Violence, Alcohol, and Completed Suicide: A Case-Control Study
Publication: American Journal of Psychiatry
Abstract
OBJECTIVE: Violent behavior may represent a risk factor for suicide. The authors tested the hypothesis that violent behavior in the last year of life is associated with completed suicide, even after controlling for alcohol use disorders. METHOD: The authors analyzed data from the 1993 National Mortality Followback Survey, a nationally representative survey conducted by telephone interview with decedents’ next of kin. Data on 753 victims of suicide were compared with data on 2,115 accident victims. Decedents ranged in age from 20 to 64. Dichotomous measures of violent behavior in the past year and history of alcohol misuse were derived by using the four-item CAGE questionnaire. Multiple logistic regression was used to evaluate the interactions of violent behavior with alcohol misuse, gender, and age, respectively, in predicting suicide versus accidental death. Education and race were included as covariates. RESULTS: Violent behavior in the last year of life was a significant predictor of suicide; the relationship was especially strong in individuals with no history of alcohol misuse, those who were younger, and women. CONCLUSIONS: Violent behavior distinguished suicide victims from accident victims, and this finding is not attributable to alcohol use disorders alone. Given that violent behavior increases the risk of suicide, violence prevention initiatives may serve to decrease the risk of suicide as well.
Suicide is the eighth leading cause of death overall in the United States (1). Many have conceptualized suicide as an inherently aggressive act (2), and the role of hostility, aggressive behavior, and anger in completed suicide has been emphasized for most of the last century (3, 4). With the exception of some neurobiological studies (5), much of the research on the interrelationship of suicide and aggression has focused on personality traits and psychiatric diagnoses related to self-reported hostility and aggression. There remains a paucity of data about overtly violent behavior. In this study, we examined the relationship between violent behavior and suicide.
Research suggests a common neurobiology of suicide and other forms of aggressive behavior (6). Lower levels of central serotonin distinguish clinical samples of suicide attempters from nonattempters (7) and are associated with more lethal suicide attempts after psychiatric illness has been controlled for (8, 9). Longitudinal investigations have demonstrated associations of self-reported aggression (10) and hostility (11) with completed suicide.
It is critical to account for alcohol misuse in investigations of the potential role of violence in suicide. For example, alcohol use disorders are strongly related to aggression toward others (12) and suicide (13, 14). Therefore, the apparent link between violent behavior and suicide may be spurious, attributable to the high base rate of alcohol use disorders among suicide victims. However, traits related to aggression are associated with attempted suicide (15–17) and completed suicide (18) among individuals with alcohol use disorders. Therefore, it appears that the relationship between aggression and suicide cannot be ascribed to alcohol misuse alone.
The present study tested the hypothesis that violent behavior in the last year of life is associated with completed suicide, even after alcohol use disorders have been controlled for. Given the greater incidence of violent behavior among younger men (19), we also examined the effects of gender and age on the apparent link between violence and suicide.
Method
Because of their efficiency, retrospective case-control designs are typically employed in studies of rare outcomes or diseases (20). Suicide, with a base rate of 11.4 per 100,000 per year in the United States (21), warrants such a design. The case-control approach assumes that variables distinguishing suicide victims from an appropriate comparison group are likely to be risk factors for suicide (22).
We analyzed data from the 1993 National Mortality Followback Survey (23). The data were released in 1998 with identifying information expunged to facilitate analyses. The 1993 National Mortality Followback Survey represents a 6-year collaborative project between the National Center for Health Statistics and several federal agencies, state and local governments, colleges and universities, and private associations and organizations. The survey was the sixth in a series first initiated by the National Center for Health Statistics in the early 1960s to provide additional information related to mortality in the United States beyond that obtained through the registration of deaths.
