Suicide Among Vietnam-Era Veterans Risk of … Haden, Lewis, & Spitznagel, in press) have also been linked to increased suicidality among ... Peterson, Dawson, & Chou, 1994), includes

Suicide Among Vietnam-Era Veterans

FOOTNOTES

Author affiliation: 1 Department of Psychiatry, Washington University School of Medicine, St.

Louis, MO, 63108. 2 David M. Ledgerwood is now at the Department of Psychiatry, University

of Connecticut Health Center, Farmington, CT, 06030. 3 Department of Surgery, St. Louis

University Health Sciences Center, St. Louis, MO, 63110. 4 Surgical Service, Department of

Veterans Affairs Medical Center, St. Louis, MO, 63103. 5 Department of Mathematics,

Washington University, St. Louis, MO, 63105.

Reprint address: Rumi Kato Price, Department of Psychiatry, Washington University School of

Medicine, 40 N. Kingshighway, Suite 2, St. Louis, MO 63108-1332.

Risk of Completed and Attempted Suicide in Middle-Aged Vietnam-Era Veterans:

Results from Two National Samples

Rumi Kato Price1, David M. Ledgerwood1,2 , Nathan K. Risk1, Katherine S. Virgo3,4,

Edward L. Spitznagel5, Collins E. Lewis1

Suicide Among Vietnam-Era Veterans

Running Head: Suicide Among Vietnam-Era Veterans

Word Counts: abstract 99; text 4474 (excluding acknowledgments and references)

Tables and Figures: number of tables, 3; number of figures, 0.

MeSH keywords: suicide, suicide attempt, veterans, Vietnam, psychiatric disorders,

substance use disorders



Suicide Among Vietnam-Era Veterans - 2

Abstract

A lingering controversy exists regarding the extent of suicidal risk among Vietnam veterans.

Using two large national data sets, we examined risk for suicidality among Vietnam-era veterans

ages 40 to 60 who currently constitute 23% in this age group of U.S. men. Compared to others of

this age group, the odds ratio was 1.96 for Vietnam-era veterans’ completed suicide, and 1.51 for

attempted suicide. The excess risk appears to result from incremental differences in several

demographic and behavioral factors. A comprehensive intervention targeting physical, mental

health, and substance abuse problems would reduce suicidal risk among aging at-risk Vietnam-

era veterans.




More attention has recently been given to the prevention of adolescent suicide (Gould,

Greenberg, Velting, & Shaffer, 2003; Novick, Cibula, & Sutphen, 2003) and suicide in the

elderly (Ron, 2002; Waern, Rubenowitz, & Wilhelmson, 2003) than to suicide in mid-life.

Recent statistics for the United States show the suicide rate among the 40 to 60 age group in

2001 as 23.14 per 100,000 for males and 6.89 for females (CDC, 2003), and these rates have

been relatively stable over the past three decades (Riggs, McGraw, & Keefover, 1996). Among

males, the rate in 2001 compares to a suicide rate of 16.63 per 100,000 in youth (ages 15-24) and

35.29 per 100,000 in the elderly aged 70 years or older (CDC, 2003). These figures indicate that

the suicide rate for middle-aged men is actually higher than the rate for men in younger age

groups. A critical gap exists in our knowledge about the epidemiology and prevention of suicide

in middle-aged men, perhaps reflective of the reluctance of males in this age group to seek help

(Moller-Leimkuhler, 2003).

Approximately 9.2 million individuals served in the U.S. military during the Vietnam-era

period (August 1964 to April 1975), of whom about 3.4 million were deployed in Vietnam

(Department of Veteran Affairs, 2003). Almost three decades later, Vietnam-era veterans

currently make up about 23% of men in the age group of 40 to 60 years old in the U.S. (CDC,

2003). The estimates of completed suicide among Vietnam-era veterans, however, have varied

among reports (Pollock, Rhodes, Boyle, Decoufle, & McGee, 1990). Evidence for an excess risk

for suicide among Vietnam-era veterans is equivocal (Bullman & Kang, 1995) depending on


many factors including deployment to Vietnam (Watanabe & Kang, 1995), military branch

(Crane, Barnard, Horsley, & Adena, 1997), rank, age group, years since discharge (CDC Vietnam

Experience Study, 1987), and draft eligibility (Hearst, Newman, & Hulley, 1986). It is difficult to

assess accurately the excess risk of suicide among Vietnam-era veterans in part because of a

“healthy worker effect” (Crane et al., 1997). Veteran cohorts tend, in general, to be healthier than

their civilian counterparts because of the selection process for military service as well as the

requirement to maintain a specified level of physical fitness. Indeed, mortality studies of

veterans show that their overall mortality rate is significantly lower than the rates of comparable

civilian groups. This appears to hold true across different eras, genders (Thomas, Kang, &

Dalager, 1991), and nationalities (Crane et al., 1997). Nonetheless, suicide risk appears to be

elevated among selected subgroups of veterans, such as Vietnam veterans with combat exposure

(Farberow, Kang, & Bullman, 1990) or physical injury (Bullman & Kang, 1996), when compared

with other groups of Vietnam veterans or other veterans.

Psychiatric and environmental risk factors for suicidality have been well delineated

through decades of research (Clark & Fawcett, 1992; Blumenthal & Kupfer, 1990; Moscicki,

1997). Furthermore, general population and clinical studies have found common predictors of

both completed suicide and non-fatal suicidality, such as aging (Owens, 1990), not being married

(Petronis, Samuels, Moscicki, & Anthony, 1990), living alone (Owens, 1990), feeling hopeless

(Wetzel, 1976), having a physical illness (Perry, Jacobsberg, & Fishman, 1990), having a

psychiatric illness (Nordentoft & Rubin, 1993), particularly depression (Hintikka et al., 1998;

Kovacs, Goldston, & Gatsonis, 1993), as well as alcoholism (Murphy, 1992; Lewis, Rice,

Andreasen, Clayton, & Endicott, 1985) and drug abuse (Saxon, Aldrich, & Kuncel, 1978; Roy,


2001).

