Article Text
Abstract
Background Taipei has seen a substantial increase in suicide rates during the past decade, with a significant rise between 2004 and 2006, the time of this study period.
Methods A decompositional analytic method was used to quantify the relative contributions of age, sex and case fatality of methods to attempts and suicides.
Results From 2004 to 2006, the rate of fatal and non-fatal suicide attempts combined for population aged 15 years or above in Taipei increased by 37.3%, while the suicide rate increased by 29.2%. Three factors in these analyses contributed to the increase in suicide rates: (1) an increase in number of attempts, (2) a greater proportion of men among fatal and non-fatal attempts and (3) an increase in the use of a lethal method—burning of charcoal to produce carbon monoxide. The authors estimated that 74.5% and 25.6% among men and women, respectively, of the overall increased suicide mortality were attributable to increased ‘charcoal burning suicides.’
Conclusions The rise in suicide rate reflected an increase in attempts and an influx of working-age men joining the pool of people attempting suicide. The much larger size of the attempter pool had the effect of reducing the case fatality even as the suicide rate climbed. The increase in the number of suicide attempts and the rise in the suicide rate were age-, sex-, and method-specific. These results strongly support the concept that reducing the total number of attempts is a central element to curbing suicides.
- Suicide/self harm
- epidemiology
- mortality
- media exposure
- environmental modification
- methods
- suicide
- models
Statistics from Altmetric.com
Introduction
Suicide rates in Taiwan have increased dramatically, from 10.4 per 100 000 population in 1999 to a historical high of 19.3 in 2006 followed by a modest decline by 2009 to 17.6 per 100 000 population, indicating an overall increase of 70% during the past decade.1 In this study, we seek to extend existing research by understanding the frequency of suicide in the context of data regarding suicide attempts and information regarding those who die from suicide in terms of their age, sex and choice of methods. Using such approaches, it may be possible to more effectively identify trends that drive the increase or decline of temporal rates, which in turn may inform future preventive interventions.
Suicide rates depend on the incidence of attempts and the case fatality of each suicidal act.2 The characteristics of persons who make fatal attempts vary considerably from those who survive.3 4 Ultimately, this reflects the choice of method and its implementation.5–7
Similar to other countries, studies in Taiwan have revealed higher attempt rates among women and younger individuals and higher death rates among men and elders.8 In the context of the significant increase in suicides in Taiwan during the past decade, it remains to be defined whether the upsurge reflected an increase in attempts, a change in the makeup of the groups seeking to kill themselves and/or emergence of new methods.9
The current work is based on data collected in Taipei City, which is the largest city in Taiwan. Taipei has a population of approximately 2.6 million, and it has developed a suicide attempt surveillance registry since October 2003. Like the rest of Taiwan, Taipei has seen a substantial increase in suicide rates during the past decade, with a significant rise between 2004 and 2006, the time of our study period. We use decompositional analysis to examine factors that may contribute to the changes of suicide rate between 2004 and 2006, combining suicide attempt surveillance data and official death records. Decompositional analysis can quantify the relative contributions from changing attempt incidence patterns and the case fatality of different suicide methods among groups differing by age and gender. In this way, we are more able to assess ‘Who?’ and ‘Which method?’. Answers to these questions may offer further insights for the development of more appropriately focused suicide prevention programs.
Methods
Study population
The study population consisted of individuals who were recorded in the Taipei City suicide attempt registry between 1 January 2004 and 31 December 2006. The attempt surveillance system is part of a service-based suicide prevention program run by the Taipei City Suicide Research and Prevention Center. All hospital emergency departments in Taipei City are required to report to the Center every case of self-inflicted injury. The involved hospitals, including eight university-affiliated medical centres and eighteen regional general hospitals, are distributed throughout Taipei. The staff in the emergency department is required to complete a structured case record for all self-harm cases, including basic sociodemographic data, method of attempt, clinical characteristics and management of the patient's condition. These case notes are reported to the surveillance system of the Taipei City Suicide Research and Prevention Center.