The 1993 National Mortality Followback Survey was based on a sample of 22,957 death certificates from 2,215,000 individuals aged 15 years and older who were residing and died in the United States in 1993. Death certificates were drawn from the 1993 Current Mortality Sample, a 10% systematic sample of death certificates received at the National Center for Health Statistics from state vital statistics offices. For each sampled decedent, sex, age, and cause of death were recorded from the death certificate.
To produce more robust analyses of predictors of death from violent causes, the National Mortality Followback Survey oversampled certain causes of death (e.g., homicide, suicide, accidental injury), certain age groups (e.g., individuals who were less than 35 years old at death) and black decedents (23). Trained personnel contacted the decedent’s next of kin (or another person familiar with the decedent’s life history) by telephone or in person and asked them to participate in the survey and provide information on the decedent. Informed written consent was obtained before in-person interviews; verbal consent was obtained before telephone interviews. Questions tapped numerous domains, including mental health and health care utilization.
Subjects
We limited the study to individuals between the ages of 20 and 64 because aggression appears to be more closely associated with suicide in younger than older adults (3). Cause of death was defined by death certificate. Victims of suicide were compared with accident victims. The accidents included motor vehicle accidents (drivers, passengers, motorcycle operators, pedestrians) and poisonings, falls, and other accidents (e.g., drowning or choking).
A total of 3,689 accident victims and 1,363 suicide victims met inclusion criteria. Analyses are based on a subgroup of 2,115 (57%) of the accident victims and 753 (55%) of the suicide victims. Of those included in the analyses, 1,483 (70%) of the accident victims and 517 (69%) of the suicide victims were men; 1,693 (80%) of the accident victims and 653 (87%) of the suicide victims were white, 358 (17%) and 80 (11%), respectively, were black, and 64 (3%) and 20 (3%), respectively, were other minorities. In education, 531 (25%) of the accident victims and 157 (21%) of the suicide victims had less than 12 years of education, 971 (46%) and 329 (44%), respectively, had 12 years, and 613 (29%) and 267 (36%), respectively, had more than 12 years. The remaining accident victims and suicide victims were excluded because of missing data on violence and/or history of alcohol misuse, either because a potential informant could not be contacted, declined to participate, or did not answer one or more of the survey questions regarding these variables. Data were routinely available for demographic variables.
The 653 white suicide victims included in the study represented a higher percentage (57%) of the 1,148 white subjects in the total sample of suicide victims than the 80 black suicide victims (47% of 171) or the 20 suicide victims of other races (45% of 44) (χ2=7.9, df=2, p<0.05). This also held for accident victims: the 1,693 white accident victims represented a higher percentage (59%) of the white subjects in the total sample of accident victims (N=2,878) than the 358 black accident victims (52% of 692) or the 64 accident victims of other races (55% of 116) (χ2=12.5, df=2, p<0.01). The 632 female accident victims included in the study represented a higher percentage of the women in the total sample (61% of 1,030) than the 1,483 men (56% of 2,659) (χ2=9.5, df=1, p<0.01). Likelihood of inclusion in the study did not differ significantly between male (56% of 932) and female (55% of 431) suicide victims. There were no significant age differences in the accident victims or suicide victims that were or were not included in the study.
Measures
Violent behavior in the last year of life was defined dichotomously by response to the question, “During the last year of life, how often did…[the decedent subject] make violent threats or attempts?” Informants were provided with a 4-point scale ranging from often to never. Subjects reported as displaying these behaviors often or sometimes were coded as displaying violent behavior; those reported as displaying these behaviors rarely or never were coded as not displaying violent behavior. The question did not specify the target of violent threats or attempts or if they were directed externally or inwardly. History of alcohol misuse was coded as positive if the informant endorsed one or more items on the CAGE questionnaire, a 4-item screening tool for detecting alcohol use disorders (24); it was coded as negative if no CAGE items were endorsed.
