version 19. We report p-values and 95% confidence intervals; a p-value < 0.05 was used as a conventional reference point for evaluating the strength of evidence.
Results
Trends and burdens of pesticide suicide and paraquat suicide
Age-standardised pesticide suicide rates decreased markedly from 7.7 per 100,000 in 1987 to 2.0 per 100,000 in 2017, i.e., a 74% reduction, while the percentage of paraquat suicides among pesticide suicides remained stable (Fig. 1). This suggests that the decline was due to factors affecting all pesticides (e.g., decreased accessibility as the farming population decreased) rather than factors specific to paraquat.
Fig. 1.
Trends in pesticide suicide rates* and percentage of paraquat poisoning suicides** in Taiwan, 1987–2017. *Age-standardised pesticide suicide rates (per 100,000; red line). Data for pesticide suicides (International Classification of Diseases, Tenth Revision, code X68) were only available in 1987, 1992, 1997, and 2002 onward. **The percentage of paraquat poisoning suicides in pesticide suicides with available information about the pesticide agents involved (%; blue bars). Detailed data were only available in 1987, 1992, 1997, and 2002–2012
Table 1 shows the distribution of pesticide suicides according to the category of pesticides involved by year, excluding those using unspecified pesticides (i.e., suicides in which the detailed information of the pesticide agents involved was unavailable on the death certificate). The analysis was restricted to the years 1987, 1992, 1997, and 2002–2012 when detailed data for the causes of deaths in free text on the death certificate were available to the authors and could be used to identify and classify the pesticides involved. In each year studied, paraquat was the leading pesticide agent involved, accounting for 50.0%–65.3% of pesticide suicides using known pesticide agents. Overall, paraquat accounted for 2,268 deaths (59.9% of 3,786 pesticide suicides with known pesticide agents involved) over the period. Paraquat was used in at least 26% (2,268/8,713) of all pesticide suicides, or at least 4.0% (2,268/56,361) of total suicides, when those using unspecified pesticides were additionally included (Table S2). If the percentage 59.9% was used to estimate the number of paraquat suicides in pesticide suicides without detailed information on pesticides involved (n = 4927; Table S2), the total estimated number of paraquat suicides would increase to 5,220, accounting for 9.3% (5,220/56,361) of all suicides.
Table 1.
Number (n) and percentage (%) of suicides by the category of pesticides* in Taiwan, 1987, 1992, 1997, 2002–2012
| Year | Herbicide | Insecticide | Rodenticide | Other pesticides | Total | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Paraquat | Glyphosate | Other herbicides + Unspecified herbicides | Organophosphate | Carbamate | Other insecticides + Unspecified insecticides | |||||||||||||
| n | (%) | n | (%) | n | (%) | n | (%) | n | (%) | n | (%) | n | (%) | n | (%) | n | (%) | |
| 1987 | 218 | (64.5) | 0 | (0.0) | 25 | (7.4) | 81 | (24.0) | 1 | (0.3) | 12 | (3.6) | 1 | (0.3) | 0 | (0.0) | 338 | (100.0) |
| 1992 | 125 | (64.1) | 3 | (1.5) | 8 | (4.1) | 47 | (24.1) | 1 | (0.5) | 6 | (3.1) | 5 | (2.6) | 0 | (0.0) | 195 | (100.0) |
| 1997 | 124 | (52.1) | 7 | (2.9) | 6 | (2.5) | 87 | (36.6) | 5 | (2.1) | 7 | (2.9) | 2 | (0.8) | 0 | (0.0) | 238 | (100.0) |
| 2002 | 178 | (62.9) | 7 | (2.5) | 4 | (1.4) | 69 | (24.4) | 21 | (7.4) | 2 | (0.7) | 2 | (0.7) | 0 | (0.0) | 283 | (100.0) |
| 2003 | 158 | (65.3) | 3 | (1.2) | 3 | (1.2) | 58 | (24.0) | 16 | (6.6) | 4 | (1.7) | 0 | (0.0) | 0 | (0.0) | 242 | (100.0) |
| 2004 | 135 | (64.3) | 4 | (1.9) | 3 | (1.4) | 46 | (21.9) | 12 | (5.7) | 8 | (3.8) | 2 | (1.0) | 0 | (0.0) | 210 | (100.0) |
| 2005 | 171 | (58.2) | 17 | (5.8) | 8 | (2.7) | 69 | (23.5) | 23 | (7.8) | 3 | (1.0) | 2 | (0.7) | 1 | (0.3) | 294 | (100.0) |
| 2006 | 155 | (61.3) | 9 | (3.6) | 11 | (4.3) | 40 | (15.8) | 29 | (11.5) | 9 | (3.6) | 0 | (0.0) | 0 | (0.0) | 253 | (100.0) |
| 2007 | 165 | (56.3) | 5 | (1.7) | 17 | (5.8) | 49 | (16.7) | 47 | (16.0) | 10 | (3.4) | 0 | (0.0) | 0 | (0.0) | 293 | (100.0) |
| 2008 | 158 | (53.4) | 7 | (2.4) | 20 | (6.8) | 55 | (18.6) | 31 | (10.5) | 24 | (8.1) | 1 | (0.3) | 0 | (0.0) | 296 | (100.0) |
| 2009 | 173 | (60.5) | 7 | (2.4) | 28 | (9.8) | 29 | (10.1) | 26 | (9.1) | 21 | (7.3) | 2 | (0.7) | 0 | (0.0) | 286 | (100.0) |
| 2010 | 149 | (50.0) | 6 | (2.0) | 29 | (9.7) | 53 | (17.8) | 30 | (10.1) | 27 | (9.1) | 3 | (1.0) | 1 | (0.3) | 298 | (100.0) |
| 2011 | 170 | (64.6) | 5 | (1.9) | 20 | (7.6) | 32 | (12.2) | 23 | (8.7) | 11 | (4.2) | 1 | (0.4) | 1 | (0.4) | 263 | (100.0) |
| 2012 | 189 | (63.6) | 6 | (2.0) | 11 | (3.7) | 33 | (11.1) | 40 | (13.5) | 14 | (4.7) | 1 | (0.3) | 3 | (1.0) | 297 | (100.0) |
| All study years | 2268 | (59.9) | 86 | (2.3) | 193 | (5.1) | 748 | (19.8) | 305 | (8.1) | 158 | (4.2) | 22 | (0.6) | 6 | (0.2) | 3786 | (100.0) |
*Herbicide, insecticide, rodenticide, and other pesticides in pesticide suicides with information available on the pesticides involved (pesticide suicides using unspecified pesticides excluded)
Organophosphates and carbamates were the second and third most common categories of pesticides involved in pesticide suicides. The percentage of organophosphates decreased from 24.0% in 1987 to 11.1% in 2012, while the percentage of carbamates increased from 0.3% to 13.5% during the period (Table 1). Overall, organophosphates accounted for 19.8% (748/3,786) of pesticide suicides with known pesticide agents, followed by carbamates (8.1%; n = 305), other/unspecified herbicides (5.1%; n = 193), other/unspecified insecticides (4.2%; n = 158), and glyphosate (2.3%; n = 86). Pesticide suicides involving other categories of pesticides were less common.
