Poor social integration and suicide: fact or artifact? A case-control study

BACKGROUND: Sociological studies have shown that poor social integration confers suicide risk. It is not known whether poor integration amplifies risk after adjusting statistically for the effects of mental disorders and employment status. METHOD: A case-control design was used to compare 86 suicides and 86 living controls 50 years of age and older, matched on age, gender, race, and county of residence. Structured interviews were conducted with proxy respondents for suicides and controls. Social integration was defined in reference to two broad levels of analysis: family (e.g. sibship status, childrearing status) and social/ community (e.g. social interaction, religious participation, community involvement). RESULTS: Bivariate analyses showed that suicides were less likely to be married, have children, or live with family. They were less likely to engage in religious practice or community activities and they had lower levels of social interaction. A trimmed logistic regression model showed that marital status, social interaction and religious involvement were all associated with suicide even after statistical adjusting for the effects of affective disorder and employment status. Adding substance abuse to the model eliminated the effects of religious involvement. CONCLUSIONS: The association between family and social/community indicators of poor social integration and suicide is robust and largely independent of the presence of mental disorders. Findings could be used to enhance screening instruments and identify problem behaviors, such as low levels of social interaction, which could be targeted for intervention.     

Suicide and undetermined death in south east Scotland. A case-control study using the psychological autopsy method

BACKGROUND: Mental disorders are major risk factors for suicide. Not all those who suffer from mental disorders kill themselves. Additional information is required to differentiate higher and lesser risk patients. METHODS: Retrospective case-control comparison was made of cases of suicide/undetermined death with living controls using psychological autopsy in South East Scotland. Cases and controls were matched for age, sex and mental disorder. Informants were those closest to cases and controls. The subjects were 45 cases of suicide/undetermined death and 40 living controls. RESULTS: Cases and controls did not differ significantly in severity of mental disorder. The main factors independently associated with undetermined death or suicide were: a history of deliberate self-harm (adjusted OR 4 1); physical ill health (adjusted OR 7.8); and engagement by mental health services (adjusted OR 0.01). Other antecedents associated with increased risk (criminal record, police involvement, financial problems and failure to vote) and those associated with decreased risk (contact with a doctor and in-patient care) did not exert effects after controlling for confounding. CONCLUSIONS: Controls were receiving more care of whatever kind. Treatment of mental disorder comorbid with physical illness and a history of deliberate self-harm may be especially important. Factors that separate those with mental disorder at high risk from those at lesser risk relate to care levels provided, which may be a function of engagement by and with health services. The role of mental health professionals is beneficial in suicide prevention. The focusing of that role towards engaging alienated or 'difficult' patients should be addressed.     

Suicide in current psychiatric in-patients: a case-control study The National Confidential Inquiry into Suicide and Homicide

BACKGROUND: Few controlled studies have investigated factors associated with suicide in current in-patients. We aimed to identify psychosocial, behavioural and clinical risk factors, including variations in care, for in-patient suicide. METHOD: We conducted a national population-based case-control study of people who died by suicide between 1 April 1999 and 31 December 2000 while in psychiatric in-patient care in England. Cases were 222 adult mental health in-patients who died by suicide matched on date of death with 222 living controls. RESULTS: Nearly a quarter of suicides took place within the first week of admission; most of these died on the ward or after absconding. After the first week, however, most suicides occurred away from the ward, the majority of patients having left the ward with staff agreement. Previous deliberate self-harm, recent adverse life events, symptoms of mental illness at last contact with staff and a co-morbid psychiatric disorder were associated with increased risk for suicide. Being off the ward without staff agreement was a particularly strong predictor. Those patients who were detained for compulsory treatment were less likely to die by suicide. Independent predictors of in-patient suicide were male sex, a primary diagnosis of affective disorder and a history of self-harm. Being unemployed or on long-term sick leave appeared to be independently protective. CONCLUSION: Prevention of in-patient suicide should emphasize adequate treatment of affective disorder, vigilance in the first week of admission and regular risk assessments during recovery and prior to granting leave. Use of compulsory treatment may reduce risk.     