Statistical Analyses
We used multiple logistic regression to predict case status (suicide victims versus accident victims). A correction for extrabinomial variation was included in each analysis (25). When this correction is used, statistical significance is tested by using F statistics rather than the usual chi-square statistics. Independent variables of primary interest were violence in the last year, history of alcohol misuse, gender, and age. Race and education were included as covariates, and all variables were entered simultaneously. We tested the main effects for all variables in the model and interactions of violence with alcohol misuse, gender, and age in predicting case status. In addition to performing the overall significance test for multilevel categories, we also examined the parameter estimates and their associated significance tests to help determine the nature of the different effects. All tests were two-tailed, and alpha was set at 0.05. Case weights, provided in the database (23), were assigned to each case and comparison record to adjust for oversampling and nonresponse.
Results
The history of violence in the past year of the suicide and accident victims as well as the variables (gender, age, history of alcohol misuse) investigated for their interaction with violence in predicting suicide are shown in Table 1. These data are unweighted and thus were not entered in the regression analyses. They are presented in order to provide a concrete indication of the greater level of violence among the suicide victims compared with the accident victims.
Results of regression analyses are shown in Table 2. There were significant interactions between violence and alcohol misuse, violence and gender, and violence and age in predicting suicide. The association of violence and suicide was stronger among individuals without a history of alcohol misuse (Table 1). It was also stronger among women and younger individuals. Suicide risk was also associated with higher education (F=11.7, df=1, 2856, p<0.001), and white race (F=14.9, df=1, 2856, p<0.001).
Discussion
Results indicate that violence in the last year of life is more frequent among suicide victims than accident victims. Alcohol misuse, gender, and age are important considerations in explaining the violence-suicide link.
Violence and Alcohol
The relationship between violence and suicide cannot be attributed entirely to a violent subgroup of alcoholics. Indeed, individuals displaying violent behavior but not in the context of alcohol misuse were especially more likely to take their own lives than to die accidentally. Agitated depression, altered serotonergic function (6), personality disorder characterized by impulsivity and aggression (26–28), and/or interpersonal crises (29, 30) may be driving their violent behavior, placing them at high risk for suicide. Still, violent behavior distinguishes a subgroup of individuals reported to abuse alcohol who are at heightened risk for suicide. These individuals may represent a subtype of alcoholism marked by impulsive and aggressive behavior (31).
Violence and Age
The relationships of younger adults are generally less stable and more prone to violence than those of older adults (32, 33), and this may be especially true of individuals vulnerable to suicide. Relationship disruptions are more likely to precede suicide in younger than in older individuals (29, 30). Risk for spousal homicide (34) and homicide followed by suicide (35) increase during relationship disruptions, often punctuating a pattern of domestic violence (34, 35). Although data on relationship violence and completed suicide are preliminary (36), improved understanding of such patterns may be a key component of suicide prevention, especially in younger adults. Given that the present analyses controlled for age and aggressive behavior, the age-related patterns of aggression and suicide were not entirely attributable to a higher rate of aggressive behavior in younger versus older adults (19).
Violence and Gender
Violence in women appears to confer greater risk for suicide than violence in men. Harris and Barraclough (37) reported a similar phenomenon with respect to alcohol use disorders in women. In a meta-analysis, they calculated higher relative risks for suicide associated with alcohol use disorders in women than men and speculated that “women who break the convention opposing female heavy drinking may have a psychology which greatly increases their suicide risk” (p. 207). This may also apply to women displaying violent behavior. Among individuals with histories of violence, gender does not appear to protect against suicide risk.
Strengths and Limitations
Because of its large sample size and the representative nature of the data, the National Mortality Followback Survey offers a valuable opportunity to examine the interrelationship of violence and suicide and to consider patterns related to age, gender, and alcohol misuse. Proxy respondents have been demonstrated to be a valid source of data in epidemiologic research (38). The CAGE questionnaire is an effective screening tool for detecting alcohol use disorders in clinical (24, 39) and community populations (40). Indeed, the proportion of suicide victims determined to have histories of alcohol misuse in this study according to the CAGE questionnaire (48.5%) is comparable to estimates based on retrospective studies of adult suicides conducted in the United States (41–43). Despite representing a single dichotomous item, violence was a powerful discriminator of suicide and accident victims. We chose accident fatalities to control for sudden death (44). It is important to note that misclassification of driver suicides as accidents is infrequent and has a trivial influence on suicide statistics (45).