The percentage of pesticide suicides using unspecified pesticides was shown in Table S2; they were the largest group in each year, although its percentage decreased from 65.8% in 1987 to 40.2% in 2012, suggesting improved data quality and completeness over time. Supplementary analyses showed no statistical evidence for a correlation between the percentage of unspecified pesticide suicides and the proportion of paraquat suicides among specified pesticide suicides (Spearman’s correlation coefficients = -0.064, p = 0.77) or the age-standardised paraquat suicide rate (Spearman’s correlation coefficients = -0.364, p = 0.088) (Figure S1), indicating no differential practices of recording paraquat as the causative agent involved in pesticide suicides across cities/counties.
Among a total of 70,039 suicides, there were 9,006 pesticide suicides and 2,869 paraquat suicides in Taiwan between 2002 and 2017 (Table 2). Pesticide and paraquat suicides accounted for 12.9% and 4.1% of total suicides (2002–2017), respectively. The percentage of pesticide and paraquat suicides in all suicides was higher in males (versus females) and highest in the older group aged 65 + years and rural areas, where pesticide poisoning accounted for one in four suicides (24.4%) and paraquat poisoning accounted for one in twelve suicides (8.1%), respectively. The sensitivity analysis using certified suicides only showed similar results (Table S3).
Table 2.
The number (n) and percentage (%) of pesticide and paraquat suicides in Taiwan, 2002–2017
| All suicides | Suicides by pesticide poisoning | Percentage of pesticide poisoning in all suicides | Suicides by paraquat poisoning | Percentage of paraquat poisoning in all suicides | |||
|---|---|---|---|---|---|---|---|
| n | n | (%a) | %b | n | (%a) | %b | |
| Overall | 70,039 | 9006 | (100.0) | 12.9 | 2869 | (100.0) | 4.1 |
| Sex | |||||||
| Male | 47,471 | 6473 | (71.9) | 13.6 | 2081 | (72.5) | 4.4 |
| Female | 22,568 | 2533 | (28.1) | 11.2 | 788 | (27.5) | 3.5 |
| Age group | |||||||
| 15–24 years | 3863 | 176 | (2.0) | 4.6 | 70 | (2.4) | 1.8 |
| 25–44 years | 25,849 | 2035 | (22.6) | 7.9 | 794 | (27.7) | 3.1 |
| 45–64 years | 24,453 | 3462 | (38.4) | 14.2 | 1054 | (36.7) | 4.3 |
| 65 + years | 15,874 | 3333 | (37.0) | 21.0 | 951 | (33.1) | 6.0 |
| City/county | |||||||
| Urban | 33,769 | 1697 | (18.8) | 5.0 | 569 | (19.8) | 1.7 |
| Semi-rural | 18,095 | 2877 | (31.9) | 15.9 | 835 | (29.1) | 4.6 |
| Rural | 18,175 | 4432 | (49.2) | 24.4 | 1465 | (51.1) | 8.1 |
aThe column-wise percentage of pesticide and paraquat suicides by sex, age group, and city/county with different levels of agricultural workers (low, medium, and high). The cities/counties were classified according to the percentage of agricultural workers as urban (0.19%–1.16%), semi-rural (1.21%–7.66%), or rural (9.31%–26.94%)
bThe row-wise percentage of pesticide and paraquat poisoning in all suicides, overall and in specific sex, age, and city/county groups
The geographical distribution of pesticide suicides and paraquat suicides
Figure 2 shows the maps of the age-standardised rates and percentages of pesticide and paraquat suicides across 23 cities/counties in Taiwan in 2002–2017. Detailed data for each city/county were presented in Table S4. For both pesticide and paraquat suicides, the highest rates and percentages were found in counties in Central and East Taiwan, whilst the lowest rates and percentages were found in major cities across Taiwan (Fig. 2). There were 23-fold and 62-fold variations in pesticide and paraquat suicide rates across cities/counties, respectively (ranged 0.35–7.95 per 100,000 for pesticide suicide and 0.06–3.84 per 100,000 for paraquat suicide). The percentages of pesticide suicides in all suicides ranged 2.4%–32.6% and those for paraquat suicides ranged 0.4%–15.6% across cities/counties (Table S4). There was a strong positive association of the percentages of agricultural workers with the rates of pesticide suicides (Spearman’s correlation coefficients = 0.76, p < 0.001) and paraquat suicides (Spearman’s correlation coefficients = 0.68, p < 0.001), respectively.
Fig. 2.
Maps of pesticide and paraquat suicide rates and percentages in all suicides in Taiwan, 2002–2017. A Pesticide suicide rate; B Pesticide suicide percentage; C Paraquat suicide rate; and D Paraquat suicide percentage. Note: 1. Taipei City; 2. Keelung City; 3. Taipei County; 4. Kaohsiung City; 5. Taichung City; 6. Hsinchu City; 7. Taoyuan County; 8. Tainan City; 9. Chiayi City; 10. Hsinchu County; 11. Taichung County; 12. Kaohsiung County; 13. Yilan County; 14. Miaoli County; 15. Hualien County; 16. Tainan County; 17. Penghu County; 18. Changhua County; 19. Nantou County; 20. Pingtung County; 21. Taitung County; 22. Chiayi County; 23. Yunlin County
Characteristics of pesticide suicides and paraquat suicides
Table 3 shows the comparison of characteristics between pesticide and non-pesticide suicides. Compared to non-pesticide suicides, pesticide suicides were more likely to be male (versus females; adjusted odds ratio [aOR] = 1.13, 95% confidence interval [CI] 1.07–1.19), from the older group aged 65 + years (aOR = 0.54 [95% CI 0.48–0.60] for 15–24-year-olds; aOR = 0.48 [95% CI 0.45–0.51] for 25–44-year-olds; aOR = 0.79 [95% CI 0.75–0.84] for 45–64-year-olds), and from semi-rural and rural areas (versus urban areas; aOR = 3.29 [95% CI 3.09–3.50] for semi-rural areas and 4.80 [95%CI 4.52–5.10] for rural areas, respectively). The sensitivity analysis using certified suicides only showed similar results (Table S5).
Table 3.
