An Examination of Potential Misclassification of Army Suicides: Results from the Army Study to Assess Risk and Resilience in Servicemembers

Debate continues about the accuracy of military suicide reporting due to concerns that some suicides may be classified as accidents to minimize stigma and ensure survivor benefits. We systematically reviewed records for 998 active duty Army deaths (510 suicides; 488 accident, homicide, and undetermined deaths; 2005‐2009) and, using research criteria, reclassified 8.2% of the nonsuicide cases to definite suicide (1), suicide probable (4), or suicide possible (35). The reclassification rate to definite suicide was only 0.2% (1/488). This low rate suggests that flagrant misclassification of Army deaths is uncommon and surveillance reports likely reflect the “true” population of Army suicides.     

A consideration of the validity and reliability of suicide mortality data

The question of the validity and reliability of suicide statistics may be considered at three levels: (1) Are suicide deaths misidentified or differentially identified across jurisdictions or over time? (2) To what degree are suicide deaths misidentified? and (3) Is the degree to which suicides are misidentified sufficient to threaten the validity of research based on suicide statistics? There is general agreement that suicides are likely to be undercounted, both for structural reasons (the burden-of-proof issue, the requirement that the coroner or medical examiner suspect the possibility of suicide) and for sociocultural reasons. There is also substantial anecdotal and empirical evidence suggesting that the mode of death for some true suicides is in fact certified as other than suicide. Overall, it does not seem that very many true nonsuicides are incorrectly certified as suicides. There is not, however, much agreement as to the degree to which true suicides are undercounted. At least some of the inconsistencies in the findings of different investigators arise because the validity of suicide certification seems to vary from place to place. But the source of apparent conflicts in many of the findings is undoubtedly the lack of a "gold standard" against which the verdicts of any given death certification process can be measured. At best, we can estimate that the sensitivity with which coroners and medical examiners certify true suicides varies from approximately 55% to 99%. A central question in estimating the sensitivity of suicide certification is this: What proportion of true suicides are either equivocal or likely to go unsuspected by the coroner or medical examiner? Very little has been done to investigate this issue. Yet the sensitivity of suicide certification clearly varies for equivocal versus unequivocal suicides. As shown in Table 1.2, specificity is also at issue when it comes to certifying equivocal cases. The final question--whether the degree of undercounting of suicide deaths is so great that it threatens the validity of research based on official statistics--is at the crux of the general concern about suicide certification. There are examples of studies in which conclusions based on crude comparisons of reported suicide statistics appear to be invalid. For the most part, these are comparisons among nations with substantially differing death certification procedures. When official statistics are interpreted with a degree of caution and an understanding of the source and direction of biases likely to affect the published rates, however, it seems unlikely that major conclusions based on these statistics will be in error.(ABSTRACT TRUNCATED AT 400 WORDS)     

Suicide surveillance in the U.S. Military--reporting and classification biases in rate calculations

The military has a well-defined population with suicide prevention programs that have been recognized as possible models for civilian suicide prevention efforts. Monitoring prevention programs requires accurate reporting. In civilian settings, several studies have confirmed problems in the reporting and classification of suicides. This analysis evaluated whether suicides were underreported or misclassified under accident or undetermined manner of death in the military system. We reviewed all 1998 and 1999 military deaths using official death reports and compared these data with additional sources, most importantly the DoD Medical Mortality Registry. We assessed for evidence of expressed suicidal intent and past psychiatric history among deaths classified as undetermined and accidents due to gunshot, overdose, drowning, falls, or asphyxia. Using sources other than official records, we found 17% more suicides than were reported, and an additional 4% of deaths that were suspicious for suicide. This study suggests that reporting and classification errors may account for 21% additional suicides in the military. These findings are comparable to rates seen in civilian studies and add to the literature regarding the problems inherent in using administrative death classification data for medical surveillance purposes.     