Among veteran populations, psychiatric disorders in general (Fontana & Rosenheck,

1995; Thompson, Katz, Kane, & Sayers, 2000), but more specifically post-traumatic stress

disorder (PTSD) (Kramer, Lindy, Green, Grace, & Leonard, 1994; Drescher, Rosen, Burling, &

Foy, 2003) and depression (Farberow et al., 1990), are associated with both suicide and non-fatal

suicidality. In addition, predictors of suicidality among combat veterans include physical trauma

(Bullman & Kang, 1996) and survivor guilt (Hendin & Hass, 1991; Hyer, McCranie, Woods, &

Boudewyns, 1990). Comorbidity of PTSD, depression, and other psychiatric difficulties appears

to be associated with greater exposure to traumatic events (Engdahl, Dikel, Eberly, & Blank,

1998; Beckham et al., 1998). Both alcoholism (Windle, 1994; Fu et al., 2002) and drugs (Price,

Risk, Haden, Lewis, & Spitznagel, in press) have also been linked to increased suicidality among

Vietnam veterans.

No studies, to our knowledge, have systematically compared veterans with non-veteran

men of the same age group to identify common and unique risk and protective factors that predict

both completed and attempted suicide. The current study analyzes two large general-population

data sets to examine death by suicide and non-fatal attempted suicide among Vietnam-era

veterans between the ages 40 to 60, compared with all other men of the same age group. The first

data set, the 1993 National Mortality Followback Survey (NMFS), contains information about

deceased individuals, and provides information about factors that may have contributed to the

completed suicide (CDC, 1998). The second data set, the National Longitudinal Alcohol

Epidemiologic Survey (NLAES) (Grant, Peterson, Dawson, & Chou, 1994), includes measures of

suicidal ideation and suicide attempts from in-person interviews with individuals (Grant &


Hasin, 1999). Both data sets include large numbers of men between the ages of 40 to 60, and

both surveys were conducted during the same time period. Furthermore, each data set includes a

sufficient number of similar measures of risk and protective factors for suicidality, thus providing

a unique opportunity to compare common and unique factors for completed and attempted

suicide.

The current study has practical and clinical significance not only for the well-being of

veterans but also for their family members and the healthcare system at large. The rate of suicide

in this sizable population of Vietnam-era veterans is expected to increase as they grow older.

Examining the issue of suicide as a possible long-term adverse consequence of war will allow the

mental health system to better understand the needs of military service veterans and prepare for

the potentially increasing morbidity of currently active U.S. service men (NYT, 2004) in the

future.

MATERIALS AND METHODS

Data sets

The 1993 National Mortality Followback Survey (NMFS) collected data on individuals

across the United States who were 15 years and older and who died in 1993. Three linked records

for 22,957 decedents constitute the NMFS data set, including death certificate information, a

proxy respondent questionnaire, and medical examiner records for deaths. Death certificates from

the Current Mortality Sample represented a 10% probability sample of death certificates received

by the National Center for Health Statistics (which in turn was about 10% of all deaths occurring

in the U.S. in 1993) (NMFS, 1998). The proxy interviews were completed with next-of-kin, or, if


next-of-kin was unavailable, another person who was familiar with the decedent. African

Americans and those who died of external causes including suicide were among several

oversampled groups.

The National Longitudinal Alcohol Epidemiologic Survey (NLAES) was conducted in

1992 by the Bureau of the Census under the design and sponsorship of the National Institute on

Alcohol Abuse and Alcoholism. The NLAES is a national probability household survey of the

population aged 18 years and older in the contiguous 48 states. One adult was selected at random

from each household resulting in a representative sample of the civilian, non-institutionalized

population (n = 42,862) (Grant et al., 1994). This large cross-sectional national survey contains

sufficient information on risk and protective factors for non-fatal suicidality (Grant & Hasin,

1999).

The portions of the two data sets available in the public domain without respondents’

identifiers were analyzed. In both data sets, the Vietnam-era veteran category is defined as U.S.

men who served in the military between August 1964 through April 1975. This included both

those who were, and who were not, deployed to Southeast Asia during the Vietnam conflict. The

age group of 40 to 60 years was selected for the current study. This age group represented 76.2%

and 81.3% of the Vietnam-era veterans in the NMFS and NLAES respectively. A total of 2,825

men were selected from the NMFS data set, consisting of 494 (17.5%) Vietnam-era veterans and

2,331 “other” men (all men except Vietnam-era veterans). The sample excluded 19.4% (n = 675)

of the deceased in this age category of men due to missing proxy interviews. The unweighted rate

of completed suicide was 9.0% in the excluded group compared to 8.9% among those included in

the current study. From the NLAES, a total sample of 5,462 men were selected, consisting of


1,277 (23.4%) Vietnam-era veterans and 4,185 “other” men.

We did not include men of the entire age range of Vietnam-era veterans because of the

study’s focus on middle-aged men. Such inclusion criteria would have reduced the advantage of

examining common factors, since predictors of suicidality are substantially different across age

groups (Conwell et al., 1998). The excluded age groups of Vietnam-era veterans in the NMFS

sample were 14.2% who were younger than 40 and 9.6% who were older than 60. In the NLAES,

13.4% were younger than 40 and 5.3% were older than 60. Given the small numbers of Vietnam-

era veterans in these age categories (e.g., n = 62 deceased veterans ages 61 or older in NMFS),

analyses separately of these sub-samples were judged under-powered.

Measures

NMFS. Cause of death was determined by the death certificate. Medical

examiner/coroner reports were gathered for individuals whose cause of death was homicide,

suicide, accident or undetermined (NMFS, 1998). Two non-fatal suicidality questions concerning

the deceased were available from the proxy interviews: communication of suicide intent (“At any

time in the last month of life, did (name) ever talk about taking his own life?”) and morbid

thought (“During the last month of life, did (name) express a wish to die, or say that he wished

that death would come quickly?”). The veteran’s status and a number of variables known to be

associated with suicide were included in the proxy questionnaire.