Comparison of characteristics in pesticide suicides versus non-pesticide suicides in Taiwan, 1987, 1992, 1997, 2002–2012
| All suicides (N = 56,361) |
Pesticide suicides (N = 8,713) |
Non-pesticide suicides (N = 47,648) |
χ2 | df | p | Unadjusted | Adjusted* | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | (%) | n | (%) | n | (%) | OR | 95% CI | p | OR | 95% CI | p | ||||||
| Sex | 25.3 | 1 | < 0.001 | ||||||||||||||
| Male | 38,290 | (67.9) | 6121 | (70.3) | 32,169 | (67.5) | 1.14 | (1.08 | , 1.19) | < 0.001 | 1.13 | (1.07 | , 1.19) | < 0.001 | |||
| Female | 18,071 | (32.1) | 2592 | (29.7) | 15,479 | (32.5) | 1.00 | 1.00 | |||||||||
| Age group | 780.0 | 3 | < 0.001 | ||||||||||||||
| 15–24 years | 3722 | (6.6) | 454 | (5.2) | 3268 | (6.9) | 0.50 | (0.45 | , 0.56) | < 0.001 | 0.54 | (0.48 | , 0.60) | < 0.001 | |||
| 25–44 years | 21,760 | (38.6) | 2389 | (27.4) | 19,371 | (40.7) | 0.44 | (0.42 | , 0.47) | < 0.001 | 0.48 | (0.45 | , 0.51) | < 0.001 | |||
| 45–64 years | 18,402 | (32.7) | 3160 | (36.3) | 15,242 | (32.0) | 0.75 | (0.71 | , 0.79) | < 0.001 | 0.79 | (0.75 | , 0.84) | < 0.001 | |||
| 65 + years | 12,477 | (22.1) | 2710 | (31.1) | 9767 | (20.5) | 1.00 | 1.00 | |||||||||
| Season | 1.7 | 3 | 0.63 | ||||||||||||||
| Spring | 14,477 | (25.7) | 2197 | (25.2) | 12,280 | (25.8) | 0.96 | (0.90 | , 1.02) | 0.20 | 0.95 | (0.88 | , 1.01) | 0.11 | |||
| Summer | 15,171 | (26.9) | 2357 | (27.1) | 12,814 | (26.9) | 0.99 | (0.92 | , 1.05) | 0.65 | 0.95 | (0.89 | , 1.02) | 0.13 | |||
| Autumn | 13,907 | (24.7) | 2144 | (24.6) | 11,763 | (24.7) | 0.98 | (0.91 | , 1.04) | 0.47 | 0.94 | (0.88 | , 1.01) | 0.09 | |||
| Winter | 12,806 | (22.7) | 2015 | (23.1) | 10,791 | (22.6) | 1.00 | 1.00 | |||||||||
| City/county | 3173.7 | 2 | < 0.001 | ||||||||||||||
| Urban | 26,703 | (47.4) | 1826 | (21.0) | 24,877 | (52.2) | 1.00 | 1.00 | |||||||||
| Semi-rural | 14,746 | (26.2) | 2896 | (33.2) | 11,850 | (24.9) | 3.33 | (3.13 | , 3.54) | < 0.001 | 3.29 | (3.09 | , 3.50) | < 0.001 | |||
| Rural | 14,912 | (26.5) | 3991 | (45.8) | 10,921 | (22.9) | 4.98 | (4.69 | , 5.29) | < 0.001 | 4.80 | (4.52 | , 5.10) | < 0.001 | |||
OR = odds ratio for pesticide suicide compared to non-pesticide suicide; CI = confidence interval. The cities/counties were classified according to the percentage of agricultural workers as urban (0.19%–1.16%), semi-rural (1.21%–7.66%), or rural (9.31%–26.94%). ORs, 95% CIs, and p values were highlighted in bold when p value was < 0.05
*Adjusted for all variables in the table. The Cox & Snell R² value is 0.067, and the Nagelkerke R² value is 0.116
The characteristics of paraquat suicides and non-paraquat pesticide suicides were generally similar, except that paraquat suicides were younger compared to non-paraquat pesticide suicides (aOR = 2.43 [95% CI 1.71–3.44] for 15–24-year-olds and 2.00 [95% CI 1.66–2.40] for 25–44-year-olds; Table 4). There was no difference in the distributions of sex or season between the two groups, but paraquat suicides were marginally less likely to be from rural areas (aOR = 0.84 [95%CI 0.71–1.01] versus urban areas). The sensitivity analysis using certified suicides only showed similar results (Table S6). Additionally, in the sensitivity analysis that excluded two cities/counties (Taichung City and Taichung County) with relatively low data quality (indicated by a high proportion of pesticide suicides lacking the information on the pesticide agents involved [Table S4]), the results were similar. However, the lower odds of paraquat suicides occurring in rural areas, compared to non-paraquat pesticide suicides, became more pronounced (aOR = 0.81 [95% CI 0.68–0.97]) (Table S7).
Table 4.
Comparison of characteristics in paraquat versus non-paraquat pesticide suicides in Taiwan, 1987, 1992, 1997, 2002–2012
| Paraquat suicides (N = 2,268) |
Non-paraquat pesticide suicides (N = 1,326) |
χ2 | df | p | Unadjusted | Adjusted* | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | (%) | n | % | OR | 95% CI | p | OR | 95% CI | p | ||||||
| Sex | 2.1 | 1 | 0.14 | ||||||||||||
| Male | 1644 | (72.5) | 931 | (70.2) | 1.12 | (0.96 | , 1.30) | 0.14 | 1.08 | (0.93 | , 1.26) | 0.33 | |||
| Female | 624 | (27.5) | 395 | (29.8) | 1.00 | 1.00 | |||||||||
| Age group | 83.1 | 3 | < 0.001 | ||||||||||||
| 15–24 years | 152 | (6.7) | 47 | (3.5) | 2.48 | (1.75 | , 3.52) | < 0.001 | 2.43 | (1.71 | , 3.44) | < 0.001 | |||
| 25–44 years | 736 | (32.5) | 278 | (21.0) | 2.03 | (1.70 | , 2.44) | < 0.001 | 2.00 | (1.66 | , 2.40) | < 0.001 | |||
| 45–64 years | 742 | (32.7) | 511 | (38.5) | 1.12 | (0.95 | , 1.31) | 0.19 | 1.10 | (0.93 | , 1.29) | 0.26 | |||
| 65 + years | 638 | (28.1) | 490 | (37.0) | 1.00 | 1.00 | |||||||||
| Season | 1.7 | 3 | 0.65 | ||||||||||||
| Spring | 568 | (25.0) | 335 | (25.3) | 1.00 | (0.82 | , 1.22) | 1.00 | 1.02 | (0.83 | , 1.24) | 0.89 | |||
| Summer | 583 | (25.7) | 359 | (27.1) | 0.96 | (0.79 | , 1.16) | 0.66 | 0.97 | (0.79 | , 1.17) | 0.73 | |||
| Autumn | 598 | (26.4) | 326 | (24.6) | 1.08 | (0.89 | , 1.32) | 0.43 | 1.10 | (0.90 | , 1.34) | 0.35 | |||
| Winter | 519 | (22.9) | 306 | (23.1) | 1.00 | 1.00 | |||||||||
| City/county | 11.5 | 2 | 0.003 | ||||||||||||
| Urban | 510 | (22.5) | 279 | (21.0) | 1.00 | 1.00 | |||||||||
| Semi-rural | 674 | (29.7) | 338 | (25.5) | 1.09 | (0.90 | , 1.33) | 0.38 | 1.05 | (0.86 | , 1.28) | 0.62 | |||
| Rural | 1084 | (47.8) | 709 | (53.5) | 0.84 | (0.70 | , 1.00) | 0.044 | 0.84 | (0.71 | , 1.01) | 0.056 | |||
OR = odds ratio for paraquat suicide compared to non-paraquat pesticide suicide; CI = confidence interval. The cities/counties were classified according to the percentage of agricultural workers as urban (0.19%–1.16%), semi-rural (1.21%–7.66%), or rural (9.31%–26.94%). ORs, 95% CIs, and p values were highlighted in bold when p value was < 0.05
*Adjusted for all variables in the table. The Cox & Snell R² value is 0.027, and the Nagelkerke R² value is 0.036
Discussion
The annual age-standardised pesticide suicide rates decreased by 74% from 7.7 to 2.0 per 100,000 in 1987–2017. In years when detailed data were available (1987, 1992, 1997, and 2002–2012), paraquat was the leading agent involved in pesticide suicides, accounting for approximately 60% of pesticide suicides for which the information on the specific pesticide agents involved was available. It was estimated that paraquat accounted for 4.0%–9.3% of all suicides in Taiwan. Geographically, the percentage of paraquat suicides in all suicides varied markedly (0.4%–15.6%) across cities/counties. Compared to non-pesticide suicides, pesticide suicides were more likely to be male, from the older group aged 65 + years, and from rural areas. By contrast, paraquat suicides were younger than non-paraquat pesticide suicides, with no difference in the distributions of sex and season in the two groups.