Musical creativity and suicide

The different abilities involved in artistic creativity may be mirrored by differences among mental disorders prevalent in each artistic profession, taking poets, painters, and composers as examples. Using suicide rates as a proxy for the prevalence of mental disorders in groups of artists, we investigated the percentage of deaths by suicide in a sample of 4,564 eminent artists who died in the 19th and 20th centuries. Of the sample, 2,259 were primarily involved in activities of a linguistic nature, e.g., poets and writers; 834 were primarily visual artists, such as painters and sculptors; and 1,471 were musicians (composers and instrumentalists). There were 63 suicides in the sample (1.3% of total deaths). Musicians as a group had lower suicide rates than literary and visual artists. Beyond socioeconomic reasons, which might favour interpretations based on effects of health selection, the lower rate of suicides among musicians may reflect some protective effect arising from music.     

Suicides in Beijing,China, 1992–1993

Analyzing the raw data of suicides reported to the Beijing Public Security Bureau in 1992 and 1993, this study reveals for the first time to the academic public in the West the rate, gender difference, the timing, causes, and means of Chinese suicides. In comparison with suicide patterns in Western societies, mainly reported by Durkheim (1897/1951) , Diekstra (1990) and the National Center for Health Statistics (1991) , the findings of this study suggest more differences than similarities. The comparatively low suicide rate (4.8 per 100,000 population), the reversed gender effect (55.4% of suicides are female and 44.6% male), alleged causes, and reported means of suicides in China are explained in terms of history, culture, and social forces. However, the direct relationship of warm seasons to Chinese suicide rates is consistent with what is known about Western societies.     

Teen suicide and changing cause-of-death certification, 1953-1987.

This study examines whether the purported tripling in teenage suicides since the 1950s represents a real increase or is simply an artifact of the increased skill of medical examiners in distinguishing youth suicides from fatal accidents. This study examines firearms and poisoning deaths, which together account for 75% of all certified youth suicides, from 1953 through 1987, and concludes: (a) a past undercount of youth suicides is likely, (b) the increase in youth suicide that has occurred is less dramatic than reported and resembles increases in adult suicide, and (c) the suicide increase indicated among youths and adults occurred from 1964 to 1971 and has since stabilized.     

Certification change versus actual behavior change in teenage suicide rates, 1955-1979

National data on firearm suicides and accidental firearm deaths for persons aged 15-19 from 1955 to 1979 were examined to test hypotheses of increases in reported suicide rates as resulting from (1) improved accuracy in the determination and certification of suicide in equivocal firearm deaths; (2) actual increases in the rate of firearm suicides among teens; or (3) some combination of these factors. Data presented support the hypothesis of certification changes as the primary factor influencing suicide rates during the first 12-year segment, but suggest a period of actual growth combined with continued certification change from 1967 to 1979.     

Clusters of Suicidal Events Among Young People: Do Clusters from One Time Period Predict Later Clusters?

We sought to compare clusters of suicidal events between two different time periods and examine the extent to which earlier clusters predict later clusters. We included data on suicides and suicide attempts from New South Wales between July 2001 and June 2012 and Western Australia between January 2000 and December 2011. Suicide attempts included admissions to hospital for deliberate self‐harm and suicides were deaths due to deliberate self‐harm. We combined data on suicides and suicide attempts and grouped them into two equal time periods. We detected clusters in each period using Poisson discrete scan statistics adjusted for socio‐economic status. We estimated the predictive values of earlier clusters on later clusters. The results showed that clusters from earlier time period had a moderate power (36%) in predicting later clusters. During the later time period, some additional cluster areas (14%) were found and some earlier cluster areas subsided (64%). Historical clusters predict 36% of the subsequent clusters, which is probably not sufficient for targeting interventions. Our study highlights the need for other strategies to detect emerging clusters, for example, up‐to‐date data.     

Suicides among family members of elderly suicide victims: an exploratory study

This exploratory study compares elderly suicides with (n=13) and without (n=72) family member suicide. Previous episodes of suicidal behavior were more common among suicides who lost first-degree relatives by suicide (100% vs. 65%, p = .009). Six persons had lost an offspring by suicide prior to their own deaths. Substance use disorder was more prevalent among those with offspring suicide than those without (100% vs. 25%, p = .000). While informants reported that offspring suicides played a central role in the suicide of the elderly study case, sibbling suicides were not considered precipitating factors.     