NLAES. This survey was designed to assess a variety of psychiatric disorders and

substance use disorders, with an emphasis on alcohol-related behaviors, according to the

Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) (APA, 1994).

Three questions were included in the depression section: morbid thought (“During that time


when your mood was at its lowest/you enjoyed or cared least about things, did you ...feel like you

wanted to die?”), suicidal ideation (“... think about suicide from time to time?”) and attempted

suicide ( “... attempt suicide?”). These questions were asked for lifetime conditioned on

depressed mood, but the section also assessed past-year suicidality without depressed mood.

Other variables selected for the current study were also obtained from self-report by the

probands.

Differences between NMFS and NLAES. The outcome measures of the two data sets

were different: while non-fatal suicidality was assessed in both surveys, completed suicide was

assessed only in the NMFS, and attempted suicide only in the NLAES. Furthermore, the NMFS

identified predictors of completed suicide compared to other deaths, whereas the NLAES

identified predictors of attempted suicide in comparison to no suicide attempts.

Ten predictor measures available in both the NMFS and the NLAES were selected after

an initial series of analyses. These included age, race, education, employment, marital status,

living alone, alcohol problems, drug use problems, depression, and health problems. The

measures differed in three important ways. First, the NMFS information was collected from

death documents and proxy interviews, and the NLAES information was obtained from live

probands. Second, the measures used for each survey differed in scope and time frame (Table 1).

The measures of alcohol problems, drug use, and depression were assessed for the past year in

the NMFS, whereas they were assessed for lifetime in the NLAES. Third, because the NMFS

data were based on proxy interviews, precise diagnostic categorization was not possible. Thus,

alcohol problems represented a syndrome approximating alcohol dependence; similarly,

depression represented a syndrome suggestive of major depression. DSM-IV diagnoses of


alcohol dependence and major depression were available from the NLAES data set.

Data analysis

Taking advantage of large sample sizes of the two surveys, prevalence rates of suicidality

were obtained separately for the Vietnam-era veterans and “others.” Because the NMFS survery

contains only the deceased respondents, estimates of the number of suicides were separately

obtained by combining the information from the NMSF data and vital statistics for the entire

population corresponding to the two groups of men in the age group 40 to 60.

Weighted estimates of prevalence rates, and coefficients for logistic regressions were

obtained with standard errors adjusted by the SUDAAN Taylor series linearization method

(Shah, Barnwell, & Bieler, 1996). The differences in prevalence rates of suicidality, as well as

demographics and behavioral characteristics, between the two groups were assessed for each

logistic regression coefficient for Vietnam-era veterans using the Wald P2 statistic. Odds ratios

and confidence intervals for Vietnam-era veterans’ excess risk for suicidality compared to

“other” men of the same age group were also presented. For multivariate logistic regression

analyses, the outcome was completed suicide in the NMFS and lifetime attempted suicide in the

NLAES. The analyses were conducted separately for Vietnam-era veterans and “others.”

Confidence intervals around odds ratios were adjusted with use of the Taylor series method.

Even though the age group of 40 to 60 years old was selected, the age distribution was still

different between Vietnam-era veterans and “other” men and between the two data sets, thus age

by decade was controlled as a covariate in the multivariate analyses. Finer age groups were

determined to be unnecessary based on preliminary analyses. The model fit was assessed by an F

statistic and a C index. The former assesses the extent of the model’s deviation from the null


hypothesis of no fit; the latter assesses the extent of the model’s predictive power which varies

between 0.5 (under the null hypothesis) and 1.0 (perfect prediction). The magnitude of the

differences in odds ratios of predictors were individually assessed using the estimates of

individual beta coefficients and their standard errors. Additionally, analyses were repeated using

the combined two groups to assess the excess risk of suicidality for Vietnam-era veterans

controlling for other predictors.

RESULTS

Increased risk for suicidality among Vietnam-era veterans

NMFS. The estimated proportion of completed suicides among Vietnam-era veterans

ages 40 to 60 in 1993 is 22.64 per 100,000. The estimated proportion of completed suicides in

the “other” men ages 40 to 60 in 1993 is 20.43 per 100,000. These numbers are based upon the

weighted numbers of suicides in the NMSF data for these two categories and the corresponding

total numbers in the population. While the veteran population did have a slightly higher death

rate (ratio = 1.11), the ratio was not significantly different (95% confidence interval (CI) = 0.64,

1.58). The standard error of the ratio was computed using the delta method (Beers, 1957).

Among the deceased men in the age group of 40 to 60, however, 5.6% of Vietnam-era

veterans’ deaths were due to suicide compared to 2.9% of “other” men’s deaths being due to

suicide (Table 2). The odds ratio (OR) was 1.96 (95% CI = 1.23, 3.12), significant at p < 0.05

level. Based on proxy interviews, 11.1% of Vietnam-era veterans communicated about their

suicidal intent in the last month compared to 4.8% of “others” (OR = 2.50, 95% CI = 0.78, 8.03).

The percentage who expressed morbid thoughts in the last month was 25.5% in Vietnam-era


veterans compared to 15.3% among “other” men (OR = 1.89, 95% CI = 0.85, 4.20). Neither

measure of non-lethal suicidality was significantly different between the two groups.

NLAES. In the 40 to 60 age group, 2.7% of Vietnam-era veterans reported an attempted

suicide sometime in their lives compared to 1.8% of “other” men (OR = 1.51, 95% CI = 0.95,

2.39). Suicidal ideation was expressed by 10.7% of Vietnam-era veterans compared to 8.4% of

“other” men (OR=1.31, 95% CI = 1.03, 1.66). Morbid thoughts were expressed by 11.0% of

Vietnam-era veterans compared to 9.1% of “other” men. Only the percentage of patients with

suicidal ideation differed significantly between the two groups.

Demographic and behavioral characteristics

In both the NMFS and NLAES data, Vietnam-era veterans compared to “other” men of

the same age group were younger and more likely to have completed high school. In the NLAES

(but not NMFS), Vietnam-era veterans were also significantly more likely to be employed, to

have a DSM-IV diagnosis of alcohol dependence, and to have a history of illicit drug

experimentation. Vietnam-era veterans, however, reported fewer health problems than other men.