Any self-inflicted injury, whether fatal or non-fatal, is registered to the surveillance system irrespective of apparent intent. For purposes of the current study, we included any non-fatal deliberate self-harm as ‘attempted suicide’, given the lack of specificity of the record form and an absence of discriminating criteria, even as some cases may have represented ‘non-suicidal self-injury’. In our analyses, persons younger than 15 years of age were excluded in view of their small numbers and, also, those who were not Taipei City residents.
Information on suicides in Taipei City issued in 2004 and 2006 was ascertained from the Death Certification System (DSC) of the Department of Health in Taiwan. The cause of death was classified according to the International Classification of Diseases Ninth Revision, Clinical Modification (ICD-9-CM) recorded in DSC. Codes under external cause of injury were used (codes E950–E959). The method of suicide was grouped into five categories: (1) intentional self-poisoning by other gases and vapours (E952, of which approximately 90% were carbon monoxide poisoning due to charcoal burning), (2) hanging (E953), (3) jumping (E957), (4) solid/liquid poisoning (E950) and (5) others. To avoid double counting, we excluded mortality cases in the suicide attempt surveillance system, as these cases were also recorded in the DSC.
Data analyses
Age-, sex- and method-specific number of suicide attempts and suicides in Taipei City in 2004 and 2006 were recorded and analysed. The case fatality was calculated by using the number of suicides as the numerator divided by the total number of suicidal acts (ie, case fatality = suicides/[suicides + suicide attempts]×100%). The age- and method-specific attempt rate was calculated by the number of suicidal acts divided by the respective population size and is expressed in number per 100 000 population. Hence, the suicide rate is the product of the attempt rate and the case fatality.
A decompositional method10 11 was used to quantify the change of the suicide rate into the relative contributions of age-, gender- and method- specific factors to account for (1) the increase of suicidal acts and (2) changing case fatality during the 3-year study period, 2004–2006 (for details of decompositional method, see online appendix 1).
Ethical permission for the study was obtained from Taipei City Hospital (IRB no: TCHIRB-991222-E).
Results
In 2004, 2237 non-fatal attempts (with a rate of 103.9 per 100 000 population) were registered in the surveillance system, while 3141 non-fatal attempts (143.9 per 100 000 population) were recorded in 2006, an increase of 38.5%. At the same time, suicide rates in Taipei City increased from 16.1 to 20.7 per 100 000 population aged 15 years or above, an increase of 29.2%. From 2004 through 2006, the rate of fatal and non-fatal suicide attempts combined in Taipei City increased by 37.3% (in 2004—non-fatal, N=2237 and fatal, N=346; in 2006—non-fatal, N=3141 and fatal, N=453). This increase of suicide and non-fatal attempts was evident for all groups regardless of sex (see figure 1).
Overall, the fatal and non-fatal attempt episodes increased for men and women during the 3-year study period—men showed a 38.6% rise and women, 40.3%, nearly 1:1. The more frequent use of charcoal burning (133.0% increase for men; 126.8% for women), as a method to kill oneself, contributed disproportionately to the overall increase (table 1). That is, the net effect of the major increase in charcoal use reduced the relative contributions of other methods, even for those whose overall number increased. For example, from 2004 to 2006, the number of fatal and non-fatal episodes of solid/liquid poisoning among men increased from 405 to 530. The proportion of this method, however, decreased from 46.1% to 43.5%.
The overall increase in the suicide rates was 36.2% for men and 16.8% for women during the period 2004–2006. At the same time, the increase in deaths from charcoal burning was much more substantial, that is, 91.9% for men and 87.0% for women. This increased number of deaths from charcoal burning was observed among all ages for men and women. At the same time, we found a slight decreasing trend in the number of suicides using solid/liquid poisoning (−5.6% and −10.0% for men and women, respectively), even as the number of related attempts had increased from 2004 through 2006. There was also a significant decrease in deaths among women from jumping (34 to 27; −20.6%), although their decrease in attempts was minimal (46 to 43; −6.5%).
As can be seen from table 2, the case fatality was the highest for hanging attempts (87% in 2004 and 85% in 2006), followed by jumping (76% in 2004 and 66% in 2006) and charcoal burning (57% in 2004 and 47% in 2006). The case fatality for solid/liquid poisoning was much lower—1.8% in 2004 and 1.2% in 2006. The case fatality for both sexes decreased significantly for charcoal burning attempts.
