Does PTSD Service Connection Affect Disease Course and Functioning?
Maureen Murdoch MD MPH
Minneapolis VA Health Care System, Minneapolis, MN
Funding Period: October 2003 - September 2006
Men denied Social Security Disability Insurance (SSDI) benefits die at rates about twice that of age-matched controls, tend not to return to gainful employment, and suffer substantially higher impoverishment levels than successful SSDI beneficiaries. Homeless, mentally ill veterans denied SSDI benefits appear to be as ill and as disabled as successful beneficiaries but, once denied benefits, have poorer outcomes (e.g., life quality, homelessness, reliance on welfare and food stamps) than successful beneficiaries. It is unknown if mentally ill veterans denied VA disability benefits might also have poorer outcomes compared to successful claimants.
1) To examine the effects of receiving/not receiving PTSD disability benefits on long-term outcomes, including PTSD symptom severity, social adjustment, functioning, income, employment, health care utilization and costs, homelessness, and mortality. 2) To see if veteran subgroups with historically low rates of PTSD service connection (i.e., women, African Americans, and veterans living in certain regions) are at especially high risk for poor outcomes.
Panel study of 3,337 nationally representative veterans (60% service connected for PTSD; 17% African-American) who applied for VA PTSD disability benefits between 1994 and 1998. Women (51%) were oversampled. Subjects completed surveys about PTSD symptoms and functioning between 1998 and 2000 (Time 1) and between 2004 and 2006 (Time 2). Health care utilization, costs, homeless-ness, and mortality were abstracted from administrative data. The primary analysis considered veterans’ PTSD service-connected (SC) status “fixed” at the time of their original claim. A secondary analysis examined outcomes according to whether veterans SC status changed over time. We used stratification by propensity scores to adjust for PTSD service-connected (SC+) and non-PTSD service-connected (SC-) veterans’ baseline/Time 1 differences.
Of 3,337 original panel members, 339 had died (10%) since 2000. Of those still living, 2,564 (85%) returned Time 2 surveys. Mean reported PTSD symptom severity, social adjustment, and physical functioning at Time 2 did not differ by veterans’ PTSD SC status after adjustment. After adjustment, an estimated 36% of SC+ veterans, compared to 64% of SC- veterans reported low income at Time 2 (p < 0.0001), even though SC+ veterans were almost half as likely as SC- veterans to report working for pay after adjustment (p < 0.0001). While SC+ veterans were more likely than SC- veterans to use VA mental health services at both time points, the proportion of SC+ veterans using such services dropped 10 percentage points by Time 2. Before adjustment, SC+ veterans were less likely than SC- veterans to have ever been identified as homeless in VA administrative data (10% v. 17%); with adjust-ment, the association interacted with veterans’ baseline propensity for service connection and varied according to whether their SC status changed. Mortality did not differ by PTSD SC status before or after adjustment (p = 0.34). There was a regional interaction between PTSD SC status and probability of working at Time 2 (p = 0.002), but no significant race*by*SC status interactions on study outcomes. Original SC+ veterans who later converted to SC- (7%) had slightly less severe Time 2 PTSD symptoms than the 82% of veterans who remained SC+ or the 38% of original SC- veterans who later became SC+, and they had a clinically important trend toward better Time 2 social adjustment (p = 0.07).
This study is the first to examine PTSD SC status’ long-term impact on clinical and functional outcomes. Its most consistent effects were on income and employment. Effects on PTSD symptoms, social adjustment, and identification as homeless were either insignificant, confounded by veterans’ baseline/Time 1 characteristics, or varied according to whether/how SC status changed. Three veteran subgroups with historically lower rates of service connection did not appear to be at unique risk of adverse outcomes. The connections between PTSD SC status, VA health care utilization patterns, and long-term clinical course require more study. Additional research is also needed to understand why some veterans change their PTSD SC status with time
DRA: Mental, Cognitive and Behavioral Disorders, Military and Environmental Exposures, Health Systems
Keywords: Access, Deployment (Pre Gulf War I), Deployment Related, Policy - VA or other, PTSD
MeSH Terms: none