Relative importance of risk and protective factors for completed and attempted suicide

NMFS. Results of logistic regression analyses run separately for Vietnam-era veterans

and “other” men ages 40 to 60 are presented in Table 3. The overall model fit indicates the model

contained important predictors for deceased veterans’ suicide (F = 5.07, df = 10, p < .0001).

Among Vietnam-era veterans, deceased African American veterans were only one-fifth as likely

as non-African American veterans to have died by suicide. Those who lived alone were over

three times more likely to die by suicide. Vietnam-era veterans with alcohol problems were

nearly five times as likely to die by suicide as those without alcohol problems. Depression was


also significantly associated with completed suicide, with depressed Vietnam-era veterans more

than three times as likely to die by suicide as non-depressed Vietnam-era veterans. Health

problems and employment were also significantly associated with completed suicide, but both

were in the opposite direction than expected: employment increased the risk of completed suicide

and health problems were a protective factor for completed suicide. However, the results are

confounded, because most deaths were natural deaths that were positively associated with lack of

employment and health problems, both of which were assessed for the year prior to death.

Predictors of completed suicide among “other” men showed patterns similar to those for

Vietnam-era veterans, with a significant overall model fit (F = 10.68, df = 10, p < .0001). Being

African American was a significant protective factor in that the risk was roughly one-eighth the

risk of non-African Americans. Drug use, even at an experimental level, tripled the risk for

completed suicide as compared with non-drug users. Again, health problems and employment

were significant predictors; however, the results were confounded by a majority of non-suicide

deaths resulting from serious medical illnesses.

NLAES. The overall models are statistically significant for both Vietnam-era veterans

and “other” men (F = 6.05 and F = 10.68 respectively; df = 10, p < .0001 for both). The C index

values of 0.87 and 0.82, respectively, indicate that the predictive power of the NLAES models

was better than in the NMFS models (C = 0.71 for Vietnam-era veterans and C = 0.66 for other

men). Among Vietnam-era veterans, employed veterans were less than half as likely as non-

employed veterans to attempt suicide (association in the expected direction), and married

veterans were significantly less likely to attempt suicide than unmarried veterans. Drug use and

depression were significant risk factors for attempted suicide, placing a veteran at about three and


half times the risk of non-drug using, or non-depressed veterans. Veterans with health problems

were at about five times the risk of attempting suicide as compared to veterans without such

problems.

Results for “other” men were similar for employment, depression and health problems,

which were all significant predictors in this group. However, being an African American was a

significant protective factor in that the risk for attempted suicide was roughly one-sixth the risk

level for non-African Americans. On the other hand, marital status was not a significant

protective factor. While drug use did not appear to increase risk for attempted suicide

significantly, clinical alcohol dependence increased the odds by almost three times.

Unique predictors among Vietnam-era veterans

The levels of risk for completing suicide shown from the NMFS were significantly

different between Vietnam-era veterans and “other” men for alcohol problems (p = 0.04) and

depression (p = 0.02) (Table 3, 5th column). Those factors are significantly more important risk

factors for completed suicide among Vietnam-era veterans in comparison to risk factors for

“other” men. For NLAES, age uniquely affected the risk of attempted suicide. For Vietnam-era

veterans, being 50 years old or older decreased the risk by more than two-thirds, but age did not

make a difference for other men (p = 0.07) (Table 3, far right column). Additionally, being

married and not having health problems were more important protective factors against suicide

attempt (p = 0.05 and p = 0.06, respectively). However, there were no significant differences in

the levels of risk of attempted suicide for any factor.

Excess risk for Vietnam-era veterans controlling for demographic and behavioral factors

Combining both Vietnam-era veterans and “other” men, deceased Vietnam-veterans were


found to be 50% (OR = 1.50) more likely to have died from suicide than deceased “other” men,

controlling for the demographic and behavioral factors assessed separately for the two groups

previously. However, the estimate was non-significant (95% CI = 0.87, 2.59). Excess risk for

suicidal attempts for Vietnam-era veterans was 20% (OR = 1.20), and again the estimate was

non-significant (95% CI 0.73, 1.97).

DISCUSSION

Among men 40 to 60 years old in the U.S., deceased Vietnam-era veterans were more

likely to have died from suicide compared to other deceased men of the same age group.

Significant risk factors for completed suicide among Vietnam-era veterans included living alone,

alcohol problems, and depression. The sole significant protective factor was being African

American. The risk for attempted suicide among Vietnam-era veterans, while elevated, was not

significantly different from the risk of “other” men of the same age group. Significant risk factors

for attempted suicide among Vietnam-era veterans included drug use, depression, and health

problems, whereas significant protective factors were being older, employed, and married.

Predictors for completed and attempted suicide among Vietnam-era veterans were similar to

those of “other” men of this age group. However, stronger associations of alcohol problems and

depression with completed suicide were found for Vietnam-era veterans. Marginal differences

were found for age, marital status, and health problems as predictors of attempted suicide

between the two groups. When controlling for risk and protective factors simultaneously,

Vietnam-era veteran status did not significantly increase the risk for suicide and suicide attempts.

The findings are consistent with other studies of male veterans that did not focus on combat


veterans (e.g., Thoresen, Mehlum, & Moller, 2003).

Employment and health problems in NMFS showed associations with suicide in the

opposite direction than expected. Employment was a risk factor for completed suicide and health

problems was protective against completed suicide. These anomalies are a result of the nature of

the data set. Most of the deceased in the sample died from natural causes. Thus, physical illnesses

leading to unemployment toward the end of life had a stronger association with non-suicidal than

suicide deaths.