Strengths and limitations
This study has several strengths, including its use of data from the cause-of-death mortality data files that allow a comprehensive and countrywide assessment of suicides by pesticide self-poisoning. Access to detailed information on the pesticides used in self-poisoning, as recorded on the death certificates, enabled us to investigate the categories of pesticides commonly involved in pesticide suicides. However, there are several limitations of the study. First, this study employed an ecological design and primarily descriptive analyses to examine long-term temporal patterns using annual data, rather than applying more advanced methods, such as regression modelling, or utilising finer-grained data, such as monthly records, which were not available. Therefore, causal inferences regarding the relationship between pesticide availability and suicide cannot be drawn from the findings. Second, suicide data were obtained from cause-of-death mortality files derived from death certificates, which are not designed for detailed analysis of the specific pesticides involved in suicide. Under Taiwanese law, any suspected unnatural death (e.g., accident, suicide, or homicide) must be referred to the prosecutor’s office for a medico-legal investigation. However, autopsies and toxicological examinations are not routinely conducted and are typically reserved for cases with uncertain causes of death or suspected homicide. In suicide cases, the cause of death is generally determined based on interviews with family members, evidence collected at the scene, and relevant documentation, such as medical records, when available. In addition, Taiwan’s cause-of-death data files are coded using ICD-10, which does not allow the recording of detailed information on the pesticides involved. Consequently, there is a lack of systematic data collection regarding the specific pesticides used in pesticide-related suicides. A substantial proportion of the agents involved in pesticide suicides were unspecified during the study period, particularly in the earlier years. Although the high lethality of paraquat poisoning may reduce the likelihood of underreporting compared with other pesticides, the lack of comprehensive autopsy and toxicological examinations means that some suicides recorded as involving unspecified pesticides may in fact have involved paraquat. Conversely, it is possible that paraquat, which is better known to the public in Taiwan as a pesticide used in self-poisoning than other pesticides, may have been more likely to be reported than other pesticides. However, although the proportion of unspecified pesticides decreased over time, the percentage of paraquat suicides amongst those deaths that involved a specified pesticide remained essentially unchanged during the study period. In addition, the lack of statistical association between the percentage of unspecified pesticide suicides and the proportion of paraquat suicides in specified pesticide suicides across cities and counties suggests that the practice of recording paraquat as the causative agent involved in pesticide suicides is similar across geographical areas. Future studies based on comprehensive information about the pesticides involved, such as those based on forensic toxicology records [23], would allow a more detailed investigation of specific pesticides used in self-poisoning. Third, our ability to assess annual trends is constrained by the availability of pesticide suicide data only in selected years before 2002, and by the lack of available paraquat suicide data from 2012 to 2017, which limits evaluation of paraquat’s contribution to total pesticide suicides. Fourth, we categorised Taiwan’s administrative areas into tertiles based on the percentage of agricultural workers as a proxy for rurality. While this allows for within-Taiwan comparisons, caution is warranted when making cross-territory comparisons. Fifth, official mortality statistics often underestimate suicide incidence [24]. Although cause‑of‑death coding for injuries and poisonings in Taiwan has high accuracy [25], some deaths officially recorded as undetermined intent or accidental may in fact be suicides [22]. Therefore, our primary analyses also included deaths of undetermined intent, accidental pesticide poisoning, and accidental suffocation [22]. Sixth, the study examined the epidemiology of pesticide suicides in Taiwan in 1987–2017 only. The impact of the 2018–2020 paraquat ban was not assessed in this study and has been reported elsewhere [26]. Lastly, the analysis was completed after paraquat had already been banned in Taiwan, which would limit its impact on domestic regulatory policies. However, we believe the findings remain highly valuable for informing pesticide regulation in other countries, with the potential to prevent deaths from self-poisoning with paraquat and other highly hazardous pesticides globally.
Comparison with previous studies
Our study showed a 74% decrease in pesticide suicide rates in Taiwan over three decades (1987–2017), whereas overall suicide increased by 15% (age-standardised rate of certified suicide increased from 10.9 to 12.5 per 100,000) during the same period [27]. Changes in trends in pesticide suicide rates were found to be associated with certain sociodemographic and pesticide accessibility (e.g., the percentage of agricultural population) factors [2, 18, 28]. Additionally, pesticide suicides decreased after restricting access to highly hazardous pesticides in several countries [17]. A previous study indicated that the reduction in pesticide suicides in Taiwan from 1987 to 2010 was mainly associated with the decrease in the agricultural population, which had easier access to pesticides than their counterpart [18]. During this period, bans were imposed on several pesticides, including selected high-strength formulations of organophosphates and carbamates, but not on the leading pesticide involved in suicide (i.e., paraquat). Bans implemented during this period were not found to be associated with a reduction in pesticide suicides. By contrast, studies from other countries showed a reduction in pesticide suicide following bans on high-lethality pesticides commonly used in self-poisoning. In South Korea, pesticide suicide rates declined markedly following the 2011–2012 paraquat ban [2, 28, 29]. Deaths from pesticide poisoning were on the rise in Japan before 1986 [30] but showed a marked 92% reduction from 1986 to 2019, following its policy to de-register high-concentration paraquat products [31]. In low- and middle-income Asian countries, such as Sri Lanka, India, Bangladesh, and China, where pesticide suicide constituted a larger proportion of total suicides, deaths from pesticide poisoning have declined markedly over the last 10–20 years. Evidence suggests that bans on highly hazardous pesticides contributed to reductions in both pesticide and overall suicides [32–35]. The WHO now considers pesticide regulation to be the most effective approach to suicide prevention in countries and areas where pesticides are an important means of suicide [15].
