Suicide

OWID presents work from many different people and organizations. When citing this entry, please also cite the original data source. This entry can be cited as:

Esteban Ortiz-Ospina, Lindsay Lee and Max Roser (2016) – ‘Suicide’. Published online at OurWorldInData.org. Retrieved from: https://ourworldindata.org/suicide/ [Online Resource]

Death by suicide is an extremely complex issue that causes pain to hundreds of thousands of people every year around the world. The objective of this data entry is to contribute to an informed open debate about ways to prevent suicide. If you are dealing with suicidal thoughts you can receive immediate help by visiting resources such as Suicide.org, or by calling 1-800-SUICIDE in the US.

Every suicide is a tragedy. According to estimates from the World Health Organisation (WHO), over 800,000 people die due to suicide every year. This corresponds to an age-standardized suicide rate of around 11.5 per 100,000 people – a figure equivalent to someone dying of suicide every 40 seconds. Yet suicides are preventable with timely, evidence-based interventions.

The mortality data from the WHO suggest that the prevalence and characteristics of suicidal behaviour vary widely between different communities, in different demographic groups and over time. One important source of heterogeneity, both globally and within countries, is gender: suicide rates are much higher for males, particularly in high-income countries.

Suicide is an extremely complex issue, and while it is not possible to pin down its causes, there are some risk factors that have been identified, mainly through correlations. Mental health, specifically depression, is widely recognized as the most important risk factor. Here we discuss evidence of this, and other important correlates that shed light on possible interventions to prevent suicide.


# Empirical View

# Historical perspective

Self-murder was considered a serious offence in almost all Western European countries from the Middles Ages until (at least) the French Revolution.1 In England, one of the last European countries to decriminalise suicide, self-murder was a crime up until 1961. In many countries, such a Singapore, it is still considered a crime today. Because of this, historical data on this topic is not easily available; indeed, as we discuss in the Data Quality section, the stigma surrounding suicide makes measurement difficult still today.

# Recent suicide mortality data

The World Health Organisation (WHO) published in 2014 the report Preventing suicide: a global imperative, with the aim of increasing awareness of the public health significance of suicide, as well as making suicide prevention a high priority on the global public health agenda. Using the estimates from the WHO, the following visualisation presents suicide rates in different countries across the world over time. While most countries do not exhibit clear trends, it is evident that there are marked cross-country differences in levels.

The fact that cross-country figures do not show clear trends is indicative of the fact that the prevalence and characteristics of suicidal behaviour vary widely between different communities, in different demographic groups and over time. The following two visualization focus on gender differences. The first chart presents trends in suicide rates by gender for a number of countries – it shows that the rates for males do not track those of females. The second visualization shows global differences in gender ratios (that is, the ratio of the suicide rate of males by the suicide rate of females) – as it can be appreciated, in the majority of countries the ratio is larger than one (with male suicide rates being more than 3 times larger than rates for females in most high-income countries). You can switch to the chart view of this graph to explore country-specific trends in more detail.

Gender differences are also markedly different by age. The following graph, produced by Bartolote and Fleischmann (2002)2 with WHO data, shows that gender gaps are not constant across ages – they are largest for middle-aged individuals (45-54 years old) and the elderly (75+ years old).

Distribution of suicide rates (per 100,000) by gender and age, 1998 – Figure 2 in Bertolote and Fleischmann (2002)3

GenderAge Global Rates

# Attempted suicide

Reliable cross-country data about the number of attempted suicides each year is not available. In contrast to data on deaths by suicide, no country in the world reports to the WHO official statistics on attempted suicide. However, it is broadly understood that the number of suicide attempts is much higher than the number of actual suicide deaths each year. In fact, available studies from specific countries suggest that individuals who have made prior suicide attempts are at much higher risk of dying by suicide than individuals who have not made prior suicide attempts. According to the figures published by the Centers for Disease Control, in the US there are roughly 25 attempts for each completed suicide; and for young adults aged 15-24 the ratio is much higher: there are approximately 100-200 suicide attempts for each suicide death.

# Means of death by suicide

Understanding the prevalence of different means of death by suicide is important from a public policy perspective, because restricting access to the means of suicide is a key component of suicide prevention efforts. Indeed, one of the first steps in coping with suicidal thoughts at the individual level is to restrict access to elements that can be used to hurt oneself (see helpguide.org for a complete list of steps).