Some limitations are inherent in the study of completed suicide and other non-fatal

suicidality. The methodologies for data collection were rather different between the NMFS and

NLAES. To obtain the sample size of Vietnam-era veterans who completed suicide available in

NMFS (estimated n = 27) from a single data set containing both deceased and live subjects, a

survey would have to include about 470,000 live male respondents of this age group at the

baseline to be followed up in the subsequent year, including a search of over 2,800 death

certificates (with a 0.6% death rate). A survey of that scale has never been done in suicide

research, nor is it likely to be conducted in the near future.

The statistically significant finding of the study, an increased risk of completed suicide

among Vietnam-era veterans ages 40 to 60 (OR=1.96; 95% CI = 1.23, 3.12) compared to the

general population of the same age group, needs a cautionary interpretation, because the overall

mortality of the Vietnam-era veterans was likely to be lower than that of “other” men due to a

healthy worker effect. Based on the estimated numbers of suicides in 1993 among men ages 40 to

60, the excess risk of completed suicide among Vietnam-era veterans is estimated to be 16%

higher than that of the of the other men in the general population in the U.S. (but not statistically


significant). Thus, the significantly elevated risk of death by suicide means not so much that the

absolute number of suicides among Vietnam-era veterans is much higher than the number in the

general population. Rather, Vietnam-era veterans disproportionately died by suicide in 1993 even

though they are generally healthier than the rest of the U.S. male population of this age group.

At issue in suicide deaths are the validity of ICD coding (Moyer, Boyle, & Pollock, 1989)

and multiple versus single causes of death (Tardon, Zaplana, Hernandez, & Cueto, 1995;

Mackenbach, Kunst, Lautenbach, Bijlsma, & Oei, 1995). Such limitations are ameliorated by the

use of mixed methodologies (i.e., death certification, autopsy results, and proxy interview) in the

NMFS. Under-reporting is an issue with the measure of non-fatal suicidality. Research has

shown that as many as 30% of those who commit or attempt suicide (Clark & Fawcett, 1992) and

as many as 50% of depressed individuals who die by suicide (Fawcett et al., 1990) do not report

their suicidal intent. Furthermore, there is some indication that uncommunicated suicide

increases with age (Conwell et al., 1998). Nonetheless, under-reporting of suicidality applies to

almost all suicide research.

Neither the NMFS nor the NLAES included measures on combat and related trauma, or

resultant symptoms of PTSD. Thus, it was not possible to examine increased risks for combat

Vietnam veterans. In the current study, the odds ratio of depression for completed suicide was

3.14 among deceased Vietnam-era veterans compared to 0.88 among other deceased men. It is

possible (and likely) that some symptoms assessed by proxies are surrogates of PTSD symptoms,

or depression was exacerbated by PTSD in this sample. Measures of methods used when

attempting suicide were not available in NLAES, thus suicide lethality measures could not be

included in the analyses. A higher risk of death by suicide among Vietnam-era veterans found in


this study may in part be due to their greater familiarity with fire arms.

In summary, limited to men in the age group of 40 to 60 years, this study provides good

evidence for a significantly higher proportion of deaths by suicide and inconclusive results for an

elevated risk for attempted suicide among Vietnam-era veterans, compared to the rest of the male

population. However, our results demonstrate that many of the factors that are characteristics of

completed and attempted suicide among Vietnam-era veterans are similar to those for the general

population of middle-aged men. Only a few statistically significant differences were found

between Vietnam-era veterans and “other” men based on the predictors of suicidality available

across the two data sets. The excess risk of completed and attempted suicide among Vietnam-era

veterans, compared to other males of the age group 40 to 60, appears in large part to have

resulted from incremental differences in several demographic and behavioral factors. When these

factors are controlled, Vietnam-era veteran status was not a significant risk factor. Therefore, an

intervention that places an emphasis on the combination of physical, mental health, and

substance abuse problems could be useful in reducing the suicidality risk among aging Vietnam-

era veterans. Consideration should also be given to a similar strategy for the current generation of

young military men serving in global war zones to reduce the risk of their suicide in years to

come.


ACKNOWLEDGMENTS

This work was supported in part by the Independent Scientist Award (K02DA00221) and

research grant (R01MH60691) to the first author. Editorial assistance by Suzanne Culter and

clerical assistance by Miyuki Ono are both greatly appreciated.


REFERENCES

AMERICAN PSYCHIATRIC ASSOCIATION. (1994). Diagnostic and statistical manual

of mental disorders. Fourth edition (DSM-IV). Washington, DC: American Psychiatric

Association.

BECKHAM, J.C., MOORE, S.D., FELDMAN, M.E., HERTZBERG, M.A., KIRBY, A.C.,

& FAIRBANK, J.A. (1998). Health status, somatization, and severity of posttraumatic stress

disorder in Vietnam combat veterans with posttraumatic stress disorder. American Journal of

Psychiatry, 155, 1565-1569.

BEERS, Y. (1957). Introduction to the theory of error. Reading, MA: Addison-Wesley.

BLUMENTHAL, S.J., & KUPFER, D.J. (Eds.) (1990). Suicide over the life cycle. Risk

factors, assessment, and treatment of suicidal patients. Washington, DC: American Psychiatric

Press.

BULLMAN, T.A., & KANG, H.K. (1995). A study of suicide among Vietnam veterans.

Federal Practitioner, 12, 9-13.

BULLMAN, T.A., & KANG, H.K. (1996). The risk of suicide among wounded Vietnam

veterans. American Journal of Public Health, 86, 662-667.

CENTERS FOR DISEASE CONTROL AND PREVENTION. (1998). National Mortality

Followback Survey - Provisional Data 1993. Hyattsville, MD: National Center for Health

Statistics.

CENTERS FOR DISEASE CONTROL AND PREVENTION. (2003). WISQARS injury

mortality report. Office of Statistics and Programming, National Center for Injury Prevention and

Control. Retrieved from: http://webapp.cdc.gov/cgi-bin/broker.exe.


CENTERS FOR DISEASE CONTROL VIETNAM EXPERIENCE STUDY. (1987).

Postservice mortality among Vietnam veterans. JAMA, 257, 790-795.