Paraquat emerged as the predominant agent involved in pesticide suicides in Taiwan over the study period, accounting for approximately 60% of all pesticide suicides involving a known pesticide agent; the corresponding figure was 26% if pesticide suicides where the pesticide agents involved were not specified were included and assumed to be not paraquat, thus resulting in a possible underestimation. Earlier epidemiologic data (1985–1993) from the Taiwan National Poison Center indicated that paraquat was responsible for 54.4% of poison-related fatalities [36]. In a later study based on suicide data for 2006–2008 in Taiwan, paraquat was the leading pesticide involved in pesticide suicides, accounting for 28.5% (= 471/1651) of all cases [20]. Similar to Taiwan, paraquat is the main product responsible for pesticide poisoning deaths in some countries. Paraquat accounts for 16.3% of poisoning deaths in Australia [7], approximately one-third of fatal poisonings involving known pesticides in Sri Lanka [10], where paraquat has been banned since 2008–2011, and 71.2% of pesticide poisoning deaths in South Korea in 2011 [37], prior to the complete ban of paraquat in 2012. In these countries, prevention strategies targeting paraquat, e.g., a national ban, would contribute to a decrease in paraquat suicides, and pesticide suicides as a whole.
We found a marked geographical difference in paraquat suicide rates within Taiwan, with the highest rate (3.84 per 100,000) being 62 times greater than the lowest rate (0.06 per 100,000) across cities/counties. In addition, there was a notable disparity in the proportion of paraquat suicides among all suicides between urban areas and rural areas, ranging from 0.4% to 15.6%. Similar findings regarding the urban-rural disparity in pesticide suicide rates have been reported in South Korea [38] and China [5, 39]. The high pesticide suicide rates in rural areas could be associated with easy access to pesticides, lower quality of medical service, an ageing population (with physical vulnerability and a tendency to use immediately available lethal means), social isolation, socioeconomic disadvantages, and attitudes towards seeking help for mental illnesses [38, 40–43]. The high correlation between the percentage of agricultural workers and the pesticide suicide rate shown in this study suggested that the accessibility of pesticides influenced pesticide suicide rates. Similarly, previous research reported an association between the percentage of the agriculture workforce and regional pesticide suicide rates even after controlling for socioeconomic disadvantage indicators [19].
Males and older people had a greater burden of pesticide suicides and paraquat suicides compared to their counterparts. The sex- and age-specific patterns of pesticide suicides in Taiwan were similar to those observed in South Korea [38], Australia [7], and Germany [44], where males and older people constituted the populations most affected. By contrast, individuals who died by pesticide self-poisoning on the Greek island of Crete consisted mainly of middle-aged males [45]. In China, pesticide suicide was previously found to be most prevalent among young females [46]; however, there was a more pronounced decline in pesticide suicides among females than males over the last two decades [5, 47, 48]. Compared to low-and middle-income countries, industrialised countries such as Taiwan have an ageing rural population. The higher likelihood of older people dying from pesticide self-poisoning (compared to their younger counterparts) in this study, relative to other suicide methods, could be attributable to the higher accessibility to pesticides and case fatality ratio of pesticide poisoning among older people living in rural areas. The higher male-to-female ratio in pesticide suicides, compared to non-pesticide suicides, is consistent with our previous findings [14]. In addition to the propensity of males to choose methods of higher lethality [49], the male predominance of pesticide self-poisoning in Taiwan could also be attributable to males’ higher agricultural participation than their female counterparts, which led to higher accessibility to pesticides [50].
Our study found no difference in the distributions of sex and season between paraquat suicide and non-paraquat pesticide suicide. However, individuals who died by paraquat self-poisoning tended to be younger than those who died from non-paraquat pesticide self-poisoning. Similarly, a recent study from Taiwan found that, among patients who presented to hospitals with pesticide self-poisoning, those who ingested paraquat were younger and had a higher education level than those ingesting other pesticides [14]. By contrast, a recent hospital-based study from Malaysia did not find an association of sex and age with the types of pesticide involved in pesticide self-poisonings [51]. Our finding of similar sociodemographic profiles, except for the age pattern, in paraquat suicides and non-paraquat pesticide suicides suggested that most people who died by pesticide self-poisoning might turn to use pesticides for self-harm because of easy access to pesticides. The high case fatality ratio of paraquat poisoning, even with small ingested volumes, which is common among young people with low suicidal intent [52], could explain the relative preponderance of young individuals in paraquat suicides compared to non-paraquat pesticide suicides. Another possible explanation for the younger age of paraquat suicides is that younger people’s choice of pesticides for self-poisoning is more likely to be influenced by the ‘cognitive availability’ —that is, their awareness or knowledge of it as a means of suicide [53]. As paraquat has been the most commonly used pesticide in self-poisoning and received more media coverage than other pesticides in Taiwan, it is likely to be more prominent in the minds of younger people when considering suicide methods. The higher odds of paraquat suicide compared to non-paraquat pesticide suicide in urban areas may also be explained by the greater ‘cognitive availability’ of paraquat relative to other pesticides among people living in these areas, as urban residents might be more aware of paraquat ingestion as a means for suicide than the ingestion of other pesticides. However, the hypothesis needs support from future studies that investigate the reasons why individuals who self-poison using pesticides choose specific pesticide agents.
Conclusions
Paraquat was the leading pesticide involved in suicides in Taiwan, accounting for 59.9% of pesticide suicides where the specific pesticide agents involved could be identified. Paraquat accounted for at least 4.0% of all suicides in the pre-ban period. The percentage of paraquat suicides among all suicides showed marked geographical variations and accounted for a significant proportion (approximately 16%) of all suicides in counties with the highest proportion of agricultural population. The burden of paraquat suicides was higher among males, individuals aged 65 + years, and rural areas than their counterparts. The findings indicated the potential of the paraquat ban as an effective approach to preventing suicide, particularly in high-risk areas and demographic groups.