Unfortunately, given the extreme stigma attached to suicide in most countries – and the fact that it is often considered inappropriate to enquire or publish details in cases of death by suicide because of the effect that this may have on others –, there is little systematic evidence documenting the prevalence of different methods. According to the WHO, the best way for researchers and policymakers to assess the patterns of suicide methods is to rely on data published in the scientific literature. A systematic review of world data for 1990−2007 estimated that around 30% (within a plausible range of 27− 37%) of global suicides are due to pesticide self-poisoning, most of which occur in low and middle income countries.4 This makes it one of the most common methods of suicide globally.

Other academic research has documented community-specific epidemics, such as the case of use of barbecue charcoal to produce carbon monoxide as a means of suicide in China – in this specific instance, the epidemic began in 1998 in Hong Kong, from where it rapidly spread to Taiwan, where it became the most common method of suicide within eight years.5

# Suicide and the internet

The internet is a powerful resource for those seeking help when dealing with suicidal thoughts. Understanding how people seek for such help can be very informative about possible ways to prevent suicide. Specifically, aggregated data stored by search engines can be a powerful source of information to understand and prevent suicide. Google Trends, for instance, allows anyone to explore the popularity of specific search terms that may be linked to specific suicidal behavior – almost in real time.

The following visualization presents three scatter plots documenting the relationship between the popularity of the search terms “suicide help”, “how to suicide” and “how to overdose”. These charts plot Google Trends data on the worldwide popularity of the above-mentioned search terms, where popularity is measured by Google as the ratio between the number of searches for a term in any given week, and the number of all searches in that week (see the documentation from Google for further details). Each point in these scatter plots corresponds to a different week in the period January 2004 – June 2016. As it can be seen, there are striking positive correlations suggesting that (i) search for help tends to go together with search for information on means to suicide; and (ii) search for information on means to suicide tends to go together with queries on drug abuse. Clearly, the individuals using Google are not representative of the world population as a whole – certainly not if we restrict our analysis to queries in the English language.

But these results are indicative of the potential informativeness of data generated by search engines and emphasize how important it is to make help available via the web.

Google Trends - Suicide Queries


# Correlates

Suicide is an extremely complex issue, and while it is not possible to pin down its causes, there are some risk factors that have been identified in the academic literature, mainly through correlations. As always, correlation does not imply causation; but correlation is a valuable starting point to detect risk factors and inform suitable policy interventions.

Depression and other mood disorders are widely recognized as the most important risk factors for suicide. In this entry we discuss evidence of this, and other factors that correlate with suicide.

As mentioned before, the objective of this data entry is to contribute to an informed open debate about ways to prevent suicide. If you are dealing with suicidal thoughts you can receive immediate help by visiting resources such as Suicide.org, or by calling 1-800-SUICIDE in the US.

# Mental health

Depression and other mood disorders are widely recognized among the most important risk factors for suicide. Bertolote and Fleischmann (2002), for instance, provide a systematic review of studies reporting diagnoses of mental disorders for individuals dying from suicide, and discuss the implications of psychiatric diagnosis for suicide prevention.7 They report that 98% of those who died by suicide had a diagnosable mental disorder. The following visualization, taken from their paper, provides details regarding the types of mental disorders in question. The two pie-charts correspond to different population sub-groups: the left chart corresponds to individuals who died by suicide and had been admitted to mental hospitals (labelled as ‘psychiatric inpatient population’), and the right chart corresponds to individuals who died by suicide but had not been admitted to mental hospitals (labelled as ‘general population’). Two points are worth emphasizing: (i) mood disorders – mainly depression – account for the largest share of diagnosed disorders in suicide cases, notably for those individuals who were diagnosed without being admitted to a mental hospital; and (ii) suicide is found associated with a variety of mental disorders. As noted by Bertolote and Fleischmann (2002), each one of the listed mental disorders entails a different therapeutic approach, and hence ‘blanket approaches’ are probably unsound.

# Distribution of psychiatric diagnoses of people who died by suicide among psychiatric inpatients (left) and the general population (right) – Figures 4 and 5 in Bertolote and Fleischmann (2002)

Bertolote and Fleischmann mental disorder

# Economic recessions

Various studies have analyzed the correlation between economic recessions and suicide rates. Chang et al. (2009)8 used WHO and Taiwanese mortality statistics to explore whether there was a correlation between the Asian economic crisis of 1997-1998 and suicide rates. The following graphs summarize their results. As it can be seen, male rates in 1998 rose notably in Japan, Hong Kong and Korea, while rises in female rates were less marked in the same countries. Similar patterns in suicide rates were not seen in Taiwan and Singapore, where the economic crisis had a smaller impact on the economy.