Suicide and attempted suicide among older adults in Western Australia

BACKGROUND: Suicide rates are high in later life. Risk factors include male sex and depressive illness. This study investigated the relationship between suicidal behaviour and contact with mental health services among the elderly in Western Australia. METHODS: Record linkage was used to obtain records of hospital admissions and mental health service contacts for all suicide attempts and deaths in the period 1980-95. Standardized incidence ratios were calculated for the elderly, general population and people with mental health service contacts. Cox regression was used to evaluated potential risk factors for elderly people who were in contact with mental health services. RESULTS: People over 60 years of age accounted for 15% of suicides and 4.6% of attempted suicides. Suicide rates were 3.3 times higher in males and 4.4 times higher in females when compared to the general population of elderly people. For attempted suicide, the rate was 5.8 times higher in males and 6.6 times higher in females with prior contact with mental health services. Highest risk of suicide was found in patients with diagnoses of affective psychoses (RR = 3.7), adjustment reaction (RR = 3.2) or depressive disorder (RR = 2.8). The diagnosis of cancer was associated with decreased risk of suicide (RR = 3.6) and attempted suicide (RR = 1.9). CONCLUSIONS: Suicide rates are high among the elderly in Western Australia. Suicide is significantly associated with the diagnosis of mood disorder. Suicide attempts are less common, and are associated most strongly with mood and personality disorders. The decreased risk of self-harm behaviour among patients with cancer warrants further investigation.     

自杀与其它伤害死亡全国性对照研究

目的:了解国内的自杀现状及其特征.方法:用自制的意外死亡研究问卷按"心理解剖"的调查方法从全国20个农村和3个城市疾病监测点抽取895例自杀和701例其它伤害死亡案例.对每个案例的家属及周围知情者单独进行持续约3小时的调查.结果:自杀案例中67%为服毒自杀(大多数为存放在家中的物品);63%自杀当时有精神障碍(39.8%为抑郁症);27%有过自杀未遂;91%从未看过精神科医生;近一年最常见的负性生活事件为经济困难(40%)、患急重病(38%)或夫妻不和(35%);自杀前两天内25%有突出的人际关系冲突;自杀当时10%有他人在现场.与其它伤害死亡者比较自杀者女性多、年龄偏大、读书年限少、家庭收入低并出现慢性和急性负性生活事件多.结论:我国自杀为多因素相互作用的结果,相当一部分属于冲动行为,与其它国家比自杀当时患精神障碍的比例低.     

Life events and psychosocial factors in elderly suicides--a case-control study.

BACKGROUND: Stressful life events, such as family conflicts, separation, bereavement, somatic illness and financial problems are common antecedents of suicide. Studies on suicide among younger persons dominate the literature, despite the fact that a large proportion of suicides occur among elderly persons. METHODS: The occurrence of stressful life events was investigated among elderly suicide cases and population controls. The study was conducted in the southwestern part of Sweden and included 85 persons (46 males and 39 females) 65 years and above who had committed suicide from January 1994 to May 1996. Population controls (84 males and 69 females) were randomly selected. Interviews were carried out with the controls and with informants for the suicide cases. Questions on sociodemographic background, mental and somatic health status, and life events (0-6, 7-12 and 13-24 months preceding suicide/interview) were included in the interviews. RESULTS: Somatic illness, family discord and financial trouble were significant risk factors during all three time periods. Other risk factors were mental disorder, lower education, feelings of loneliness and previous suicide in the family. Factors associated with a decreased risk included active participation in organizations and having a hobby. Variables that remained in the multivariate logistic regression model were mental disorder (men, odds ratio (OR) = 62.4, 95% CI 17.9-217.5; women, OR = 55.9, 95% CI 14.1-222.3) and family discord (men, OR = 10.0. 95% CI 1.7-59.8; women, OR = 9.2, 95% CI 1.9-44.8). CONCLUSIONS: Mental disorder and family discord were the two major risk factors for suicide among elderly men and women.     

Mental healthcare status and psychiatric diagnoses of train suicides

BACKGROUND: The objective of this study is to investigate mental healthcare status and psychiatric disorders in train suicides. METHODS: Data of 4 published train suicide studies were combined with a study of 57 train suicides in The Netherlands. RESULTS: 53% of all train suicides received psychiatric care at the time of suicide, with 49% of them being inpatients. These values are higher than those found in general suicides. When compared to general suicides, functional non-affective psychoses are overrepresented by 25% vs 14%. The percentage of affective disorders approximates that of general suicides namely: 39% vs 42%. "Other diagnoses" are underrepresented by 23% vs 40%. LIMITATIONS: All psychiatric diagnoses were based on clinical data, which may not accurately reflect the patient's psychiatric condition at the time of suicide. CONCLUSIONS: Train suicides receive mental healthcare more often than general suicides and are more often characterized by severe psychopathology. The study indicates that patients with affective and psychotic disorders in particular should be targeted in order to prevent train suicide.     

