Abstract
Introduction:
Mandatory, age-based re-evaluations for posttraumatic stress disorder (PTSD) service connection contribute substantially to the Veterans Benefits Administration’s work load, accounting for almost 43% of the 168,013 assessments for PTSD disability done in Fiscal Year 2017 alone. The impact of these re-evaluations on Veterans’ disability benefits has not been described.
Materials and Methods:
The study is an 18-year, ecological, ambispective cohort of 620 men and 970 women receiving Department of Veterans Affairs PTSD disability benefits. Veterans were representatively sampled within gender; all were eligible for PTSD disability re-evaluations at least once because of age. Outcomes included the percentage whose PTSD service connection was discontinued, reduced, re-instated, or restored. We also examined total disability ratings among those with discontinued or reduced PTSD service connection. Subgroup analyses examined potential predictors of discontinued PTSD service connection, including service era, race/ethnicity, trauma exposure type, and chart diagnoses of PTSD or serious mental illness. Our institution’s Internal Review Board reviewed and approved the study.
Results:
Over the 18 years, 32 (5.2%) men and 180 (18.6%) women had their PTSD service connection discontinued; among them, the reinstatement rate was 50% for men and 34.3% for women. Six men (1%) and 23 (2.4%) women had their PTSD disability ratings reduced; ratings were restored for 50.0% of men and 57.1% of women. Overall, Veterans who lost their PTSD service connection tended to maintain or increase their total disability rating. Predictors of discontinued PTSD service connection for men were service after the Vietnam Conflict and not having a Veterans Health Administration chart diagnosis of PTSD; for women, predictors were African American or black race, Hispanic ethnicity, no combat or military sexual assault history, no chart diagnosis of PTSD, and persistent serious mental illness. However, compared to other women who lost their PTSD service connection, African American and Hispanic women, women with no combat or military sexual assault history, and women with persistent serious illness had higher mean total disability ratings. For both men and women who lost their PTSD service connection, those without a PTSD chart diagnosis had lower mean total disability ratings than did their counterparts.
Conclusions:
Particularly for men, discontinuing or reducing PTSD service connection in this cohort was rare and often reversed. Regardless of gender, most Veterans with discontinued PTSD service connection did not experience reductions in their overall, total disability rating. Cost-benefit analyses could help determine if mandated, age-based re-evaluations of PTSD service connection are cost-effective.
Keywords: Posttraumatic Stress Disorder, Gender, Cohort Studies, Veterans Disability Claims, Compensation
Introduction
With a budget of nearly $177 billion per year,1 the Department of Veterans Affairs’ (VA) Veterans Benefits Administration (VBA) oversees the United States’ third largest federal disability program. Veterans receiving compensation for disorders incurred or aggravated by their military service are said to be service connected for those disorders. The degree to which Veterans are disabled for these conditions is called their service connected disability rating. Disability ratings range from 0% (non-disabling) to 100% (totally disabled). Veterans may be service connected for multiple disorders, but their total disability rating can never exceed 100%. Higher ratings are associated with greater compensation and other benefits.
Of the 4.4 million Veterans indemnified by VA for military-related disorders in 2016, almost 20% were service connected for posttraumatic stress disorder (PTSD).2 Because PTSD is treatable and potentially curable,3–6 one might expect that at least some Veterans’ PTSD service connection would change over time. For example, if their symptoms lessened substantially, their service rating might decrease commensurately; if they recovered, they might not require PTSD service connection at all. By law, Veterans service connected for potentially nonpermanent conditions, such as PTSD, must undergo re-evaluations every two to five years.7 At these re-evaluations, Veterans’ compensation may be adjusted upward, downward, or terminated, depending on their clinical status. The re-evaluation requirement ends after 5 years for Veterans with no improvement. Veterans aged 55 and older are also exempt.8
Re-evaluations for PTSD service connection contribute substantially to the VBA’s work load. Of the 168,013 assessments for PTSD disability conducted by VBA in Fiscal Year 2017, almost 43% were re-evaluations (“Number of C and P Exams by Category,” Veterans Health Administration (VHA) Support Service Center, https://securereports2.vssc.med.gove/ReportServer, accessed Oct 24, 2017). Many Veterans find these re-evaluations stressful and disruptive.9,10 They may be asked to undergo new and often emotionally upsetting forensic evaluations, for example, and every re-evaluation carries with it the risk of monetary loss. Veterans who lose their service connection or experience a downgrade in their disability rating have the right to appeal those decisions. Such appeals are also time-consuming to the Veteran and add further still to VBA’s workload.