CLARK, D.C., & FAWCETT, J. (1992). Review of empirical risk factors for evaluation of

the suicidal patient. In B. Bongar (Ed.), Suicide. Guidelines for assessment, management and

treatment (pp. 16-48). New York: Oxford University Press.

CONWELL, Y., DUBERSTEIN, P.R., COX, C., HERRMANN, J., FORBES, N., &

CAINE, E.D. (1998). Age differences in behaviors leading to completed suicide. American

Journal of Geriatric Psychiatry, 6, 122-126.

CRANE, P.J., BARNARD, D.L., HORSLEY, K.D., & ADENA, M.A. (1997). Mortality of

Vietnam veterans: The veteran cohort study. A report of the 1996 retrospective cohort study of

Australian Vietnam veterans. Canberra: Department of Veterans' Affairs.

DEPARTMENT OF VETERANS AFFAIRS. (2003). America’s wars. Washington, D.C.:

Office of Public Affairs.

DRESCHER, K.D., ROSEN, C.S., BURLING, T.A., & FOY, D.W. (2003). Causes of

death among male veterans who received residential treatment for PTSD. Journal of Traumatic

Stress, 16, 535-543.

ENGDAHL, B., DIKEL, T.N., EBERLY, R., & BLANK, A. JR. (1998). Comorbidity and

course of psychiatric disorders in a community sample of former prisoners of war. American

Journal of Psychiatry, 155, 1740-1745.

FARBEROW, N.L., KANG, H.K., & BULLMAN, T.A. (1990). Combat experience and

postservice psychosocial status as predictors of suicide in Vietnam veterans. Journal of Nervous

and Mental Disease, 178, 32-37.


FAWCETT, J., SCHEFTNER, W.A., FOGG, L., CLARK, D.C., YOUNG, M.A.,

HEDEKER, D., & GIBBONS, R. (1990). Time-related predictors of suicide in major affective

disorder. American Journal of Psychiatry, 147, 1189-1194.

FONTANA, A., & ROSENHECK, R. (1995). Attempted suicide among Vietnam veterans:

A model of etiology in a community sample. American Journal of Psychiatry, 152, 102-109.

FU, Q., HEATH, A.C., BUCHOLZ, K.K., NELSON, E.C., GLOWINSKI, A.L.,

GOLDBERG, J., LYONS, M.J., TSUANG, M.T., JACOB, T., TRUE, M.R., & EISEN, S.A.

(2002). A twin study of genetic and environmental influences on suicidality in men.

Psychological Medicine, 32, 11-24.

GOULD, M.S., GREENBERG, T., VELTING, D.M., & SHAFFER, D. (2003). Youth

suicide risk and preventive interventions: A review of the past 10 years. Journal of the American

Academy of Child and Adolescent Psychiatry, 42, 386-405.

GRANT, B.F., PETERSON, A., DAWSON, D.A., & Chou, P. (1994). Source and accuracy

statement for the National Longitudinal Alcohol Epidemiologic Survey. Rockville, MD: National

Institute on Alcohol Abuse and Alcoholism.

GRANT, B.F., & HASIN, D.S. (1999). Suicidal ideation among the United States drinking

population: Results from the National Longitudinal Alcohol Epidemiologic Survey. Journal of

Studies on Alcohol, 60, 422-429.

HEARST, N., NEWMAN, T.B., & HULLEY, S.B. (1986). Delayed effects of the military

draft on mortality: A randomized natural experiment. New England Journal of Medicine, 314,

620-624.

HENDIN, H., & HAAS, A.P. (1991). Suicide and guilt as manifestations of PTSD in


Vietnam combat veterans. American Journal of Psychiatry, 148, 586-591.

HINTIKKA, J., VIINAMÄKI, H., KOIVUMAA-HONKANEN H.T., SAARINEN, P.,

TANSKANEN, A., & LEHTONEN, J. (1998). Risk factors for suicide ideation in psychiatric

patients. Social Psychiatry and Psychiatric Epidemiology, 33, 235-240.

HYER, L., MCCRANIE, E.W., WOODS, M.G., & BOUDEWYNS, P.A. (1990). Suicidal

behavior among chronic Vietnam theatre veterans with PTSD. Journal of Clinical Psychology,

46, 713-721.

KOVACS, M., GOLDSTON, D., & GATSONIS, C. (1993). Suicidal behaviors and

childhood-onset depressive disorders: A longitudinal investigation. Journal of the American

Acadamy of Child and Adolescent Psychiatry, 32, 8-20.

KRAMER, T.L., LINDY, J.D., GREEN, B.L., GRACE, M.C., & LEONARD, A.C. (1994).

The comorbidity of post-traumatic stress disorder and suicidality in Vietnam veterans. Suicide

and Life-Threatening Behavior, 24, 58-67.

LEWIS, C.E., RICE, J., ANDREASEN, N., CLAYTON, P., & ENDICOTT, J. (1985).

Alcoholism in antisocial and nonantisocial men with unipolar major depression. Journal of

Affective Disorders, 9, 253-263.

MACKENBACH, J.P., KUNST, A.E., LAUTENBACH, H., BIJLSMA, F., & OEI, Y.B.

(1995). Competing causes of death: An analysis using multiple-cause-of-death data from The

Netherlands. American Journal of Epidemiology, 141, 466-475.

MOLLER-LEIMKUHLER, A. (2003). The gender gap in suicide and premature death or:

Why are men so vulnerable? European Archives of Psychiatry and Clinical Neuroscience, 253, 1-

8.


MOSCICKI, E.K. (1997). Identification of suicide risk factors using epidemiologic studies.

In J.J. Mann (Ed.), Psychiatric clinics of North America (pp. 499-517). Washington, DC:

American Psychiatric Press.

MOYER, L.A., BOYLE, C.A., & POLLOCK, D.A. (1989). Validity of death certificates

for injury-related causes of death. American Journal of Epidemiology, 130, 1024-1032.

MURPHY, G.E. (1992). Suicide in alcoholism. New York: Oxford University Press.

NEW YORK TIMES. (2004). US army finds its suicide rate in Iraq higher than among

other G.I.s, March 26, A8.