Acculturation and suicide: a case-control psychological autopsy study

BACKGROUND: The relationships between acculturation and suicide were investigated in East Taiwan. METHODS: Psychological autopsy interviews were conducted for consecutive suicides from two native Taiwanese groups (Atayal and Ami) (N = 30 for each group); each of them was matched with two controls for age, sex and area of residence. The Taiwan Aboriginal Acculturation Scale was used to measure the extent of acculturation. RESULTS: A lower degree of social assimilation was significantly associated with a higher risk of suicide in the Atayal and the male groups. In multivariable regression analysis, a significant effect of low social assimilation on the risk of suicide was found in Atayal and in men, even after controlling for the effects of ICD-10 depressive episode and emotionally unstable personality disorder. Meanwhile, there was a significant trend across low, moderate and high social assimilation on suicide risk in Atayal and in men. CONCLUSIONS: For the native Taiwanese, the stress from rapid acculturation into the main Chinese society is crucial to their mental health. It might be reduced through targeted social and educational programmes.     

自杀未遂者社会心理因素和临床特点对照研究

目的：了解自杀未遂者的社会心理因素及临床特点。方法：本研究对７６名自杀未遂者进行了调查,对其中部分自杀未遂者进行了心理量表评定,并作出了精神病学诊断,以３１名正常人为对照。结果：１自杀未遂者中年轻女性多见;自杀方式绝大多数为服药或服毒;自杀未遂者中一半患精神障碍。２自杀未遂者早年与父母分离比例较高,自杀未遂者早年得到父母情感温暖与理解明显少于正常对照。３自杀未遂者具有焦虑、病态人格和疑心等性格特点。４自杀未遂者的个体素质因素是自杀未遂发生的主要原因。     

富士康深圳工厂12名员工丛集性自杀原因的回顾性分析

目的:对富士康深圳工厂丛集性自杀伤亡者进行心理解剖,分析自杀者的精神状况,探讨连续自杀的原因及相关因素。方法:对2010年1月-5月期间富士康深圳工厂的12例自杀进行回顾性分析,每例至少由2名精神科副主任及以上医师通过查阅询问笔录、现场勘查记录、尸检报告以及访视幸存者和知情人等方法分析,精神疾病诊断标准参考ICD-10精神与行为障碍分类。结果:12名自杀者均为外来劳务工,其中8名为男性,4名女性;入职时间均小于24个月,其中9名小于12个月;3例可初步诊断精神病,3例虽有个别症状但不足以诊断精神疾病,6例可初步排除精神病的诊断;8例自杀前表现了自杀危险信号;自杀原因分析显示3例可能与精神症状有关,3例可能与应激性事件有关,4例可能与心理压力过大、理想破灭绝望等原因有关。结论:富士康深圳工厂员工丛集性自杀的原因复杂,包括个体方面罹患精神疾病、面临较大压力、消极应对方式以及工作倦怠;企业缺乏自杀干预和预防知识,封闭的环境、紧张的气氛,以及媒体的密集报道可能是导致自杀事件发生和扩大的原因。     

The role of alcohol in suicide: a case-control psychological autopsy study

BACKGROUND: The purpose of the study was to estimate the proportion of alcohol abuse and dependence (AAD) among suicides and controls, and to compare the incidence of AAD documented by clinicians with diagnoses derived from a research protocol. METHOD: AAD according to DSM-IV was diagnosed on the basis of interviews with relatives of people who committed suicide and with controls. A total of 427 people who committed suicide during one year were paired by region, gender, age and nationality with controls randomly selected from general practitioners' lists. RESULTS: Alcohol abuse was found in 10% and alcohol dependence in 51% of suicide cases. The corresponding figures for controls were 7% and 14% respectively. AAD was a statistically significant predictor of completed suicides, while abstinence was a significant predictor for female suicides and former use a significant predictor for older male suicides. AAD was diagnosed in 68% of male and 29% of female suicides. Middle-aged (35-59 years) males who committed suicide had the highest risk of alcohol dependence. Among suicide cases only 29% had received a lifetime diagnosis of AAD, against 23% of controls. CONCLUSIONS: AAD was significantly more prevalent among suicides than controls. Overall, the proportion of male suicides affected by alcohol was the same in the present psychological autopsy study as in our previous findings for Estonia on the aggregate level, while the share of female suicides with an AAD diagnosis was dramatically higher on the individual level. AAD is markedly underdiagnosed by general practitioners and clinicians.     