The burdens associated with PTSD re-evaluations could be acceptable if the process identified substantial numbers of veterans who no longer needed compensation for PTSD or whose disability ratings could be appropriately reduced. Then the costs and potential harms associated with the program might be counterbalanced by the amount of money freed up to sponsor other programs. If only small numbers of Veterans discontinued or reduced their disability ratings, however, then it might make sense to re-visit the laws mandating routine re-evaluations.
In the present paper, we examine changes in the service connected status of Veterans who applied for PTSD disability benefits between 1994 and 1998 and who were potentially eligible for PTSD disability re-evaluations because of their age. Specifically, we examine the following markers of re-evaluation efficiency: 1) the percentage of Veterans whose PTSD service connection was completely discontinued or whose PTSD disability rating was reduced, 2) the percentage of reversals –that is, the percentage of individuals who had their PTSD service connection re-instated after a discontinuation or had their PTSD disability rating restored after a reduction, and 3) substitutions: If PTSD discontinuations or disability rating reductions were counterbalanced by new or increased ratings for other disorders, then Veterans’ total disability rating would not decrease, and there would be no net savings to the system.
Our second objective was to identify demographic and clinical variables associated with discontinued PTSD service connection. If found, there could be many possible explanations for subgroup differences in discontinued PTSD service connection, including variations in disease severity, differing odds of receiving or responding to evidence-based treatments, or uneven processing during re-evaluations. Identifying subgroups with particularly high or low odds of predictors of discontinued versus continued PTSD service connection would therefore allow us to focus future endeavors on understanding those differences, with an end goal of improving the processes, outcomes, and equity of VA disability policies.
Methods
The study is an ecological, ambispective cohort (ambispective = retrospective and prospective elements).
Participants
Between January 1994 and December 1998, 100,750 male Veterans and 3,866 female Veterans applied for, but did not necessarily receive, PTSD service connection. We randomly selected 2,700 men and 2,700 women (total N = 5,400) from this population to participate in a survey examining disparities in PTSD disability awards11–13. The inception cohort are the 1,654 men and 1,683 women who responded. Service connection outcomes were assessed as of Dec 31, 1999 (Time 1), April 30, 2007 (Time 2), and August 24, 2016 (Time 3). The present study is limited to cohort members who had ever been eligible for a PTSD re-evaluation because of age and were alive at Time 3. Specifically, 620 men and 970 women who were service connected for PTSD at Time 1 and under age 55 at Time 1, compensated for PTSD at Time 2 and under age 55 at Time 2, or both are included.
Service Connection Outcomes
The primary outcome was discontinued PTSD service connection as of August 24, 2016 (Time 3). The final PTSD service connection discontinuation rate is the number of men or women with discontinued PTSD service connection at Time 3, divided by all the men or women in the sample. The total PTSD service connection discontinuation rate is the total number of men or women whose PTSD service connection had ever been discontinued, divided by all the men or women in the sample.
Secondary outcomes included reduced PTSD disability ratings, reversals, and substitutions. The final PTSD disability rating reduction rate is the number of men or women with a reduced PTSD disability rating by Time 3, divided by all the men or women in the sample. The total PTSD disability rating reduction rate is the total number of men or women who ever had a reduced PTSD disability rating, divided by all the men or women in the sample.
In terms of reversals, we examined PTSD service connection that was reinstated after a discontinuation and PTSD disability ratings that were restored (or even increased) after having been reduced. The reinstatement rate for PTSD service connection is the number of men or women with reinstated PTSD service connection at Time 3 divided by all the men or women with discontinued PTSD service connection at Time 2. The restoration rate for reduced PTSD disability ratings is the number of men or women whose PTSD disability rating was restored (or increased) at Time 3 divided by all the men or women with reduced PTSD disability ratings at Time 2.