The present analyses have limitations. Data are retrospective and primarily based on telephone interviews with informants. Diagnostic data are unavailable. African Americans and other minorities were less likely to be included, raising concerns about the generalizability of our findings. The CAGE questionnaire does not provide information about the recency or severity of alcohol misuse and has traditionally been used as a self-screen rather than with informants. There is no comparison group of living subjects available in the National Mortality Followback Survey. The comparison group for the present analyses, fatal accident victims, may deviate from individuals in the community in personality characteristics, alcohol misuse, and demographic factors (46, 47). Indeed, the high base rate of alcohol misuse in the comparison group likely muted the association of alcohol and suicide. Finally, the degree to which the informants may have included the decedent’s threats of suicide or self-directed violence when asked about violent threats or attempts is not known.
Conclusions
Even after controlling for alcohol misuse, we found that violence is a potent risk factor for suicide. Given that violence and suicide appear to share many common risk factors (4), interventions effective in reducing violence may also reduce the potential for suicide. For example, providing behavioral marital therapy to treated alcoholics appears to be effective in decreasing domestic violence (48). Principal targets of this treatment, alcohol misuse and marital distress and conflict, are also putative risk factors for suicide, and so progress in these areas may also minimize risk for self-harm. Given the apparent risk for suicide among men with histories of domestic violence (36, 49), family violence courts and mandated domestic violence intervention programs may also be promising settings for suicide prevention initiatives. These efforts are likely to have their greatest impact on relatively younger men. The association between violence and suicide in women suggests that there is a need for novel prevention strategies targeting women displaying aggressive behavior. Overall, violence prevention and suicide prevention initiatives can inform one another and stand to gain by coordination.
Footnote
Received June 29, 2000; revisions received Dec. 11, 2000, and March 5, 2001; accepted March 23, 2001. From the Center for the Study and Prevention of Suicide, University of Rochester School of Medicine and Dentistry. Address reprint requests to Dr. Conner, Center for the Study and Prevention of Suicide, University of Rochester School of Medicine and Dentistry, 300 Crittenden Blvd., Rochester, NY 14642; [email protected] (e-mail). Supported by NIMH grants MH-01759, MH-01135, and MH-18911 and a grant from the American Foundation for Suicide Prevention.
References
1.
Centers for Disease Control and Prevention and National Center for Injury Prevention and Control:1997 Leading Causes of Death, United States. Hyattsville, Md, National Center for Health Statistics, 1999
2.
Buie DH, Maltsberger JT: The psychological vulnerability to suicide, in Suicide: Understanding and Responding. Edited by Jacobs D, Brown HN. Madison, Conn, International Universities Press, 1989, pp 59-72
3.
Duberstein PR, Seidlitz L, Conwell Y: Reconsidering the role of hostility in completed suicide: a life-course perspective, in Psychoanalytic Perspectives on Developmental Psychology. Edited by Masling J, Bornstein RF. Washington, DC, American Psychological Association, 1996, pp 257-323
4.
Plutchik R: Outward and inward directed aggressiveness: the interaction between violence and suicidality. Pharmacopsychiatry 1995; 28:47-57
5.
Linnoila VM, Virkkunen M: Aggression, suicidality and serotonin. J Clin Psychiatry 1992; 53:46-51
6.
Mann JJ: The neurobiology of suicide. Nat Med 1998; 4:25-30
7.
Asberg M, Nordstrom P, Traskman-Bendz L: Cerebrospinal fluid studies in suicide: an overview. Ann NY Acad Sci 1986; 487:243-255
8.