Age-standardized suicide rates (per 100,000 people aged 15 or above) in six East/Southeast Asian countries, 1985-2006 – Chang et al. (2009)9

Chang et al 1997 asian crisis

Other studies, such as Reeves, McKee and Stuckler (2014)10, Chang et al. (2013)11 and Nordt et al. (2015)12 have found similar results. Below is the descriptive evidence presented by Reeves, McKee and Stuckler (2014) from the Financial crisis of 2007–08 in Europe. As it can be seen, there is a clear change in suicide trends right after the crisis.

Age-standardized suicide rates in the European Union, 2001-2011 – Reeves, McKee and Stuckler (2014)13

Reeves et al 2014 EU

Most of the studies exploring the effect of the economy on suicide tend to focus on patterns for specific events in particular countries or regions. However, more comprehensive studies using cross-country data over several different world regions and over longer time windows find similar results. Recently, Nordt et al. (2015)14 retrospectively analysed public data for suicide, population, and economy over the period 2000-2011 across 63 countries in four world regions. Their results correspond with previous research indicating that a rise in unemployment is linked to an increase in suicides; however, they find that the relationship is non-linear – the correlation between unemployment and suicide is stronger in countries with a lower rather than with a higher pre-crisis unemployment rate.

# Gun prevalence

Data from the US indicate that firearms account for 46% of all suicides; in other high-income countries where the prevalence of firearms in the household is lower, firearms account for only 4.5% of all suicides.15

Briggs and Tabarrok (2014)16 investigate the relationship between firearm prevalence and suicide in a sample of all US states over the years 2000–2009. Below is a scatter plot providing evidence of the strong positive correlation that they find. The vertical axis shows the rate of suicide by firearms, and the horizontal axis shows the percent of households that have firearms (according to data from the Behavioral Risk Factor Surveillance System telephone survey conducted by the Center for Disease Control). The authors report that (i) firearms are very strongly related to firearm suicides; (ii) firearms are also strongly related to overall suicides – despite evidence for substantial substitution in method of suicide–; and (iii) there is evidence for a diminishing effect of guns on suicides as ownership levels increase.

Scatterplot of rate of suicide by firearm and the prevalence of guns in the household – Briggs and Tabarrok (2013)17

Briggs and Tabarrok scatterplot

# Daylight patterns

Seasonal variation in suicides has been shown in many countries, with most studies showing spring and summer peaks, contrary to the common belief that suicides peak when nights are longest (see Björkstén et al. (2009) 18 and the references therein). Björkstén et al. (2009) document the accentuation of suicides – but not homicides – with rising latitudes of Greenland in the sunny months. They find that suicides were more concentrated in the summer months north of the Arctic Circle than south of it, and most concentrated in North Greenland where 48% of suicides occurred during the period of constant light. The following visualization provides evidence supporting these results. It shows the monthly distribution of suicides in different areas of Greenland – the y-axis shows the number of suicide cases and the x-axis shows the corresponding month of the year, from January to December.

Monthly distribution of suicides in areas of Greenland – Björkstén, Kripke and Bjerregaard (2009) 19

Bjorksten et al monthly distribution of suicides

# Divorce laws

Unilateral divorce is the process in which one can seek the dissolution of one’s marriage without the consent of a spouse. Legal unilateral divorce laws make seeking divorce easier, and is seen as being especially important for people to be able to escape abusive marriages. Stevenson and Wolfers (2006)20 exploit the variation occurring from the different timing of divorce law reforms across the United States to evaluate how unilateral divorce changed family violence and whether the option provided by unilateral divorce reduced suicide and spousal homicide. The authors estimate a regression of suicide rates using the time since (until) the introduction of unilateral divorce as an explanatory variable (conditional on a series of other control variables). The following graphs summarize their results – further technical details are provided in their paper. Each panel shows the effect of unilateral divorce on suicide rates for a specific age group (with the last panel showing the aggregate effect). Specifically, it can be seen from the second row of this figure that prime-age women account for the bulk of the main effect, with unilateral divorce substantially reducing the suicide rates of women in each of the age groups from 25–65.