NORDENTOFT, M., & RUBIN, P. (1993). Mental illness and social integration among

suicide attempters in Copenhagen. Acta Psychiatrica Scandinavica, 88, 278-285.

NOVICK, L.F., CIBULA, D.A., & SUTPHEN, S.M. (2003). Adolescent suicide

prevention. American Journal of Preventive Medicine, 24, 150-156.

OWENS, M.J. (1990). Age and attempted suicide. Acta Psychiatrica Scandinavica, 82, 385-

388.

PERRY, S., JACOBSBERG, L., & FISHMAN, B. (1990). Suicidal ideation and HIV

testing. JAMA, 263, 679-682.

PETRONIS, K.R., SAMUELS, J.F., MOSCICKI, E.K., & ANTHONY, J.C. (1990). An

epidemiologic investigation of potential risk factors for suicide attempts. Social Psychiatry and

Psychiatric Epidemiology, 25, 193-199.

POLLOCK, D.A., RHODES, P., BOYLE, C.A., DECOUFLE, P., & MCGEE, D.L. (1990).

Estimating the number of suicides among Vietnam veterans. American Journal of Psychiatry,

147, 772-776.


PRICE, R.K., RISK, N.K., HADEN, A.H., LEWIS, C.E., & SPITZNAGEL, E.L. (2004).

Post-traumatic stress disorder, drug dependence and suicidal ideation among male Vietnam

veterans with a history of heavy drug use. Drug and Alcohol Dependence (in press).

RIGGS, J.E., MCGRAW, R.L., & KEEFOVER, R.W. (1996). Suicide in the United States,

1951-1988: Constant age-period-cohort rates in 40- to 44-year-old men. Comprehensive

Psychiatry, 37, 222-225.

RON, P. (2002). Depression and suicide among community elderly. Journal of

Gerontological Social Work, 38, 53-71.

ROY, A. (2001). Characteristics of cocaine-dependent patients who attempt suicide.

American Journal of Psychiatry, 158, 1215-1219.

SAXON, S., ALDRICH, S.Y., & KUNCEL, E.E. (1978). Suicide and drug abuse. In D.J.

Lettieri (Ed.), Drugs and suicide. When other coping strategies fail (pp. 167-191). Beverly Hills:

Sage.

SHAH, B.V., BARNWELL, B.G., & BIELER, G.S. (1996). SUDAAN User's Manual,

Release 7.0. Research Triangle Park, NC: Research Triangle Institute.

TARDON, A.G., ZAPLANA, J., HERNANDEZ, R., & CUETO, A. (1995). Usefulness of

the codification of multiple causes of death in mortality statistics. International Journal of

Epidemiology, 24, 1132-1137.

THOMAS, T.L., KANG, H.K., & DALAGER, N.A. (1991). Mortality among women

Vietnam veterans, 1973-1987. American Journal of Epidemiology, 134, 973-980.

THOMPSON, R., KATZ, I.R., KANE, V.R., & SAYERS, S.L. (2002). Cause of death in

veterans receiving general medical and mental health care. Journal of Nervous and Mental


Disease, 190, 789-792.

THORESEN, S., MEHLUM, L., & MOLLER, B. (2003). Suicide in peacekeepers. A

cohort study of mortality from suicide in 22,275 Norwegian veterans from international

peacekeeping operations. Social Psychiatry and Psychiatric Epidemiology, 38, 605-610.

WATANABE, K.K., & KANG, H.K. (1995). Military service in Vietnam and the risk of

death from trauma and selected cancers. Annals of Epidemiology, 5, 407-412.

WAERN, M., RUBENOWITZ, E., & WILHELMSON, K. (2003). Predictors of suicide in

the old elderly. Gerontology, 49, 328-334.

WETZEL, R.D. (1976). Hopelessness, depression and suicide intent. Archives of General

Psychiatry, 33, 1069-1073.

WINDLE, M. (1994). Characteristics of alcoholics who attempted suicide: Co-occurring

disorders and personality differences with a sample of male Vietnam era veterans. Journal of

Studies on Alcohol, 55, 571-577.

TABLE 1

Comparisons of Time Frame and Content of the Measures Between the National Mortality Followback Survey, 1993, and the National

Longitudinal Alcohol Epidemiologic Survey, 1992

NMFSa NLAESb

Time frame Measure content Time frame Measure content

Outcome Past year Completed suicide Lifetimec Attempted suicide

Predictors

Education Time of death No high school diploma Time of interview No high school diploma

Employment Past year Any work at all in the past year Time of interview Employed full-time or part-time

Marital status Time of death Being married Time of interview Being married

Living alone Past year Lived alone Time of interview Living alone

Alcohol

problems

Past year Three or more of alcohol symptoms

related to alcohol dependenced

Lifetime DSM-IV diagnosis of alcohol

dependence

Drug use Past year Any non-prescribed use or overuse of

psychoactive substances, including 11

typese

Lifetime 12 times or more use of illicit drug,

including eight types of psychoactive

substancese

Depression Past year Four or more symptoms suggestive of

clinical major depressionf

Lifetime DSM-IV diagnosis of major

depression excluding the symptom of

suicidality

Health

problems

Lifetime Two or more major health problemsg Lifetime Two or more major health problemsg

Note. Age and race measures are identical across the two surveys.

a The National Mortality Followback Survey, 1993.

b The National Longitudinal Alcohol Epidemiologic Survey, 1992.

c The combined measure includes lifetime assessment of attempted suicide conditioned on depressive mode and/or attempted suicide

in past year without report of depressive mode.

d Available symptoms include frequent drinking, spending a lot of time drunk or hungover, failure to cut down, annoyed by criticism

of drinking, felt guilty about drinking, morning drinking, or advised by a doctor to stop or decrease drinking.

e For NMFS, substances include antidepressants, cocaine, hallucinogens, heroin, marijuana, methadone, pain killers, sedatives,

steroids, stimulants, or tranquilizers; for NLAES, substances include cocaine, heroin, marijuana, methadone, pain killers, sedatives,

stimulants, tranquilizers, and others.

f Available symptoms include change in appetite, drowsy or sluggish, feelings of guilt, having trouble concentrating, long periods of

crying, trouble sleeping, and being withdrawn.

g For NMFS, health problems include Alzheimer’s, cancer, cirrhosis, diabetes, emphysema, and stroke; for NLAES, problems include

arthritis, cancer, cirrhosis, diabetes, epilepsy, emphysema, heart attack, high blood pressure, hepatitis, and stroke.

