Joint Hearing of the Committee on Homeland Security and Governmental Affairs of the U.S. Senate and the Committee on Veterans’ Affairs of the U.S. House of Representatives at 1:00 p.m. CDT.
Witness Testimony of Kara Zivin, Ph.D., Health Services Research and Development, Research Health Scientist, Veterans Health Administration, U.S. Department of Veterans Affairs
Good morning Mr. Chairman, I am honored to provide testimony to the Committee about suicide among veterans treated for depression in the VA Health System. I come before this Committee as a mental health services researcher who has conducted research on this topic. The views and opinions expressed are my own, and do not necessarily represent those of my current employer, the Department of Veterans Affairs, or the views of the VA research community.
I am here today to report on findings from a study that I conducted along with my colleagues at the Department of Veterans Affairs’ National Serious Mental Illness Treatment Research and Evaluation Center (SMITREC) and the VA’s Health Services Research and Development Center of Excellence in Ann Arbor, Michigan. We recently published a paper in the American Journal of Public Health examining suicide rates using data from the VA’s National Registry for Depression for 807,694 veterans of all ages diagnosed with depression and treated at any Veterans Affairs facility between 1999 and 2004.
In all, 1,683 of veterans in VA depression treatment died by suicide during the study observation period, representing 0.21 percent of this treatment population. When we calculated the overall suicide rate in this population over the 5.5 year observation period, it was 88.3 per 100,000 person-years (PY), approximately 7-8 times greater than the suicide rate in the general adult US population. A higher suicide rate would be expected among a population of patients in treatment for depression than the general US population, given that depression is a potent risk factor for suicide. Because most healthcare systems lack the capability of assessing suicide rates among their treatment populations, there are few points of comparisons with non-veteran treatment populations. However, at least one prior study reports a suicide rate for men receiving depression treatment in managed care settings between 1992-1994 of 118 per 100,000 PY, a suicide rate which is somewhat higher than that observed in this veteran depression treatment population.1
In our study, we observed that the predictors of suicide among veterans in depression treatment differed in several ways from those observed in the general US population. Typically, people in the general population who die by suicide are older, male, and white, and have depression and medical or substance abuse issues. In this study, we too found that depressed veterans who had substance abuse problems or a psychiatric hospitalization in the year prior to their index depression diagnosis had higher suicide rates.
However, when we divided depressed veterans into three age groups: 18 to 44 years, 45 to 64 years, and 65 years or older, we found that the younger veterans were at the highest risk for suicide. Differences in rates among depressed veterans of different age groups were striking;18-44-year-olds completing suicide at a rate of 95.0 suicides per 100,000 PY, compared with 77.9 per 100,000 PY for the middle age group, and 90.1 per 100,000 PY for the oldest age group. We did not assess whether individuals had served in combat during a particular conflict, although the existence of a military service-connected disability was considered.
In this VA treatment population, men veterans were more likely to complete suicide than women veterans. Suicide rates were 89.5 per 100,000 PY for depressed veteran men and 28.9 per 100,000 PY for veteran women. However, the differential in rates between men and women (3:1) was smaller than has been observed in the general population (4:1).
We found higher suicide rates for white depressed veterans (95.0 per 100,000 PY) than for African Americans (27.1 per 100,000 PY) and for veterans of other races (56.1 per 100,000 PY). Veterans of Hispanic origin had a lower rate (46.3 per 100,000 PY) of suicide than those not of Hispanic origin (86.8 per 100,000 PY). Adjusted hazard ratios also reflected these differences.
Surprisingly, our initial findings revealed a lower suicide rate among depressed veterans who also had a diagnosis of post-traumatic stress disorder (PTSD) compared to depressed veterans without this disorder. Depressed veterans with a concurrent diagnosis of PTSD had a suicide rate of 68.2 per 100,000 PY compared to a rate of 90.7 per 100,000 PY for depressed veterans who did not also have a PTSD diagnosis. We investigated further to examine whether specific subgroups of depressed veterans with PTSD had higher or lower suicide risks. We found that concurrent PTSD was more closely associated with lower suicide rates among older veterans rather than among younger veterans. This study does not reveal a reason for this lower suicide rate, but we hypothesize that it may be due to the high level of attention paid to PTSD treatment in the VA system, and the greater likelihood that patients with both depression and PTSD will receive psychotherapy and more intensive visits. In general, individuals with depression and PTSD diagnoses have higher levels of VA mental health services use than individuals with depression without PTSD.
Interestingly, depressed veterans who did not have a service-connected disability were more likely to complete suicide than those with a service-connected disability. This may be due to greater access to treatments among service-connected veterans, or more stable incomes due to compensation payments.
We hope our findings will help inform clinical treatment and policy initiatives to reduce suicide mortality among veterans with depression.
Thank you for this opportunity to testify. I will be pleased to answer any questions you may have.
1. Simon GE, VonKorff M. Suicide mortality among patients treated for depression in an insured population. American Journal of Epidemiology. Jan 15 1998;147(2):155-160.