Effects of unilateral divorce laws on female suicide by age group – Stevenson and Wolfers (2006)21

unilateral divorce

# Media coverage of suicide

A number of empirical studies have found that there is an increase in suicides following media reports of suicide – a relationship referred to as ‘copycat’ behaviour. Gould et al. (2003)22 provide a number of references from the literature on suicide clusters supporting the contention that suicide is “contagious”. Indeed, based on such findings, the WHO has developed guidelines for the reporting of suicides in the media.

In a review of the literature studying the link between media coverage and suicides, Stack (2003)23 analyzes a total of 293 findings from 42 academic studies, concluding that specific media content makes a substantial difference: studies measuring the effect of either an entertainment or political celebrity suicide story were 14.3 times more likely to find a copycat effect than studies that did not; and studies based on a real as opposed to fictional story were 4.03 times more likely to uncover a copycat effect.

More recent studies corroborate this result. Using data from Austria, Niederkrotenthaler (2010)24 find that repetitive reporting of the same suicide and the reporting of suicide myths were positively associated with higher suicide rates; while coverage of positive coping in adverse circumstances – as covered in media items about suicidal ideation – was negatively associated with suicide rates.

# Alcohol use

Alcohol abuse is strongly linked to death by suicide. In the US, in 2007 alcohol was involved in approximately a third of reported suicides.25 Indeed, several academic studies have found a positive and significant association between per capita alcohol consumption and male suicide rates in a number countries (see Rossow (1996)26).

# Summary – evidence-based interventions for suicide prevention

The following chart, taken from the report Preventing suicide: a global imperative, summarizes the key risk factors for suicide – as per the WHO’s analysis of available evidence – aligned with relevant interventions. Its message is clear: to prevent suicide, interventions need to take place from the universal to the individual level. One-size-fits-all approaches are likely to be ineffective.

Key risk factors for suicide aligned with relevant interventions – WHO (2014)27

risk factors and interventions


# Measurement & Data Quality

# Measurement

Suicide rates are typically reported as number of deaths per 100,000 people; that is, the number of suicides in a country in a year, divided by the population and multiplied by 100,000.

In the data published by the WHO, suicide rates are ‘age-adjusted’. This means that the figures are adjusted for the corresponding age structure of the relevant population. This is necessary because the numbers of deaths per 100,000 population are influenced by the age distribution – two populations with the same age-specific mortality rates for a particular cause of death will have different overall death rates if the age distributions of their populations are different (for more information see the WHO’s Indicator and Measurement Registry).

# Data quality

Data on suicides is deficient for two important reasons. Firstly, there is a problem with the frequency and reliability of vital registration data in many countries – an issue undermining the quality of mortality estimates in general, not just suicide. The WHO reports that, of its 172 Member States for which suicide estimates are published, only 60 have good-quality vital registration data that can be used directly to estimate suicide rates. The estimated suicide rates in the other 112 Member States (which account for about 71% of global suicides) are based on modelling assumptions. The following visualization, from the WHO, shows the distribution of mortality data quality around the world. As might be expected, good quality vital registration systems are more likely available in high-income countries.

Quality of Suicide Mortality Data – Map 2 in WHO (2014)28

WHO Suicide Data Quality

Secondly, there are problems with the accuracy of the official figures made available to WHO by its Member States, since suicide registration is a complicated process involving several responsible authorities with medical and legal concerns. In spite of the fact that in the International Classification of Diseases the category name and code of suicide has remained relatively stable, it is widely accepted that suicide as a cause of death is hidden and underreported for different reasons, notably social and religious attitudes. Moreover, the illegality of suicidal behaviour in some countries contributes to under-reporting and misclassification. According to the WHO, suicides are most commonly found misclassified as “deaths of undetermined intent” and “accidents”.


# Data Sources

The World Health Organization compiles and disseminates data on mortality and morbidity reported by its Member States, according to one of its mandates. This is the main source of cross-country data on suicide rates.

# World Health Organization
  • Data: Total deaths by suicide and suicide rates per 100,000 people overall and by age group and sex
  • Geographical coverage: 198 countries
  • Time span: 1950-2011 for some countries, more recent time periods for most countries
  • Available at: The WHO publishes data here. The same data are also made available via Gapminder.org