Supplementary Information
Additional file 1: Table S1. Number (n) and percentage (%) of suicides by the category of pesticides (herbicide, insecticide, rodenticide, and other pesticides) in Taiwan, 1987, 1992, 1997, 2002–2012. Table S2. Number (n) and percentage (%) of suicides by the category of pesticides (herbicide, insecticide, rodenticide, other pesticides, and unspecified pesticides) involved in pesticide suicides in Taiwan, 1987, 1992, 1997, 2002–2012. Table S3. The number (n) and percentage (%) of pesticide and paraquat suicides by sex, age group, and city/county with different levels of agricultural workers (low, medium, and high), and the percentage of pesticide and paraquat poisoning in all suicides (N=59,829) in Taiwan, 2002-2017 (sensitivity analysis based on certified suicides only). Table S4. Age-standardised rate, number, and percentage of pesticide suicide and paraquat suicide across 23 cities/counties in Taiwan, 2002–2017. Table S5. Comparison of characteristics in pesticide suicides (n=6,657) versus non-pesticide suicides (n=40,226) in Taiwan, 1987, 1992, 1997, 2002–2012 (sensitivity analysis based on certified suicides only). Table S6. Comparison of characteristics in paraquat suicides (n=1,698) versus non-paraquat pesticide suicides (n=1,027) in Taiwan, 1987, 1992, 1997, 2002–2012 (sensitivity analysis based on certified suicides only). Table S7. Comparison of characteristics in paraquat suicides (n=2,196) and non-paraquat pesticide suicides (n=1,262) in Taiwan (excluding Taichung City and Taichung County), 1987, 1992, 1997, 2002–2012. Figure S1. The correlations between the percentage of unspecified pesticide suicides and (A) the proportion of paraquat suicides in specified pesticide suicides and (B) age-standardised paraquat suicide rate (per 100,000) across 23 cities/counties in Taiwan, 2002–2012.
Acknowledgements
The authors thank the Department of Mental Health, Ministry of Health and Welfare, Taiwan, for their support in accessing the data.
Abbreviations
- aOR
Adjusted odds ratio
- CI
Confidence interval
- HHPs
Highly hazardous pesticides
- ICD-10
International Classification of Diseases, Tenth Revision
- ICD-9
International Classification of Diseases, Ninth Revision
- OR
Odds ratio
- WHO
World Health Organization
Authors’ contributions
Conceptualisation: HYC, CYT and SSC. Data curation: HYC, CYT and SSC. Formal analysis: HYC, CYT and CYL. Writing – original draft: HYC, CYT and SSC. Writing – review & editing: HYC, CYT, CYL, CYH, THY, DG, ME and SSC. Funding acquisition: SSC and CYH.
Funding
This work was supported by the Centre for Pesticide Suicide Prevention, University of Edinburgh [grant numbers c-12538894 and c-14689012] and the National Science and Technology Council, Taiwan [grant number NSTC 111-2314-B-002-290]. The Centre for Pesticide Suicide Prevention is funded by Coefficient Giving on the recommendation of Good Ventures Foundation. DG is supported by the NIHR Biomedical Research Centre at University Hospitals Bristol and Weston NHS Foundation Trust and the University of Bristol, England. SSC is supported by the Population Health and Welfare Research Center from the Featured Areas Research Center Program within the framework of the Higher Education Sprout Project funded by the Ministry of Education (MOE) in Taiwan (grant numbers NTU-113L900401, NTU-114L900401, NTU-115L900401). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Data availability
The data that support the findings of this study are available from Department of Mental Health, Ministry of Health and Welfare, Taiwan but restrictions apply to the availability of these data, which were used under licence for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the Department of Mental Health, Ministry of Health and Welfare, Taiwan.
Declarations
Ethics approval and consent to participate
This study was approved by Research Ethics Committee of National Taiwan University Hospital (201606036RIN). All procedures were conducted in accordance with the Declaration of Helsinki.
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.
Hwei Yuen Chang and Chao-Ying Tu contributed equally to this work.
References
- 1.Mew EJ, Padmanathan P, Konradsen F, Eddleston M, Chang SS, Phillips MR, et al. The global burden of fatal self-poisoning with pesticides 2006-15: Systematic review. J Affect Disord. 2017;219:93–104. 10.1016/j.jad.2017.05.002. [DOI] [PubMed] [Google Scholar]
- 2.Cha ES, Chang SS, Choi Y, Lee WJ. Trends in pesticide suicide in South Korea, 1983–2014. Epidemiol Psychiatr Sci. 2019;29:e25. 10.1017/S2045796019000118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gunnell D, Fernando R, Hewagama M, Priyangika WD, Konradsen F, Eddleston M. The impact of pesticide regulations on suicide in Sri Lanka. Int J Epidemiol. 2007;36(6):1235–42. 10.1093/ije/dym164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lee WJ, Cha ES. Overview of pesticide poisoning in South Korea. J Rural Med. 2009;4(2):53–8. https://www.jstage.jst.go.jp/article/jrm/4/2/4_2_53/_pdf. [Google Scholar]
- 5.Page A, Liu S, Gunnell D, Astell-Burt T, Feng X, Wang L, et al. Suicide by pesticide poisoning remains a priority for suicide prevention in China: Analysis of national mortality trends 2006–2013. J Affect Disord. 2017;208:418–23. 10.1016/j.jad.2016.10.047. [DOI] [PubMed] [Google Scholar]
- 6.Patel V, Ramasundarahettige C, Vijayakumar L, Thakur JS, Gajalakshmi V, Gururaj G, et al. Suicide mortality in India: a nationally representative survey. Lancet. 2012;379(9834):2343–51. 10.1016/S0140-6736(12)60606-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Osborne NJ, Cairns R, Dawson AH, Chitty KM, Buckley NA. Epidemiology of coronial deaths from pesticide ingestion in Australia. Int J Hyg Environ Health. 2017;220:478–84. 10.1016/j.ijheh.2017.01.009. [DOI] [PubMed] [Google Scholar]
- 8.Karunarathne A, Gunnell D, Konradsen F, Eddleston M. How many premature deaths from pesticide suicide have occurred since the agricultural Green Revolution? Clin Toxicol (Phila). 2020;58(4):227–32. 10.1080/15563650.2019.1662433. [DOI] [PubMed] [Google Scholar]
- 9.Chen YY, Lu TH, Lee HJ, Kuo CJ, Chiu CH, Chen CC. Comparison of method-specific suicide rates between Taiwan and South Korea (in Chinese). Taipei City Med J. 2006;3(10):982–91. 10.6200/tcmj.2006.3.10.04. [DOI] [Google Scholar]
- 10.Buckley NA, Fahim M, Raubenheimer J, Gawarammana IB, Eddleston M, Roberts MS, et al. Case fatality of agricultural pesticides after self-poisoning in Sri Lanka: a prospective cohort study. Lancet Glob Health. 2021;9(6):e854–62. 10.1016/S2214-109X(21)00086-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dawson AH, Eddleston M, Senarathna L, Mohamed F, Gawarammana I, Bowe SJ, et al. Acute human lethal toxicity of agricultural pesticides: a prospective cohort study. PLoS Med. 2010;7(10):e1000357. 10.1371/journal.pmed.1000357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Min YG, Ahn JH, Chan YC, Ng SH, Tse ML, Lau FL, et al. Prediction of prognosis in acute paraquat poisoning using severity scoring system in emergency department. Clin Toxicol (Phila). 2011;49(9):840–5. 10.3109/15563650.2011.619137. [DOI] [PubMed] [Google Scholar]
- 13.Lee Y, Lee JH, Seong AJ, Hong CK, Lee HJ, Shin DH, et al. Arterial lactate as a predictor of mortality in emergency department patients with paraquat intoxication. Clin Toxicol (Phila). 2012;50(1):52–6. 10.3109/15563650.2011.639716. [DOI] [PubMed] [Google Scholar]
- 14.Tu CY, Yen TH, Chang CM, Chen HY, Yen YC, Guo MC, et al. Characteristics and psychopathology of 1,086 patients who self-poisoned using pesticides in Taiwan (2012–2019): A comparison across pesticide groups. J Affect Disord. 2022;300:17–26. 10.1016/j.jad.2021.12.058. [DOI] [PubMed] [Google Scholar]
- 15.Lee YY, Chisholm D, Eddleston M, Gunnell D, Fleischmann A, Konradsen F, et al. The cost-effectiveness of banning highly hazardous pesticides to prevent suicides due to pesticide self-ingestion across 14 countries: an economic modelling study. Lancet Glob Health. 2021;9(3):e291–300. 10.1016/S2214-109X(20)30493-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Preventing suicide: a resource for pesticide registrars and regulators: World Health Organization and Food and Agriculture Organization of the United Nations. 2019. Available from: https://www.who.int/publications-detail/preventing-suicide-a-resource-for-pesticide-registrars-and-regulators.