Total disability ratings, also called the combined degree of service connection in VA parlance, incorporates all mental and physical conditions for which the Veteran is service connected and can never exceed 100%. If Veterans’ PTSD service connection was discontinued or their PTSD disability ratings were reduced, but their combined degree of service connection remained the same (or increased), we considered that a substitution. For a Veteran’s combined degree of service connection to stay the same or increase in this setting, he or she would have had to become service connected for some new disorder at a rating similar to (or higher than) their former PTSD rating, or they would have had to experience an increased rating in one of their other, pre-existing service connected conditions.
Potential Predictors
We examined the following potential predictors of discontinued PTSD service connection: age, race, gender, service era, combat and military sexual trauma exposures, married status, and having a chart diagnosis of PTSD or of persistent serious mental illness. Except for serious mental illness, all these variables are predictive for initial awards for PTSD service connection;11,13 we therefore anticipated their effects would persist with time. Comorbid serious mental illness’ impact on PTSD service connection has not been studied. However, it is strikingly prevalent in this population14 and can impede PTSD treatment.15 We anticipated that comorbid serious mental illness — defined here as bipolar disorder, schizoaffective disorder, or schizophrenia — would reduce Veterans’ odds of discontinuing PTSD service connection.
Data Sources
Veterans age, service era (categorized as during the Vietnam Conflict versus later), and service connection information at Time 1 were obtained using the Veterans Issues Tracking Adjudication Log (VITAL). We used the VA’s Beneficiary Identification Records Locator Subsystem (BIRLS) to collect service connection information at Time 2 and the Veterans Services Network Corporate Mini Master File (VETSNET File) to collect service connection information at Time 3. At Time 2, we used International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9 CM) codes housed in the VA National Patient Care Database to determine whether cohort members had been diagnosed with PTSD or with bipolar disorder, schizoaffective disorder, or schizophrenia between 1994 and 2006. We categorized veterans as having “persistent serious mental illness” if they had been diagnosed with bipolar disorder, schizoaffective disorder, or schizophrenia at least once in 3 separate calendar years.
Veterans’ self-reported married status and race were assessed at inception via survey using single-item questions. Respondents self-identified their race/ethnicity by selecting as many categories as applied from the following: (1) American Indian or Alaskan Native, (2) Asian, (3) Black or African American, (4) Hispanic or Latino, (5) Native Hawaiian or Other Pacific Islander, (6) White (7) Other. Also at inception, we assessed self-reported combat exposure using a modified, 22-item version of the Combat Exposure Index16 and military sexual assault history using the 3-item “criminal sexual misconduct” subscale from the Sexual Harassment Inventory17 plus a fourth question asking about in-service sexual assault unrelated to work. Combat and military sexual assault were dichotomized into “any” versus “none” variables then combined into a “trauma history scale,” where a history of both combat and sexual assault = 2, a history of one or the other = 1, and a history of neither = 0. The survey’s details have been reported elsewhere.11
Analysis
Because men and women were sampled independently, outcomes are reported separately by gender. We examined predictors of our main outcome, discontinued service connection for PTSD at Time 3, using one-way ANOVA, χ2 tests, or binary logistic regression. To control for potential confounding among predictors, variables associated with the primary outcome at p ≤ 0.10 on univariable analysis were included simultaneously in a binary logistic regression model. There was near-perfect collinearity between women selecting white race and not selecting black or African American or Hispanic categories. Therefore, only African American or black race and Hispanic ethnicity were entered into multivariable regression, both coded as dummy variables. In both genders, age and service during the Vietnam Conflict were highly correlated (rs > 0.80), so only service era (service during Vietnam = 0 and after Vietnam = 1) was included in logistic models. Once we identified the statistically significant predictors of discontinued PTSD service connection, we then examined the total disability ratings (combined degree of service connection) of Veterans whose PTSD service connection had been discontinued at Time 3 according to those predictors. This allowed us to better understand the impact of losing PTSD service connection on Veterans’ total compensation package according to these risk characteristics and to see if substitution patterns differed by these characteristics. The 1 male and 8 female Veterans who lost all service connection at Time 3 were given total service ratings of zero for this analysis. We used SPSS Version 19 for all analyses.