TABLE 2

Demographic and Behavioral Characteristics of Men 40-60 Years Old in the National Mortality Followback Survey and the National

Longitudinal Alcohol Epidemiologic Survey

NMFSa NLAESb

Vietnam-era

veteran

(n = 494)

Other

(n = 2,331)

Vietnam-era

veteran

(n = 1,277)

Other

(n = 4,185)

Suicidality

Completed suicide (%) 5.6* 2.9* ... ...

Attempted suicide (%) ... ... 2.7 1.8

Communication of suicide intent (%) 11.1 4.8 ... ...

Suicidal ideation (%) ... ... 10.7* 8.4*

Morbid thought (%) 25.5 15.3 11.0 9.1

Demographic

Age (% 50 years old or older) 39.8* 68.9* 19.3* 46.4*

Race (% African American) 21.8 19.2 9.5 9.7

Education (% no high school diploma) 20.7* 38.7* 6.5* 18.0*

Employment (% employed) 48.3 42.5 86.2* 83.1*

Marital status (% currently married) 54.4 53.9 79.7 81.0

Behavioral

Living alone (%) 16.2 20.1 10.9 10.5

Alcohol problems (%)c 35.6 30.0 23.6* 17.4*

Drug use (%)c 10.9 6.4 21.8* 12.3*

Depression (%)c 29.3 30.5 9.1 7.6

Health problems (%)c 23.6 31.9 13.3* 16.2*

Note. The percentages are shown separately for Vietnam-era veterans and others. They are weighted with standard errors adjusted by

the SUDAAN Taylor series method. * indicates p < .05 based on using the Wald P2 statistic testing an unadjusted logistic regression

coefficient for Vietnam-era veterans compared to “others” (two-tail test).

a The National Mortality Followback Survey, 1993, men aged 40 to 60 years old at the time of death.

b The National Longitudinal Alcohol Epidemiologic Survey, 1992, men aged 40 to 60 years old at the time of interview.

c See Table 1 for a description of each of the four behavioral measures.

TABLE 3

Comparing the Relative Magnitude of Associations of Predictors for Suicide and Suicide Attempt Between Vietnam-Era Veterans and

Other Males of Ages 40 to 60 Years Old

NMFSa (outcome = completed suicide) NLAESb (outcome = attempted suicide)

Predictors

Vietnam-era

veteran

(n = 494)

Other

(n = 2,331) p-value for

differences

of beta

coefficients

Vietnam-era

veteran

( n = 1,277)

Other

(n = 4,185) p-value for

differences

of beta

coefficientsOR 95% CI OR 95% CI OR 95% CI OR 95% CI

Age (50 years or older) 0.54 0.20, 1.50 0.60 0.38, 0.96 0.86 0.26* 0.07, 0.95 0.93 0.54, 1.59 0.07

Race (African

American) 0.20* 0.06, 0.66 0.12* 0.06, 0.22 0.47 0.57 0.17, 1.87 0.16* 0.05, 0.57

0.15

Education (no high

school diploma) 0.53 0.15, 1.78 1.43 0.87, 2.36 0.14 1.01 1.28, 3.58 1.23 0.53, 2.85 0.79

Employment (currently

employed) 3.98* 1.42, 11.11 2.01* 1.24, 3.25 0.24 0.43* 0.20, 0.93 0.40* 0.21, 0.75 0.86

Marital status

(currently married) 1.94 0.67, 5.64 1.02 0.57, 1.81 0.30 0.18* 0.05, 0.64 0.72 0.37, 1.38 0.05

Living alone 3.42* 1.16, 10.09 1.42 0.78, 2.57 0.16 0.59 0.17, 2.10 1.04 0.52, 2.07 0.44

Alcohol problemc 4.90* 2.04, 11.76 1.69 1.03, 2.78 0.04 2.05 0.79, 5.33 2.94* 1.68, 5.17 0.51

Drug usec 2.13 0.74, 6.15 3.01* 1.45, 6.26 0.60 3.55* 1.42, 8.86 1.64 0.90, 2.99 0.16

Depressionc 3.14* 1.23, 7.98 0.88 0.49-1.58 0.02 3.78* 1.65, 8.65 4.23* 2.29, 7.81 0.83

Health problemsc 0.10* 0.02, 0.51 .25* 0.14, 0.46 0.29 4.93* 2.12, 11.44 1.90* 1.08, 3.32 0.06

Model fitd F = 5.07

C = 0.71

F = 10.68

C = 0.66

F = 6.05

C = 0.87

F = 10.68

C = 0.82

Note. The adjusted odds ratios (OR) and 95% confidence intervals (CI) are shown separately for Vietnam-era veterans and “others.”

They are weighted with standard errors adjusted by the SUDAAN Taylor series method. * indicates p < .05 for the regression

coefficient for individual logistic regression. A p-value listed in the last column for each data set is the significance level of the t-value

for the difference of two regression coefficients between Vietnam-era veterans and “others” (two-tail test).

a The National Mortality Followback Survey, 1993, men aged 40 to 60 years old at the time of death.

b The National Longitudinal Alcohol Epidemiologic Survey, 1992, men aged 40 to 60 years old at the time of interview.

c See Table 1 for a description of each of the four behavioral measures.

d df = 10, p < 0.0001 for all F statistics. The C index varies from 0.5 under the null hypothesis to the maximum 1.00.

Documents

Suicide Among Vietnam-Era Veterans Risk of … Haden, Lewis, & Spitznagel, in press) have also been linked to increased suicidality among ... Peterson, Dawson, & Chou, 1994), includes