Thus, any decrease in case fatality for charcoal and poisons would reduce the suicide rate; however, the reduction was more than offset by the increasing number of attempts of charcoal burning for men and women. The disproportionate higher-than-average increase in the number of attempts of using carbon monoxide poisoning from charcoal combustion (relatively lethal) and a moderate increase in the use of oral poisons (relatively less lethal) concomitantly contributed to the increase in the suicide rate. The case fatality for hanging and jumping remained relatively stable from 2004 to 2006.
It is important to note that for any given method, the case fatality was higher among men. Case fatality generally increased with age except for hanging, which displayed high lethality for all age groups. No gender difference in case fatality was observed for highly lethal methods (ie, hanging and jumping). Case fatality was significantly greater among men than among women for charcoal burning suicide, solid/liquid poisoning and other methods of suicide.
The decompositional analyses (table 3) revealed a complex set of factors contributing to the net increase in rates. Despite the fact that women had a greater per cent increase in the total (fatal and non-fatal) number of attempts (40.3% vs 38.6% for women and men, respectively), the per cent contribution to increased suicide rates from the increased attempts in men and women was 113.0 and 34.3, respectively. It was due to the use of more lethal methods by men. The increasing use of charcoal burning suicide, a method that is relatively lethal, contributes prominently to the increase in suicide rate between 2004 and 2006 in Taipei City. Overall, the increased suicide rates during the period resulted from increasing rate of attempts for men (113.0%) and women (34.3%), and charcoal burning accounted for most of this increase (74.5% for men and 25.6% for women). In this context, the reduced case fatality for charcoal burning from 2004 to 2006 contributed −25.9% (−19.6% for men and −6.3% for women) towards a downward tendency in death rates. However, for the end result, this lower case fatality was totally offset by the effects of more attempts, relatively more men attempting and relatively more use of charcoal for gas inhalation suicide. The increase in charcoal-related deaths among men primarily reflected contributions from those 40–59 years old (43.7/74.5=58.9%) and 25–39 years old (23.6/74.5=31.7%), resulting in a 90.6% increase from this middle-aged group (ie, 25–59 years old). For women choosing charcoal burning, 82.8% was drawn from 25 to 59 years old, that is, 44.1% (11.3/25.6=44.1%) from 25 to 39 years old and 38.7% (9.9/25.6=38.7%) from 40 to 59 years old.
Discussion
Our data and analyses revealed three critical factors that contributed to the rise in suicides in Taipei during 2004–2006: (1) there were more attempts, (2) a relatively greater proportion were made by men and (3) the attempter group (especially men) chose a relatively more lethal method for their attempts—burning charcoal in a confined space to kill themselves with carbon monoxide. About 67.3% (23.6% from 25 to 39 and 43.7% from 40 to 59) of the increase in the suicide rate between 2004 and 2006 is related to the use of charcoal burning by men in the middle years of life, 25–59 years old, typically members of the workforce and a prime economically active group.
Prior studies have also described a disproportionate adoption of charcoal burning among middle-aged men, which occurred in Hong Kong and Taiwan during a downturn in the local economy.9 12–14 While the rise in suicide in Taiwan began during the late 1990s during a time of economic distress and high unemployment, it continued to rise through mid-decade up to its 2006 peak during a period when the economy had begun to recover.15 It remains uncertain whether there was a ‘lag period’ in Taiwan, between the time when economic recovery began and suicide rates started to drop—due to the slow penetration of macroeconomic forces at local levels, or whether other countervailing forces particular to Taiwan, such as local media, perceived political instability and breakdown of the family support system, contributed to the continuing increase in the number of suicides. Further longitudinal observations may help disentangle these factors.
Our findings revealed a change in the epidemiology of suicide in Taipei from 2004 to 2006 and underscored, yet again, that means matter! Similar to prior findings,9 12 16 our results point to the addition of new members to the pool of people who died by suicide. In addition, there was not merely an increase in attempts. A disproportionate number in 2006, compared to the 2004 base year, involved men, and they disproportionately chose to use charcoal as their method of killing themselves. Even as there was a modest decrease in the case fatality for charcoal, jumping and solid/liquid poisoning, these reductions were more than offset by increased numbers, relatively more men and greater use of charcoal.