Results
Descriptive statistics
At inception, the sample’s men had average age 46.8 (±5.4) years; women’s average age was 35.0 (±8.4) years. By Time 3, men’s mean age was 66.6 (±5.4) years, and women’s, 54.8 (±8.4) years. Table 1 shows the sample’s descriptive statistics stratified by gender. At inception, more than 99% of the men and women with PTSD service connection were under age 55. By Time 2, only 12.4% of men with PTSD service connection were under age 55, compared to 86.8% of the women. Besides older age, men were substantially more likely than the women to have served during the Vietnam Conflict, to have combat exposures, to be married, and to have received a VA chart diagnosis of PTSD. Women were more likely to have experienced military sexual assault and to have a diagnosis of persistent serious mental illness.
Table 1.
Sample Characteristics, Stratified by Gender
| Characteristic | Men N = 620 | Women N = 970 | p-value |
|---|---|---|---|
| Under age 55 at inception | 99.7 | 99.8 | 0.65 |
| Under age 55 at Time 2 | 12.4 | 86.8 | <0.001 |
| Race/Ethnicity | |||
| White | 73.5 | 76.2 | 0.24 |
| Black or African American | 14.5 | 15.6 | 0.57 |
| Hispanic | 7.6 | 4.8 | 0.02 |
| Other | 6.6 | 6.4 | 0.86 |
| Served during Vietnam Conflict | 89.8 | 19.4 | <0.001 |
| Married | 60.8 | 36.3 | <0.001 |
| Any history of combat exposure | 94.2 | 28.6 | <0.001 |
| Any history of military sexual assault | 5.2 | 74.2 | <0.001 |
| Military trauma history scale score | <0.001 | ||
| 2 (Combat and sexual assault) | 3.4 | 18.8 | |
| 1 (Combat or sexual assault) | 92.6 | 65.3 | |
| 0 (Neither combat nor sexual assault) | 4.0 | 16.0 | |
| Any PTSD diagnosis, 1994–2006 | 93.2 | 88.5 | 0.002 |
| Persistent serious mental illness, 1994–2006 | 8.9 | 20.9 | <0.001 |
Results are reported as column percentages (%).
PTSD = posttraumatic stress disorder
PTSD Service Connection over time
Supplementary Figure 1a and 1b show men and women’s changes in PTSD service connection over the 18-year study. Of the 590 men who were service connected for PTSD and under age 55 at Time 1, 20 (3.4%) had discontinued PTSD service connection by Time 2. Of these, half were re-instated by Time 3 (reinstatement rate = 50%). All but 45 men service connected for PTSD at Time 1 had aged out of the risk for age-based re-evaluations by Time 2, while an additional 30 men under age 55 became newly service connected for PTSD at Time 2 (not shown in Figure). Therefore, 75 men were potentially eligible for an age-based PTSD re-evaluation at Time 2; of these, 12 (16.0%) had discontinued PTSD service connection by Time 3. Men’s total PTSD service connection discontinuation rate was 32/620, or 5.2%. Their final discontinuation rate for PTSD service connection was 22/620, or 3.5%.
Of the 754 women service connected for PTSD and under age 55 at Time 1, 67 (8.9%) had discontinued PTSD service connection by Time 2. Of these, service connection was reinstated for 23 women by Time 3 (reinstatement rate = 34.3%). At Time 2, 128 women had aged out of the risk for age-based re-evaluations while an additional 216 women under age 55 became newly service connected for PTSD (not shown in Figure). Thus, a total of 788 women were potentially eligible for age-based PTSD re-evaluations at Time 2. Of these, 113 (14.3%) discontinued PTSD service connection at Time 3. Women’s total PTSD service connection discontinuation rate was 180/970, or 18.6%. Women’s final discontinuation rate for PTSD service connection at Time 3 was 157/970 or 16.2%.