厦门市18岁及以上人群自杀未遂流行病学调查

目的:了解厦门市自杀未遂的流行病学特征及相关因素。方法:采用2010年5月至2010年11月厦门市精神卫生流行病学调查样本,调查18岁及以上厦门市常住人口的自杀未遂发生情况。共调查12051人,回收有效应答问卷10758份。采用扩展的一般健康问卷(GHQ-12)量表作为筛查量表,修订的DSM-IV-TR轴I障碍定式临床检查(SCID-I/P)病人版作为诊断工具。结果:本样本的自杀未遂发生率为0.55%;农村发生率高于城市(0.82%vs.0.47%,P0.05)。多因素logistic回归分析显示:经常性参与赌博(OR=3.18)、精神障碍(OR=2.64)、因精神或心理问题住过院(OR=5.13)是自杀未遂的危险因素。自杀的最主要方式为"采用工具"(42.4%),工具来源主要为"放在家里"(61.1%),自杀最主要的原因为"家庭纠纷"(45.8%),自杀最主要的目的是"为了解脱痛苦"(64.4%)。结论:本研究显示农村自杀未遂的发生率高于城市,经常性参与赌博、精神障碍和因精神或心理问题住过院是自杀未遂主要的危险因素。     

Suicide in female nurses in England and Wales

BACKGROUND: Female nurses appear to have an increased risk of suicide but the reasons are unknown. METHOD: We have concluded a study of nurse suicides (N = 106) in England and Wales, including a psychological autopsy study (N = 42) and case-control comparison with living nurses (N = 84). RESULTS: Nearly three-quarters of the nurse suicides had previous contact with psychiatric services and almost half had been psychiatric in-patients in the past. There were particularly marked differences between the cases and controls for current psychiatric disorder (90.5% v. 7.1%, OR = 68.5), personality disorder (38.1% v. 12%, OR = 32), and history of deliberate self-harm (71.4% v. 2.4%, OR = 58.5). Family background and social factors (especially concerning interpersonal relationships) also distinguished the two groups. Smoking and serious alcohol abuse were much more frequent in the suicides. There was some indication that while many of the suicides were in contact with psychiatric services, care may not have been optimal in some cases. CONCLUSIONS: The most important strategies for suicide prevention in nurses are in prevention, detection and management of psychiatric disorders. In assessing suicide risk a history of DSH and the presence of comorbid psychiatric and personality disorders are particularly important.     

Life problems and physical illness as risk factors for suicide in older people: a descriptive and case-control study

BACKGROUND: The role of physical illness and life problems in contributing to suicide in older people is potentially important with regard to suicide prevention. METHOD: The aim of the study was to determine the life problems other than psychiatric illness contributing to suicide in older people. Semi-structured psychological autopsy interviews, covering life problems and physical illness prior to death, were conducted with informants for 100 people aged 60 years old and over who died through suicide in five English counties. Interviews were completed with informants for 54 age- and sex-matched control subjects who died through natural causes. RESULTS: The three most frequent life problems associated with suicide were physical illness, interpersonal problems, and bereavement. Physical health problems were present in 82% and felt to be contributory to death in 62%. Pain, breathlessness and functional limitation were the most frequent symptoms. Interpersonal problems were present in 55% of the sample and contributory in 31%. The corresponding figures for bereavement-related problems were 47% and 25%. In the case-control analysis, the problems found to be risk factors for suicide were problems related to a bereavement over 1 year before death (OR 3.5, 95% CI 1.2-10.6), and problems with accommodation (OR 5.0, 95% CI 1.1-22.8), finances (p=0.01), and retirement (p=0.02). CONCLUSION: Physical illness, interpersonal problems and bereavement are commonly associated with suicide in older people, but financial, accommodation, retirement and long- term bereavement-related problems may be more specific risk factors.     