Rural‐Urban Differences in Suicide Rates for Current Patients of a Public Mental Health Service in Australia

Rural versus urban rates of suicide in current patients of a large area mental health service in Australia were compared. Suicide deaths were identified from compulsory root cause analyses of deaths, 2003–2007. Age‐standardized rates of suicide were calculated for rural versus urban mental health service and compared using variance of age‐standardized rates with 95% confidence intervals. There were 44 suicides and the majority (62%) were rural. Only urban patients used jumping from heights as a method of suicide (4/17; p = 0.02). Rural patients had 2.7 times higher rates of suicide, similar to findings for rural versus urban community suicides and may reflect the underlying community rates, differences in mental health service delivery, or socioeconomic disadvantage.     

An Integrative Suicide Prevention Program for Visitor Charcoal Burning Suicide and Suicide Pact

An integrative suicide prevention program was implemented to tackle an outbreak of visitor charcoal burning suicides in Cheung Chau, an island in Hong Kong, in 2002. This study evaluated the effectiveness of the program. The numbers of visitor suicides reduced from 37 deaths in the 51 months prior to program implementation to 6 deaths in the 42 months post‐implementation period. The number of visitor suicide pacts decreased from 7 pacts (15 individuals) to 1 pact (2 individuals). No statistically significant differences in the numbers of visitor suicide attempts and resident suicides were observed in the two time periods. No statistically significant changes in visitor suicides during the study period were observed on the comparison islands. The consistency and timing of reduction in visitor suicides correlated with the development and delivery of the integrative program on the intervention island, suggesting a causal association between program delivery and reduction of visitor suicides. The possibility of displacement seems small because there was no increase in visitor suicides on the comparison islands during the study period. This integrative approach in preventing target‐specific suicides may serve as an example for other communities to develop suicide prevention programs that make use of the existing local resources.     

Alcohol and suicide death among American Indians of New Mexico: 1980-1998

The relationship between alcohol use prior to suicide was explored among American Indian decedents in New Mexico for the years 1980 through 1998. The suicide data were collected from New Mexico Vital Statistics and toxicology reports from the New Mexico Office of the Medical Investigator and matched on a case-by-case basis. Detailed analyses were undertaken for all cases of resident New Mexico Indians from the Navajo, Pueblo, and Apache cultures. Alcohol was detected in 69% of all suicides of American Indians with some variance by major tribal cultural groups (range = 62.1% to 84.4%). This is higher than in suicides among the overall New Mexico population (44.3%). The mean blood alcohol concentration (BAC) of the drinking Indian decedents at suicide was 0.198 (+/- SD of .088). Mean BACs were high for both males (0.199) and females (0.180) who had been drinking. Over 90% of the Indian decedents who had been drinking had BACs greater than the legal intoxication level of 0.08. The Navajo had the lowest percentage of cases that were alcohol involved, and their mean BAC was lower than the other two cultural groups. Alcohol use for completed suicides also varied somewhat by age, sex, method of suicide, and place of occurrence, but very little by whether the decedent was an on or off reservation resident. Analyses indicated that alcohol use prior to suicide was significantly more associated with male suicides than for females, and it was negatively correlated for those who died by overdose and also those using other drugs at suicide. Otherwise, alcohol use did not significantly differentiate American Indian suicides by age, use of firearms, hanging, use of other methods, or residence, for the presence of alcohol was a factor very commonly associated with all of these variables. Heavy alcohol consumption is, therefore, an important factor in over two thirds of all completed suicides among the Indians of New Mexico.     

Blue Monday Is Real for Suicide: A Case–Control Study of 188,601 Suicides

Many studies have reported that suicides tend to occur on Mondays. However, owing to a lack of controls, conclusive findings on the potential effects of a day of the week on suicides have been lacking. We analyzed public data for causes of death from 1997 to 2015 in the Republic of Korea. Accidental death was used as a control group. The probability of suicide on each day of the week according to age group was calculated. A total of 377,204 deaths (188,601 suicides and 188,603 accidental deaths) were used. The frequency of suicide was highest on Monday and decreased throughout the week until Saturday. Accidental death was highest on Saturday and showed no variations according to weekday. For people in their teens and 20s, the probabilities of suicide on Monday were 9% and 10% higher, respectively, than those on Sunday. As age increased, the differences in suicide probability according to the day of the week were attenuated. The so‐called Blue Monday effect is real, particularly for people in their teens and 20s. Suicide prevention strategies that aim to attenuate the burden and stress of Mondays should be planned.     