We cannot definitively explain the reduced lethality of attempts using charcoal; it could be related to the increase of public awareness in the community such that these attempts would be saved from completing the act. But even at its reduced level, it far exceeded the lethality level of solid/liquid poisoning and was greater than the overall case fatality (46.8% vs 1.2% vs 12.6%). Thus, the increase of attempts using charcoal burning contributed heavily to the increase in suicides in Taipei City.
Our observation of the growth in suicides in Taipei was fuelled by (1) an infusion of new attempters (ie, as evidenced by a change in sex and age distribution) and (2) use of a new method, which reinforces and extends prior findings—in different places and in different eras. The additive nature of charcoal burning to suicides in Taipei City tracks with observations from Hong Kong. It is also reminiscent of the rapid adoption of domestic cooking gas in Great Britain and Wales during the late 19th and early 20th centuries, where it quickly became the method of first choice, and the overall rates increased greatly, first in men and later in women. In Taiwan, burning charcoal was not described as a method for suicide before the late 1990s, similar to the emergence of this method in Hong Kong.9 During our study period (2004–2006), the use of charcoal for suicide became the second overall method of death in Taiwan (28.9%) after hanging (35.4%), and in Taipei, however, one prior study reported that charcoal burning (34.4%) had moved to first place with hanging (33.3%) as second.17
The choice of means—a critical factor in overall suicide rates—appears to relate to several key factors, particularly, sex, availability and what might be called ‘local custom’. New methods can spread quickly in the era of electronic and mass print media, as evident in recent years with charcoal. In addition, the internet has widened the scope of the word ‘local’. Availability or accessibility of specific methods is also associated with national and regional variations in choice, such as the use of firearms in the USA, jumping from tall buildings in urban Asia and the use of older generation agricultural pesticides in rural China or Sri Lanka.18–22 Women use ingestion more than men, and as we have shown in this study, men tend to be more lethal in their efforts irrespective of the method chosen. Thus, the changes in suicide in Taipei, as in other settings, reflected the contributions of multiple factors, and in this study, we have sought to disaggregate them in order to better understand their components.
Clinicians have long sought to carefully characterise individuals' efforts to kill themselves. We see comparable importance for populations, where such understanding is necessary to guide potential preventive interventions. Suicide attempts that use poisons often allow a window of time for rescue or an opportunity for reconsidering life-ending decisions. This is not the case for jumping or hanging once the act is initiated. Carbon monoxide, the gas produced during the burning of charcoal, is a poison that is potent and difficult to treat and that requires intensive medical intervention relatively soon after a person starts to succumb to its effects. We saw that the increased attempts—and deaths—in 2006 as compared to 2004 reflected to a substantial degree the heavy representation of charcoal burning among those people—mostly men—who were added to the suicide pool. These additions generally did not diminish the use of other methods; rather, this group of decedents, largely in their middle years of life, chose a highly lethal method for their attempts, and likely, many arrived in hospital emergency settings already dead.
Prior studies have indicated that higher death among men was due to gender differences in the choice of suicide methods.6 23–25 We found that gender differences in case fatality remained evident for suicide by solid/ liquid poisoning and charcoal burning. This likely reflected greater suicidal intent among men and, more specifically, a greater vigour with which they implemented their attempts. For methods of high lethality, such as hanging and jumping, men and women did not differ significantly in case fatality. One prior study from rural Taiwan has demonstrated no gender differences in case fatality when variables such as age and suicide method were considered.23 Pesticide, an agent of high lethality, is the most commonly used substance in solid/liquid poisoning in rural Taiwan; conversely, medication overdose, a method of low lethality, is the most common agent used for solid/liquid poisoning suicide in metropolitan Taipei. In other words, when a method of high lethality was chosen, gender differences in case fatality disappear; conversely, when a less lethal method was selected, the ‘who’ of the attempt made a difference as well as the ‘what’. In sum, means matter, and gender does too!