- 17.Gunnell D, Knipe D, Chang SS, Pearson M, Konradsen F, Lee WJ, et al. Prevention of suicide with regulations aimed at restricting access to highly hazardous pesticides: a systematic review of the international evidence. Lancet Glob Health. 2017;5(10):e1026–37. 10.1016/S2214-109X(17)30299-1. [DOI] [PubMed] [Google Scholar]
- 18.Chang SS, Lu TH, Eddleston M, Konradsen F, Sterne JA, Lin JJ, et al. Factors associated with the decline in suicide by pesticide poisoning in Taiwan: a time trend analysis, 1987–2010. Clin Toxicol (Phila). 2012;50(6):471–80. 10.3109/15563650.2012.688835. [DOI] [PubMed] [Google Scholar]
- 19.Chang SS, Lu TH, Sterne JA, Eddleston M, Lin JJ, Gunnell D. The impact of pesticide suicide on the geographic distribution of suicide in Taiwan: a spatial analysis. BMC Public Health. 2012;12:260. 10.1186/1471-2458-12-260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lin JJ, Chang SS, Lu TH. The leading methods of suicide in Taiwan, 2002–2008. BMC Public Health. 2010;10(1):480. 10.1186/1471-2458-10-480. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lu TH. Impact of automated coding system implementation based on ICD-10 on mortality statistics in Taiwan: a bridge coding study. Taiwan J Public Health. 2020;39(5):578–97. 10.6288/tjph.202010_39(5).109039. [DOI] [Google Scholar]
- 22.Chang SS, Sterne JA, Lu TH, Gunnell D. Hidden’ suicides amongst deaths certified as undetermined intent, accident by pesticide poisoning and accident by suffocation in Taiwan. Soc Psychiatry Psychiatr Epidemiol. 2010;45(2):143–52. 10.1007/s00127-009-0049-x. [DOI] [PubMed] [Google Scholar]
- 23.Ghimire R, Utyasheva L, Pokhrel M, Rai N, Chaudhary B, Prasad PN, et al. Intentional pesticide poisoning and pesticide suicides in Nepal. Clin Toxicol (Phila). 2022;60(1):46–52. 10.1080/15563650.2021.1935993. [DOI] [PubMed] [Google Scholar]
- 24.Ohberg A, Lonnqvist J. Suicides hidden among undetermined deaths. Acta Psychiatr Scand. 1998;98(3):214–8. 10.1111/j.1600-0447.1998.tb10069.x. [DOI] [PubMed] [Google Scholar]
- 25.Lu T-H, Lee M-C, Chou M-C. Accuracy of cause-of-death coding in Taiwan: types of miscoding and effects on mortality statistics. Int J Epidemiol. 2000;29(2):336–43. 10.1093/ije/29.2.336. [DOI] [PubMed] [Google Scholar]
- 26.Lin CY, Hsu CY, Lee MB, Chang SS. Impact of the paraquat ban on reducing suicide in Taiwan: the effect on 2020 suicide rates. J Suicidology. 2022;17(1):80–7. 10.30126/JoS.202203_(1).0011. [DOI] [Google Scholar]
- 27.The Cause of Death Statistics: Department of Statistics of Ministry of Health and Welfare, Taiwan. 2024. Available from: https://dep.mohw.gov.tw/DOS/lp-5069-113.html.
- 28.Cha ES, Chang SS, Gunnell D, Eddleston M, Khang YH, Lee WJ. Impact of paraquat regulation on suicide in South Korea. Int J Epidemiol. 2016;45(2):470–9. 10.1093/ije/dyv304. [DOI] [PubMed] [Google Scholar]
- 29.Lin CY, Hsu CY, Choi M, Gunnell D, Eddleston M, Lee WJ, Chang SS: Longer-term impact of pesticide bans on suicide in South Korea. BMJ Glob Health 2026;11(2):e022329. 10.1136/bmjgh-2025-022329. [DOI] [PMC free article] [PubMed]
- 30.Ito T, Nakamura Y. Deaths from pesticide poisoning in Japan, 1968–2005: data from vital statistics. J Rural Med. 2008;3(1):5–9. https://www.jstage.jst.go.jp/article/jrm/3/1/3_1_5/_pdf. [Google Scholar]
- 31.Eddleston M, Nagami H, Lin CY, Davis ML, Chang SS. Pesticide use, agricultural outputs, and pesticide poisoning deaths in Japan. Clin Toxicol (Phila). 2022;60(8):933–41. 10.1080/15563650.2022.2064868. [DOI] [PubMed] [Google Scholar]
- 32.Knipe DW, Metcalfe C, Fernando R, Pearson M, Konradsen F, Eddleston M, et al. Suicide in Sri Lanka 1975–2012: age, period and cohort analysis of police and hospital data. BMC Public Health. 2014;14:839. 10.1186/1471-2458-14-839. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Yan Y, Jiang Y, Liu R, Eddleston M, Tao C, Page A, et al. Impact of pesticide regulations on mortality from suicide by pesticide in China: an interrupted time series analysis. Front Psychiatry. 2023;14:1189923. 10.3389/fpsyt.2023.1189923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Chowdhury FR, Dewan G, Verma VR, Knipe DW, Isha IT, Faiz MA, et al. Bans of WHO Class I Pesticides in Bangladesh-suicide prevention without hampering agricultural output. Int J Epidemiol. 2018;47(1):175–84. 10.1093/ije/dyx157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Indira M, Chowallur MJ, M KA, Lin CY, Chang SS, Sooraj M, et al. Effect of pesticide ban on suicide trend - a 20-year study from a tertiary care center in Central Kerala from 2001 to 2020. Clin Toxicol (Phila). 2022;60(11):1214–9. 10.1080/15563650.2022.2129671. [DOI] [PubMed] [Google Scholar]
- 36.Yang CC, Wu JF, Ong HC, Hung SC, Kuo YP, Sa CH, et al. Taiwan National Poison Center: epidemiologic data 1985–1993. J Toxicol Clin Toxicol. 1996;34(6):651–63. 10.3109/15563659609013825. [DOI] [PubMed] [Google Scholar]
- 37.Moon JM, Chun BJ, Cho YS. The characteristics of emergency department presentations related to acute herbicide or insecticide poisoning in South Korea between 2011 and 2014. J Toxicol Environ Health A. 2016;79(11):466–76. 10.1080/15287394.2016.1172529. [DOI] [PubMed] [Google Scholar]
- 38.Cha ES, Khang YH, Lee WJ. PLoS ONE. 2014;9(4):e95299. 10.1371/journal.pone.0095299. Mortality from and incidence of pesticide poisoning in South Korea: findings from National Death and Health Utilization Data between 2006 and 2010. [DOI] [PMC free article] [PubMed]