PTSD disability ratings over time
Six men had PTSD disability ratings reduced between Time 1 and 2; no men had their PTSD ratings reduced between Time 2 and 3 (total PTSD disability rating reduction rate = 6/620, or 1.0%). Of the 6 men with reduced PTSD disability ratings at Time 2, three had their rating restored or increased by Time 3 (restoration rate = 50%). The final PTSD disability rating reduction rate was therefore 3/620 or 0.4%. Fourteen women had their PTSD disability ratings reduced between Time 1 and 2, and nine had their PTSD ratings reduced between Time 2 and 3 (total PTSD disability rating reduction rate = 23/970, or 2.4%). Of the 14 women with reduced PTSD disability ratings at Time 2, eight had their rating restored or increased by Time 3 (restoration rate = 57.1%). The final PTSD disability rating reduction rate was 15/970, or 1.5%.
Substitutions
Among Veterans with discontinued PTSD service connection or reduced PTSD disability ratings at Time 3, Table 2 shows their other service connected outcomes across time. As the first set of columns show, among those with discontinued PTSD service connection, 18 of the 22 men and 145 of the 157 women retained service connection for at least one other disorder at Time 2. By Time 3, 21 men and 149 women had service connection for other disorders. For these Veterans, despite losing PTSD service connection, mean total disability ratings (combined degree of service connection) increased between Time 1 and Time 2 and again between Time 2 and Time 3.
Table 2.
Service connection outcomes among those with discontinued PTSD service connection or reduced PTSD ratings at Time 3
| Service Connection Outcomes | Veterans with discontinued PTSD service connection at Time 3 | Veterans with reduced PTSD service ratings at Time 3 | ||
|---|---|---|---|---|
| Men N = 22 | Women N = 157 | Men N = 3 | Women N = 15 | |
| Service connected for disorders other than PTSD at Time 2, n (%) | 18 (81.8%) | 145 (92.4%) | 3 (100%) | 10 (66.7%) |
| Service connected for disorders other than PTSD at Time 3, n (%) | 21 (95.4%) | 149 (94.9%) | 3 (100%) | 13 (86.7%) |
| Total disability ratings (combined degree of service connection) | ||||
| At Time 1, mean (SD) | 46.4 (31.1) | 36.7 (28.3) | 73.3 (25.2) | 66.0 (31.4) |
| At Time 2, mean (SD) | 53.9 (29.9)a | 52.9 (31.7)a | 53.3 (20.8) | 78.7 (25.9) |
| At Time 3, mean (SD) | 68.1 (29.8)b | 63.8 (31.9)b | 73.3 (15.3) | 74.0 (18.8) |
Among the 18 men and 145 women who had any service connection at all at Time 2.
Among the 21 men and 149 women who had any service connection at all at Time 3.
The next set of columns show that all 3 men with reduced PTSD disability ratings maintained service connection for other disorders at Times 2 and 3. For the 15 women with reduced PTSD disability ratings at Time 3, 10 were service connected for other disorders at Time 2, and 13 were service connected for other disorders by Time 3. Men’s average total disability ratings (combined degree of service connection) decreased at Time 2 compared to Time 1, but were restored completely by Time 3. For women who maintained service connection for disorders other than PTSD, mean total disability ratings (combined degree of service connection) increased between Time 1 and 2 and then dipped slightly at Time 3 compared to Time 2.
Predictors of Discontinued PTSD Service Connection at Time 3
Table 3 shows the characteristics of men and women with discontinued PTSD service connection at Time 3 compared to their counterparts’. For men, service after the Vietnam Conflict and no PTSD chart diagnosis were statistically significantly associated with discontinued PTSD service connection at Time 3. Statistically significant predictors of discontinued PTSD service connection for the women included African American or black race, Hispanic ethnicity, no combat or military sexual assault history, no PTSD chart diagnosis, and persistent serious mental illness.
Table 3.