Patterns of suicide among hospitalized mentally disordered offenders

The case records of 29 mentally disordered offenders who committed suicide in a state hospital were reviewed. Information was collected on demographic and clinical characteristics; time, method and location of the suicides; and the presence of suicide indicators. Comparisons were made between the characteristics of the suicide patients and those of the current hospital population. Major findings were: 80 percent of the suicides were committed by psychotic patients and 66 percent were by patients with diagnoses which included paranoia; 48 percent of the suicide patients had been charged with or convicted of murder or attempted murder; most suicides occurred in individual patient rooms; nearly 50 percent of the patients had a history of suicide attempts; and other common indicators of suicide risk were often masked by psychotic symptomatology. Several recommendations for identifying and managing potentially suicidal patients in this population are presented.     

Incidence of suicidal ideation and behavior in the United States, 1994.

Completed suicides reflect only a portion of the impact of suicidal behavior; sublethal behaviors cause morbidity and can signal treatable problems such as depression. There is no national quantification of nonlethal suicidal behaviors. The present study used a random-digit-dialed telephone survey to estimate the 12-month incidence of suicidal ideation, planning, and attempts among U.S. adults. Of 5,238 respondents, 5.6% (representing about 10.5 million persons) reported suicidal ideation, 2.7% (about 2.7 million) made a specific suicide plan, and 0.7% (about 700,000) made a suicide attempt (estimate = 1.1 million attempts). Hence, suicidal behaviors are not uncommon and occur along a continuum ranging from ideation to completed suicides. Preventing nonlethal precursor behaviors may prevent deaths.     

Prior Health Care Utilization Patterns and Suicide Among U.S. Army Soldiers

Suicides among U.S. Army soldiers are increasing and, in January 2009, outpaced deaths due to combat. For this study, 1,873 army suicides identified through death, inpatient, and emergency room records were matched with 5,619 controls. In multivariate models, older, male, White, single, and enlisted soldiers with a prior injury (OR = 2.04, 95% CI = 1.64‐2.54), alcohol (OR = 3.41, 95% CI = 2.32‐4.99), or mental health hospitalization (OR = 6.62, 95% CI = 4.77‐9.20) were at increased risk for suicide. Risk was greatest immediately following diagnoses, but remained elevated even after 5 or more years of follow‐up. Most injury hospitalizations were unintentional but, nonetheless, significantly associated with suicide. Interactions indicate soldiers with both mental health and injury history are particularly vulnerable.     

Premature mortality due to suicide, homicide, and motor vehicle accidents in health service delivery areas: comparison of status Indians in British Columbia, Canada, with all other residents

From each of 15 health regions, potential years of life lost (PYLL) before age 75 for Status Indians is compared for select causes of death with all other residents. Mortality data from 1991 to 2001 for rates of PYLL (standardized to the 1991 population) are from tables of the British Columbia Vital Statistics Agency and First Nations and Inuit Health Branch in 2002. PYLL rate differences and rate ratios were compared for two groups with significance of the former indicated by the 95% confidence interval. Overall, the rates of PYLL for suicide, homicide, and deaths due to motor vehicle accidents were about 224%, 340%, and 248% higher among Status Indians than all other residents. Rates of PYLL for homicide and deaths from motor vehicle accidents among Status Indian women exceeded those of other residents who were men. For suicide, Status Indian men ranked first and all other male residents of British Columbia ranked second.     

The Growing Use of Firearms by Suicidal Older Women, 1979–1992: A Research Note

Suicide among older women (65 +) has received very little attention despite increasing numbers of suicides in this population. An examination of national mortality data from the National Center for Health Statistics for the years 1979 through 1992 shows an increasing trend in rates of suicide among older women and a declining trend among women under 65. Over the 14-year period, firearms replaced poisoning as the most prevalent method of suicide by women 65 and over. The results seem consistent with the assertion that the availability, familiarity, and cultural acceptability of firearms may play a role in the choice of suicide method among older women. Although violent death and the use of firearms are generally associated with males in our society, the trends reported here indicate that greater attention to firearm suicides among older women is warranted.