- 39.Phillips MR, Li X, Zhang Y. Suicide rates in China, 1995-99. Lancet. 2002;359:835–40. 10.1016/S0140-6736(02)07954-0. [DOI] [PubMed] [Google Scholar]
- 40.Hirsch JK. A review of the literature on rural suicide: risk and protective factors, incidence, and prevention. Crisis. 2006;27(4):189–99. 10.1027/0227-5910.27.4.189. [DOI] [PubMed] [Google Scholar]
- 41.Judd F, Cooper AM, Fraser C, Davis J. Rural suicide–people or place effects? Aust N Z J Psychiatry. 2006;40(3):208–16. 10.1080/j.1440-1614.2006.01776.x. [DOI] [PubMed] [Google Scholar]
- 42.Eddleston M, Dissanayake M, Sheriff MH, Warrell DA, Gunnell D. Physical vulnerability and fatal self-harm in the elderly. Br J Psychiatry. 2006;189(3):278–9. 10.1192/bjp.bp.105.018671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Conwell Y, Thompson C. Suicidal behavior in elders. Psychiatr Clin North Am. 2008;31(2):333–56. 10.1016/j.psc.2008.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Moebus S, Boedeker W. Frequency and trends of hospital treated pesticide poisonings in Germany 2000–2014. Ger Med Sci. 2017;15. 10.3205/000254. [DOI] [PMC free article] [PubMed]
- 45.Kastanaki AE, Kraniotis CF, Kranioti EF, Nathena D, Theodorakis PN, Michalodimitrakis M. Suicide by pesticide poisoning: findings from the island of Crete. Greece Crisis. 2010;31(6):328–34. 10.1027/0227-5910/a000042. [DOI] [PubMed] [Google Scholar]
- 46.Conner KR, Phillips MR, Meldrum S, Knox KL, Zhang Y, Yang G. Low-planned suicides in China. Psychol Med. 2005;35(8):1197–204. 10.1017/s003329170500454x. [DOI] [PubMed] [Google Scholar]
- 47.Liu Y, Lan Z, Yin Y, Liu NH, Tong Y. Trends in suicide rates and the case-fatality of pesticide self-poisoning in an agricultural county in china, 2009 to 2014. J Affect Disord. 2021;283:52–9. 10.1016/j.jad.2021.01.024. [DOI] [PubMed] [Google Scholar]
- 48.Scholin L, Knipe D, Bandara P, Eddleston M, Sethi A. Banning highly hazardous pesticides saves the lives of young people, particularly females, in low- and middle-income countries. BMC Public Health. 2023;23(1):2249. 10.1186/s12889-023-17071-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Spicer RS, Miller TR. Suicide acts in 8 states: incidence and case fatality rates by demographics and method. Am J Public Health. 2000;90(12):1885–91. 10.2105/ajph.90.12.1885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Agriculture Employed Persons by Gender: Ministry of Agriculture, Taiwan. 2017. Available from: https://echart.moa.gov.tw/index.php?cid=24.
- 51.Chan LF, Chin SJ, Loo TH, Panirselvam RR, Chang SS, Chang HY, et al. Surveillance of pesticide poisoning in an East and a West Malaysian hospital: characteristics of pesticide poisoning and the early impact of a national Paraquat ban. BMC Psychiatry. 2023;23(1):472. 10.1186/s12888-023-04974-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Eddleston M, Gunnell D, Karunaratne A, de Silva D, Sheriff MH, Buckley NA. Epidemiology of intentional self-poisoning in rural Sri Lanka. Br J Psychiatry. 2005;187:583–4. 10.1192/bjp.187.6.583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Florentine JB, Crane C. Suicide prevention by limiting access to methods: a review of theory and practice. Soc Sci Med. 2010;70(10):1626–32. 10.1016/j.socscimed.2010.01.029. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Additional file 1: Table S1. Number (n) and percentage (%) of suicides by the category of pesticides (herbicide, insecticide, rodenticide, and other pesticides) in Taiwan, 1987, 1992, 1997, 2002–2012. Table S2. Number (n) and percentage (%) of suicides by the category of pesticides (herbicide, insecticide, rodenticide, other pesticides, and unspecified pesticides) involved in pesticide suicides in Taiwan, 1987, 1992, 1997, 2002–2012. Table S3. The number (n) and percentage (%) of pesticide and paraquat suicides by sex, age group, and city/county with different levels of agricultural workers (low, medium, and high), and the percentage of pesticide and paraquat poisoning in all suicides (N=59,829) in Taiwan, 2002-2017 (sensitivity analysis based on certified suicides only). Table S4. Age-standardised rate, number, and percentage of pesticide suicide and paraquat suicide across 23 cities/counties in Taiwan, 2002–2017. Table S5. Comparison of characteristics in pesticide suicides (n=6,657) versus non-pesticide suicides (n=40,226) in Taiwan, 1987, 1992, 1997, 2002–2012 (sensitivity analysis based on certified suicides only). Table S6. Comparison of characteristics in paraquat suicides (n=1,698) versus non-paraquat pesticide suicides (n=1,027) in Taiwan, 1987, 1992, 1997, 2002–2012 (sensitivity analysis based on certified suicides only). Table S7. Comparison of characteristics in paraquat suicides (n=2,196) and non-paraquat pesticide suicides (n=1,262) in Taiwan (excluding Taichung City and Taichung County), 1987, 1992, 1997, 2002–2012. Figure S1. The correlations between the percentage of unspecified pesticide suicides and (A) the proportion of paraquat suicides in specified pesticide suicides and (B) age-standardised paraquat suicide rate (per 100,000) across 23 cities/counties in Taiwan, 2002–2012.
Data Availability Statement
The data that support the findings of this study are available from Department of Mental Health, Ministry of Health and Welfare, Taiwan but restrictions apply to the availability of these data, which were used under licence for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the Department of Mental Health, Ministry of Health and Welfare, Taiwan.