The high rates of suicide and repeat suicide attempts among patients discharged from psychiatric hospitalization are documented in the literature (
1–
3). Studies have shown that most posthospitalization suicides occur during the first few weeks after discharge and that many suicides occur before the first follow-up appointment (
4,
5). Some of the primary factors that may contribute to posthospitalization suicide risk include reduced clinical supervision, noncompliance with aftercare treatment, and brief hospitalizations that do not allow sufficient time for effective treatment of suicidality (
3). A return to stressful conditions, disconnection from care resources, and social isolation may also increase risk (
3,
6).
A study by Valenstein and colleagues (
7) examined suicide rates among 887,859 U.S. military veterans who were in treatment for depression at Department of Veterans Affairs (VA) hospitals between 1999 and 2004. The suicide rate within the first 12 weeks after a psychiatric hospital stay was 568 per 100,000 person-years, which was approximately five times the base rate for the overall treatment population and 54 times the rate for the general U.S. population during the same period. The risk for suicide among active-duty U.S. service members discharged from a psychiatric hospitalization, however, has yet to be examined. The rising rate of suicide (
8) and the increase in psychiatric hospitalizations (
9) in the U.S. military underscore the need to determine risk among service members in psychiatric care so that targeted interventions and prevention programs are implemented.
The purpose of this study was to conduct a retrospective analysis to determine suicide rates among active-duty U.S. service members who had psychiatric hospitalizations at military treatment facilities. Survival analysis was used to examine time-dependent patterns of suicide after hospital discharge.
Methods
The cohort consisted of 68,947 active-duty U.S. service members who had psychiatric hospitalizations at military treatment facilities between January 1, 2001, and December 31, 2011 (the surveillance period). Data on demographic characteristics, personnel data, and medical data were obtained from the Defense Medical Surveillance System, and Medical Expense and Performance Reporting System codes were used to identify psychiatric inpatient facilities within military treatment facilities. Person-level demographic data and primary psychiatric diagnoses (at most recent hospitalization during the surveillance period) were examined for both the entire cohort and the subcohort of persons who died by suicide. For the comparison with the overall population rate, person-years were calculated from the service member’s entry into active service until death, separation from service, or end of the surveillance period.
A survival analysis was conducted with the generalized linear model procedure in SAS, version 9.2. Person-years for this analysis began at hospital discharge. As in other published studies, risk periods (days after release from hospitalization) were delineated as 1–30 days, 31–60 days, 61–90 days, 121–150 days, 151–180 days, 181–365 days, and over one year. Gender, age at hospital admission, service (Air Force, Army, Navy, and Marine Corps), and number of prior psychiatric hospitalizations were treated as covariates. Incidence rates, crude risk ratios, and adjusted risk ratios for each of the risk periods were computed with “over one year after discharge” as the reference group.
This analysis was conducted by duly constituted public health authorities who have been directed to conduct public health surveillance on the incidence and prevalence of suicide and risk factors contributing to suicide in military members. Per regulation (45 CFR 46.101/102), this activity does not constitute research, and institutional review board examination was not required.
Results
Descriptive statistics
Of the 304 deaths in the cohort group, 153 (50%) were from suicide, 93 (31%) were from accidents, 30 (10%) were from natural causes, and 28 (9%) were from combat or homicide or were pending and undetermined. As shown in
Table 1, most service members who died by suicide were white, non-Hispanic, married men in the 20- to 29-year age group. The most frequent primary psychiatric diagnoses among those who died by suicide were adjustment reaction, episodic mood disorder, and depressive disorder not elsewhere classified. Two percent of those who died by suicide had a primary diagnosis of posttraumatic stress disorder.
Suicide rate and survival analysis
The overall suicide rate in this cohort of active-duty U.S. military personnel who had been discharged from a psychiatric hospitalization was 71.6 per 100,000 person-years. The rate of suicide in the general active-duty U.S. military population during 2001–2011 (baseline rate) was 14.2 per 100,000 person-years. Thus suicides among U.S. service members who had had a psychiatric hospitalization were five times more likely than among the general active-duty U.S. military population.
The results of the survival analysis are shown in
Table 2. Suicide rates were highest within the first 90 days after hospital discharge, with the highest rate in the first 30 days. When adjusted for number of hospitalizations, gender, age, and service, the risk of suicide was approximately 8.2 times higher in the first 30 days, five times higher in the 31- to 60-day and 61- to 90-day posthospitalization periods, and twice as high for all remaining risk periods, compared with the risk at more than one year posthospitalization.
Discussion
The results of this study indicate that the risk of suicide was considerably higher among U.S. military personnel who were recently released from psychiatric hospitalizations than in the general active-duty U.S. military population. The suicide rates were highest within 90 days after discharge and especially high within the first 30 days after discharge. These findings are consistent with other published studies that indicate that risk is highest in the 30 days after hospital discharge.
The results of this study highlight the importance of safety planning, continuity of care, and evidence-based prevention interventions both during and after psychiatric inpatient treatment. Several suicide prevention interventions that specifically address suicide risk among psychiatric inpatients are currently being evaluated. Collaborative assessment and management of suicidality (
10), for example, is a structured, collaborative approach to suicide risk assessment, risk reduction, treatment planning, and therapeutic alliance–building that has been adapted and tested for inpatient psychiatric settings (
11). Postadmission cognitive therapy is another intervention that is currently being tested for inpatients who have been hospitalized for a recent suicide attempt (
12). This collaborative brief intervention is administered at individual therapy sessions during hospitalization and also involves collaboration with the unit’s social work team to identify outpatient resources for patients. An intervention that specifically addresses posthospitalization suicide risk is the caring letters concept (
6,
13). The caring letters concept entails sending brief and compassionate correspondence to recently discharged patients and is the only intervention shown in a randomized controlled trial to reduce suicide mortality rates (
14). A caring letters program has been pilot-tested at a large military treatment facility (
15), and a large randomized control trial to investigate this specific program is under way at multiple military and VA hospitals (
6).
Some limitations to this study should be noted. First, the analysis was limited to active-duty military personnel who were treated within the Military Health System (MHS) and did not include service members who were hospitalized outside the MHS network or who were discharged from military service during the surveillance period. Thus the data may underestimate the true rate of suicide in the cohort. Second, the small number of suicide cases limited a more thorough examination of risk based on risk periods, demographic characteristics, and diagnostic categories. Also, rates based on small counts may be unstable and should be interpreted with caution. Future research that includes inpatient treatment encounters from all inpatient facilities (within and outside the MHS network) would provide a larger population to address these issues.
Conclusions
Active-duty U.S. service members who were recently discharged from a psychiatric hospitalization were found to be at high risk of suicide. This study highlights the importance of targeted suicide interventions during and after inpatient stays, as well as close follow-up and monitoring, especially within 30 days of hospital release. Structured transition programs and aggressive follow-up after hospital discharge are recommended.
Acknowledgments and disclosures
Opinions and assertions expressed in this article should not be construed as reflecting official views, policies, or positions of the Department of Defense or the United States Government.
The authors report no competing interests.