Analysis of the seasonal pattern in suicide

BACKGROUND: Suicide has been attributed to social and psychological factors but also to geophysical effects. Of the latter, changes in solar radiation and geomagnetic activities may contribute to the frequency and the seasonal pattern of suicides. METHODS: We studied with a population-based, nationwide analysis all the individuals who committed suicide (n=27,469) in Finland during the period of 1979 to 1999. The daily data on the number of suicides, and the mean and maximum levels of geomagnetic activity were compiled and modelled with Poisson regression using the number of inhabitants in each province as the denominator. Time series analysis of monthly numbers of suicides was carried out using a seasonal-trend decomposition procedure. RESULTS: There was a strong seasonal effect on suicide occurrence (P<0.00001), the risk of suicide being greatest in spring. The seasonal effect was most pronounced when the number of suicides was relatively low. High levels of solar radiation activity were associated with the increased risk of suicide (P=0.00001), but the effect of geomagnetic activity was weak. LIMITATIONS: No individual data on alcohol consumption or mental disorders were available. CONCLUSIONS: Suicide occurrence varies markedly by season and needs attention where prevention is concerned.     

Mental disorders and suicide in Japan: a nation-wide psychological autopsy case-control study

BackgroundThe purpose of the present nationwide psychological autopsy case–control study is to identify the association between mental disorders and suicide in Japan, adjusting for physical conditions.MethodsA semi-structured interview was conducted of the closest family members of 49 suicide completers and 145 gender-, age-, and municipality-matched living controls. The interview included sections of socio-demographic characteristics, physical conditions, and a psychiatric interview producing DSM-IV diagnoses of mental disorders prior to suicide (or at survey). We compared prevalences of mental disorders between the two groups, using conditional logistic regression.ResultsA significantly higher proportion with any mental disorder was found in the suicide group (65.3%) compared to the control group (4.8%) (p = 0.003, odds ratio [OR] = 7.5). The population attributable risk proportion associated with mental disorder was 0.24. Mood disorder, particularly major depressive disorder, was the most strongly associated with suicide (p < 0.001). Anxiety disorder, alcohol-related disorder, and brief psychotic disorder were also significantly associated with suicide (p < 0.05). These patterns were unchanged after adjusting for serious chronic physical conditions.LimitationsThe present study had some limitations, such as small sample size, sampling bias and information bias.ConclusionsMost mental disorders, particularly mood disorder, were significantly associated with a greater risk of suicide in Japan, independent of physical conditions. Mental disorders are a major target of suicide prevention programs in Japan.     

Potential for suicide prevention in primary care? An analysis of factors associated with suicide.

BACKGROUND: General practitioners (GPs) need to be aware of the risk factors for suicide. GP records may provide clues to identifying the relative importance of such risk factors. AIMS: To identify, in suicide cases and matched controls, the patterns of consultation, diagnosis, and treatment of mental illness, and recording of risk factors for suicide. To examine the usefulness of data routinely collected by GPs in computerized databases to investigate treatment of patients in general practice prior to suicide. METHOD: Case control study using GP records from the General Practice Research Database (GPRD). Three controls selected for each case, matched for age, sex, and duration of registration with practice. Information extracted of the prevalence of major disease; diagnosis of, and treatment, or referral for, mental illness; frequency of recording of recent life events; and consultations with the GP in the 12 months prior to death. RESULT: Of the 339 suicide cases recorded, 80% were male, which is similar to the national percentage for this age group. Females were more likely than males to have a history of mental illness and to have been diagnosed and treated for mental illness in the 12 months before death (59% versus 35%), and women were more likely to have previously attempted suicide (47% versus 27%). There was no significant difference between males and females in period of time since last contact with GP practice, but females consulted more frequently. Twenty-nine per cent of cases had not consulted their GP in the six months prior to death. In multivariate analysis, the following were identified as independent risk factors: history of attempted suicide; untreated serious mental illness (odds ratio > 20); recent (past 12 months) marital life event; alcohol abuse; frequent consultations with GP; and previous mental illness. Recording of life events by GPs was poor. CONCLUSIONS: Females at risk of suicide are more likely than males to have been diagnosed and treated for mental illness. It is likely that GPs are under-diagnosed and under-treating males at risk. Data from the GPRD give comparable results to those from other studies. The GPRD is a potentially useful tool for research into relatively uncommon events in general practice.     