Similar to other findings,6 23–25 older people tend to have higher case fatality after adjusting for suicide methods. This may be explained, in part, by elders' physical vulnerabilities and higher levels of intent. The only method where we did not find this effect was in hanging, the highly lethal for all groups.
We have noted several potential limitations in our methods, and there are others. We are aware that some people who deliberately harm themselves are not admitted to emergency care or a hospital. Failure to count these people, who made non-fatal attempts in the community, would inflate the estimated case fatality. However, for the results on decompositional analysis, we dealt with the difference in suicide rate between two time points. Unless there was a significant discrepancy in the pattern of under-reporting between 2004 and 2006, we would not anticipate an effect on the results. Moreover, we do not know of any significant changes in the reporting system between 2004 and 2006. At the same time, community deaths were much less likely to be missed, as these would have been captured in Taiwan's death reporting mechanisms.
For decompositional analysis, we considered case fatality and age-, sex- and method-specific incidence of suicide attempt in the explanation of the rising suicide rates; other factors that may contribute to the increase during the study period were not examined. However, the two factors we selected are most relevant for explaining the difference in suicide rates, as suggested by prior research.2–4 Last, there is no specific code for charcoal burning as a method for suicide in the ICD-9-CM system. We know that using E952 as synonymous to charcoal burning modestly overcounts the numbers of attempts and suicides.
Despite their limitations, our data and methods clearly reveal that the upsurge in suicide rates in Taipei City was the result of changes in the quantity and quality of suicide attempts—with more attempts, relatively more men and use of more lethal methods. While the influence of sex, age and method has been studied in Taiwan before,9 26 the use of decompositional analysis allows us to quantify the relative impact of changes in attempts and case fatality on suicide rates These results indicate that the future design and deployment of suicide prevention programs must consider the heterogeneity of potential target groups. The realisation that especially lethal methods have a disproportionate impact on changing death rates challenges planners to consider how to implement preventive interventions outside clinical settings, as many of the people using such means do not arrive alive in hospital emergency rooms. Interventions need to be built upon community-integrated approaches in order to better capture such populations.27 Other more specific approaches include deterring the use of charcoal by placing it behind store counters, which requires a customer to request help for access. The first study of this approach in Hong Kong demonstrated a positive effect of reducing deaths;28 this requires replication in different regions to ascertain its effectiveness and potential for wider application. It will also be important in the future to debunk the misimpression that suicide using charcoal burning is a painless and peaceful way of dying; attempters sometimes receive severe burns when placing themselves close to the burning coals.29 30
In conclusion, this study reinforces our understanding that to better understand the factors contributing changes in suicide rates, it is essential to explore quantitative aspects—charting changes in attempt rates—and qualitative dimensions—defining who is involved and what they choose to do. These types of findings, as we have discussed, provide important lessons for shaping future preventive interventions. They can also serve as a basis for defining future outcome measures to assess potential public health impact.
What is already known on this subject
Suicide has become a significant public health issue in Taiwan, associated with an extraordinarily rapid rise in rates during the past decade. As part of efforts to prevent suicide, it is essential to determine specific factors that may have contributed to the recent dramatic increase in deaths.
What this study adds
Three factors predominately contributed to the increase in suicide rates: (1) an increased incidence of attempts, (2) relatively more men in the attempter pool and (3) use of more lethal methods, specifically, burning charcoal in a confined space.
About 67% of the increase in the suicide rate between 2004 and 2006 was related to the use of charcoal burning by men in the middle years of life.
These results reinforce that reducing the number of attempters is crucial to curb the rise of suicide rate and that the increase in the number of attempted suicides can be age-, sex-, and method-specific. Some focused programs are needed when designing suicide prevention programs.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Download Supplementary Data (PDF) - Manuscript file of format pdf
Footnotes
Funding The research was supported by the department of Health, Taipei City Government and the RGC General Research Fund (2010/2011, HKU 784210M).
Competing interests None.
Ethics approval Ethics approval was obtained from Taipei City Hospital (IRB no: TCHIRB-991222-E).
Provenance and peer review Not commissioned; externally peer reviewed.