Mann JJ, Malone KM: Cerebrospinal fluid amines and higher lethality suicide attempts in depressed inpatients. Biol Psychiatry 1997; 41:162-171
9.
Ninan PT, van Kammen DP, Linnoila M, Bunney WE Jr, Goodwin FK: CSF 5-hydroxyindoleacetic acid levels in suicidal schizophrenic patients. Am J Psychiatry 1984; 141:566-569
10.
Angst J, Clayton P: Premorbid personality of depressive, bipolar, and schizophrenic patients with special reference to suicidal issues. Compr Psychiatry 1986; 27:511-531
11.
Romanov K, Hatakka M, Keskinen E, Laaksonen H, Kaprio J, Rose RJ, Koskenuvo M: Self-reported hostility and suicidal acts, accidents, and accidental deaths: a prospective study of 21,443 adults aged 25-59. Psychosom Med 1994; 56:328-336
12.
Graham K, Leonard KE, Room R, Wild C, Pihl RO, Bois C, Single E: Alcohol and aggression: current directions in research on understanding and preventing intoxicated aggression. Addiction 1998; 93:659-676
13.
Inskip HM, Harris EC, Barraclough B: Lifetime risk of suicide for affective disorder, alcoholism and schizophrenia. Br J Psychiatry 1998; 172:35-37
14.
Murphy GE, Wetzel RD: The lifetime risk of suicide in alcoholism. Arch Gen Psychiatry 1990; 47:383-392
15.
Roy A, Lamparski D, DeJong J, Moore V, Linnoila M: Characteristics of alcoholics who attempt suicide. Am J Psychiatry 1990; 147:761-765
16.
Whitters AC, Cadoret R, Troughton E, Widmer RB: Suicide attempts in antisocial alcoholics. J Nerv Ment Dis 1987; 175:624-626
17.
Windle M: Characteristics of alcoholics who attempted suicide: co-occurring disorders and personality differences with a sample of male Vietnam era veterans. J Stud Alcohol 1994; 55:571-577
18.
Berglund M: Suicide in alcoholism: a prospective study of 88 suicides: the multidimensional diagnosis at first admission. Arch Gen Psychiatry 1984; 41:888-891
19.
Robins LN: The epidemiology of aggression, in Impulsivity and Aggression. Edited by Hollander E, Stein D. Chichester, UK, John Wiley & Sons, 1995, pp 43-55
20.
Gordis L: Epidemiology. Philadelphia, WB Saunders, 1996, pp 124-140
21.
Centers for Disease Control and Prevention: Suicide Deaths and Rates per 100,000. Hyattsville, Md, National Center for Health Statistics, 2000
22.
Hawton K, Appleby L, Platt S, Foster T, Cooper J, Malmberg A, Simkin S: The psychological autopsy approach to studying suicide: a review of methodological issues. J Affect Disord 1998; 50:269-276
23.
Centers for Disease Control and Prevention: National Mortality Followback Survey—Provisional Data, 1993. Hyattsville, Md, National Center for Health Statistics, 1998
24.
Mayfield D, McLeod G, Hall P: The CAGE questionnaire: validation of a new alcoholism screening instrument. Am J Psychiatry 1974; 131:1121-1123
25.
McCullagh P, Nelder JA: Generalized Linear Models. London, Chapman & Hall, 1989
26.
Duberstein PR, Conwell Y: Personality disorders and completed suicide: a methodological and conceptual review. Clin Psychol Sci and Practice 1997; 4:359-376
27.
Rich CL, Runeson BS: Similarities in diagnostic comorbidity between suicide among young people in Sweden and the United States. Acta Psychiatr Scand 1992; 86:335-339
28.
Foster T, Gillespie K, McClelland R, Patterson R: Risk factors for suicide independent of DSM-III-R axis I disorder. Br J Psychiatry 1999; 175:175-179
29.