Diurnal variations in suicide by age and gender in Italy

BACKGROUND: Recent Italian statistics on suicide distribution by time of day also report data on gender and age of victims, factors which have been shown to influence the seasonal distribution of suicide and which could also affect the influence of biological circadian rhythms on suicidal behaviour. This study aims to identify and evaluate any diurnal variations that may be present in suicide occurrence by age and gender in Italy, considering data from 1994 to 1997. METHODS: The null hypothesis that there are no variations in the distribution of suicides by time of day (or over the three major periods of the day: morning, afternoon, evening/night) was tested with the chi2 goodness-of-fit test and with ANOVA. RESULTS: A clear diurnal variation in the distribution of suicides over time can be observed for both genders, with a peak in the late morning (08:00-11:00 h), and a subsequent decrease to a trough in the night hours. This trend varies with age for both genders: in particular, the age groups 45-64 and 65+ show a clear suicide peak in the morning (08:00-11:00 h), whereas younger people have a peak number of suicides in the late afternoon (16:00-19:00 h). Adults (25-44 years old) show an intermediate trend, with a less pronounced peak between the morning and early afternoon hours. The observed trend is more marked among males; however, the distribution of suicides by time of day is clearly congruent by age between both genders. CONCLUSIONS: Diurnal variation in suicide occurrence by age group may be affected by factors distributed unevenly across age groups. In particular, age distribution of disorders leading to suicidal ideation, and the sensitivity of biological systems of different age groups to environmental cues may affect each group's risk of suicide. Socio-relational factors are also likely to contribute to diurnal variation in suicide risk by age and gender. Children and adolescents can generally be presumed to be at school during the morning, therefore their opportunity for self-harm is restricted to afternoon hours. The elderly, on the other hand, may find themselves alone in the morning, when family and friends spend more time away from home due to daily work activities. LIMITATIONS: Data are based on time of death and not on presumed time of the suicidal act. For suicides committed by certain methods (e.g., poisoning) there may be a considerable difference between time of act and time of death. CLINICAL RELEVANCE: The existence of a temporal window in suicide risk implies an improvement in the surveillance of people at risk of suicide and greater attention to chronobiological factors affecting those suffering from mental disorders leading to suicide ideation.     

Risk factors for suicide in China's youth: a case-control study

BACKGROUND: Suicide is the most common cause of death among youth in China. METHOD: A case-control psychological autopsy study in 23 geographically representative disease surveillance points around China collected information from family members and close associates of 114 persons aged 15-24 years who died by suicide (cases) and 91 who died of other injuries (controls). RESULTS: Among the 114 suicides 61% were female, 88% lived in rural villages, 70% died by ingesting pesticides (most commonly stored in the home), 24% previously attempted suicide, and 45% met criteria of a mental illness at the time of death. Multivariate logistic regression identified several independent risk factors: severe life events within 2 days before death (OR 31.8, 95% CI 2.6-390.6), presence of any depressive symptoms within 2 weeks of death (OR 21.1, 95% CI 4.6-97.2), low quality of life in the month before death (OR 9.7, 95% CI 2.8-34.1), and acute stress at time of death (moderate: OR 3.1, 95% CI 0.8-11.9; high: OR 9.1, 95% CI 1.2-66.8). A significant interaction between mental illness at time of death and gender indicated that diagnosis was an important predictor of suicide in males (OR 14.0, 95% CI 2.6-76.5) but not in females (OR 0.3, 95% CI 0.0-3.6). Prior suicide attempt was related to suicide in the univariate analysis (OR 57.5) but could not be included in the multivariate model because no controls had made prior attempts. CONCLUSIONS: Suicide prevention efforts for youth in China must focus on restricting access to pesticides, early recognition and management of depressive symptoms and mental illnesses, improving resiliency, and enhancing quality of life.     

Suicide and occupation: the impact of socio-economic, demographic and psychiatric differences

BACKGROUND: To explore the risk of suicide associated with occupation while evaluating the impact of socio-economic, demographic and psychiatric differences. METHOD: A nested case-control study with 3195 suicides and 63 900 matched controls. Information on causes of death, occupation, psychiatric admission, marital status and socio-economic factors was obtained from routine registers. RESULTS: Across the 55 occupations investigated, the risk of suicide ranged from 2.73 [95% confidence interval (CI) 1.77-4.22] among doctors to 0.44 (95% CI 0.27-0.72) among architects and engineers compared with primary school teachers. With the exception of doctors and nurses, most of the excess risk of suicide associated with particular occupations is explained by the social and economic characteristics of people in those occupations. Much, but not all, of the excess risk in doctors and nurses is due to their increased use of self-poisoning, a method for which they have the knowledge to use effectively. Occupation has little association with suicide among people who suffer from a psychiatric illness, except for doctors, where the excess risk is 3.62 (p=0.007). CONCLUSIONS: Most of the considerable variation in suicide risk across occupations is explained by socio-economic factors, except for doctors and nurses. Apart from in doctors, the risk of suicide has little association with occupation among people who suffer from a psychiatric illness. Restriction of access to lethal means is an important strategy in suicide prevention.