Heikkinen ME, Isometsä ET, Aro HM, Sarna SJ, Lönnqvist JK: Age-related variation in recent life events preceding suicide. J Nerv Ment Dis 1995; 183:325-331
30.
Rich CL, Warstadt GM, Nemiroff RA, Fowler RC, Young D: Suicide, stressors, and the life cycle. Am J Psychiatry 1991; 148:524-527; correction, 148:960
31.
Zucker RA, Ellis DA, Fitzgerald HE, Bingham CR: Other evidence for at least two alcoholisms, II: life course variation in antisociality and heterogeneity of alcoholic outcome. Dev Psychopathol 1996; 8:831-848
32.
Leonard KE, Senchak M: Prospective prediction of husband marital aggression within newlywed couples. J Abnorm Psychol 1996; 105:369-380
33.
Straus MA, Gelles RJ: Societal change in family violence from 1975 to 1985 as revealed by two national surveys. J Marital Fam Ther 1986; 48:465-479
34.
Wilson M, Daly M: Spousal homicide risk and estrangement. Violence Vict 1993; 8:3-16
35.
Morton E, Runyan CW, Moracco KE, Butts J: Partner homicide-suicide involving female homicide victims: a population-based study in North Carolina, 1988-1992. Violence Vict 1998; 13:91-106
36.
Conner KR, Duberstein PR, Conwell Y: Domestic violence, separation, and suicide in young men with early onset alcoholism: re-analyses of Murphy’s data. Suicide Life Threat Behav 2000; 30:354-359
37.
Harris EC, Barraclough B: Suicide as an outcome of mental disorders. Br J Psychiatry 1997; 170:205-228
38.
Nelson LM, Longstreth WT Jr, Koepsell TD, van Belle G: Proxy respondents in epidemiologic research. Epidemiol Rev 1990; 12:71-86
39.
King M: At risk drinking among general practice attenders: validation of the CAGE questionnaire. Psychol Med 1986; 16:213-217
40.
Cherpitel CJ: Screening for alcohol problems in the US general population: a comparison of the CAGE and TWEAK by gender, ethnicity, and services utilization. J Stud Alcohol 1999; 60:705-711
41.
Conwell Y, Duberstein PR, Cox C, Herrmann JH, Forbes NT, Caine ED: Relationships of age and axis I diagnosis in victims of completed suicide: a psychological autopsy study. Am J Psychiatry 1996; 153:1001-1008
42.
Rich CL, Fowler RC, Young D: Substance abuse and suicide: the San Diego study. Ann Clin Psychiatry 1989; 1:79-85
43.
Robins E: The Final Months: A Study of the Lives of 134 Persons Who Committed Suicide. New York, Oxford University Press, 1981
44.
Clark DC, Horton-Deutsch S: Assessment in absentia: the value of the psychological autopsy method for studying antecedents of suicide and predicting future suicides, in Assessment and Prediction of Suicide. Edited by Maris RW, Berman AL, Maltsberger JT, Yufit RI. New York, Guilford, 1992, pp 144-182
45.
Ohberg A, Penttila A, Lonnqvist J: Driver suicides. Br J Psychiatry 1997; 171:468-472
46.
Mann RE, Smart RG, Anglin L: Alcohol-related measures as factors in traffic fatalities. J Stud Alcohol 1996; 57:646-651
47.
Tsuang MT, Boor M, Fleming JA: Psychiatric aspects of traffic accidents. Am J Psychiatry 1985; 142:538-546
48.
O’Farrell TJ, Van Hutton V, Murphy CM: Domestic violence before and after alcoholism treatment: a two-year longitudinal study. J Stud Alcohol 1999; 60:317-321
49.
Stawar TL: Suicidal and homicidal risk for respondents, petitioners, and family members in an injunction program for domestic violence. Psychol Rep 1996; 79:553-554
Information & Authors
Information
Published In
History
Published online: 1 October 2001
Published in print: October 2001
Authors
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).