Characteristics associated with discontinued PTSD service connection at Time 3
| Characteristic | Men N = 620 | Women N = 970 | ||||
|---|---|---|---|---|---|---|
| Discontinued PTSD Service connection at Time 3? | p-value | Discontinued PTSD Service connection at Time 3? | p-value | |||
| No n = 598 | Yes n = 22 | No n = 813 | Yes n = 157 | |||
| Race/Ethnicity | ||||||
| White | 73.4 | 77.3 | 0.69 | 78.0 | 66.9 | 0.003 |
| Black or African American | 14.5 | 13.6 | 0.91 | 14.3 | 22.3 | 0.01 |
| Hispanic | 7.9 | 0.0 | 0.17 | 4.3 | 7.6 | 0.08 |
| Other | 6.9 | 0.0 | 0.20 | 6.6 | 5.1 | 0.47 |
| Served during Vietnam Conflict | 90.8 | 63.6 | <0.001 | 20.5 | 13.4 | 0.04 |
| Married at inception | 61.0 | 54.5 | 0.54 | 36.5 | 35.0 | 0.72 |
| Military trauma history scale score | 0.44 | <0.001 | ||||
| 2 (Combat and sexual assault) | 3.3 | 4.5 | 20.7 | 8.9 | ||
| 1 (Combat or sexual assault) | 92.8 | 86.4 | 65.8 | 62.4 | ||
| 0 (Neither combat nor sexual assault) | 3.8 | 9.1 | 13.5 | 28.7 | ||
| Any PTSD diagnosis, 1994–2006 | 94.1 | 68.2 | <0.001 | 92.7 | 66.2 | <0.001 |
| Persistent serious mental illness, 1994–2006 | 8.7 | 13.6 | 0.42 | 19.8 | 26.8 | 0.05 |
Results are reported as column percentages (%). p-value tests for statistically significant differences between those with and without discontinued PTSD service connection within gender.
Multivariable regression results (Table 4) did not differ substantially from the unadjusted results, especially for the men. For women, service after the Vietnam Conflict was no longer statistically significantly associated with discontinued PTSD service connection after controlling for other variables in the model. The remaining predictors identified on univariate analysis retained statistical significance. Hosmer-Lemeshow χ2 for model fit for the men was 0.26 (df = 1, p = 0.61) and 6.61 (df = 6, p = 0.36) for the women.
Table 4.
Predictors of discontinued PTSD service connection at Time 3. Results of multivariable logistic regression.
| Predictor | Men N = 620 | Women N = 970 | ||||
|---|---|---|---|---|---|---|
| AOR for Discontinued PTSD SC at Time 3 | 95% CI | p-value | AOR for Discontinued PTSD SC at Time 3 | 95% CI | p-value | |
| African American or black race | -- | -- | -- | 1.65 | 1.03, 2.63 | 0.04 |
| Hispanic ethnicity | -- | -- | -- | 2.36 | 1.12, 4.98 | 0.03 |
| Served after the Vietnam Conflict | 4.94 | 1.92, 12.72 | 0.001 | 1.40 | 0.84, 2.36 | 0.20 |
| No combat or military sexual assault history | -- | -- | -- | 1.96 | 1.41, 2.70 | <0.001 |
| No chart diagnosis of PTSD, 1994–2006 | 6.54 | 2.42, 17.54 | <0.001 | 6.82 | 4.31, 10.80 | <0.001 |
| Persistent serious mental illness, 1994–2006 | -- | -- | -- | 2.26 | 1.47, 3.49 | <0.001 |
AOR = Adjusted Odds Ratio. CI = Confidence Intervals. PTSD = posttraumatic stress disorder. SC = service connection.
--- Not in model.
Substitution patterns in Veterans with discontinued PTSD service connection by predictive subgroups
Table 5 shows the change in mean total disability ratings (combined degree of service connection) for the 22 men and 157 women with discontinued PTSD service connection at Time 3, stratified by the predictors of same. Male Veterans without a chart diagnosis of PTSD had a slight dip in their mean total disability ratings at Time 2 compared to Time 1. Otherwise, despite having no service connection for PTSD at Time 3, all other subgroups experienced net increases in their total disability ratings at each study point, including at Time 3. Although African American or black race, Hispanic ethnicity, no combat or military sexual assault history, and persistent serious illness were each associated with higher odds of discontinued PTSD service connection, by Time 3, women with these characteristics who lost their PTSD service connection had higher mean total disability ratings than did their counterparts without these characteristics. Men and women without a PTSD chart diagnosis had lower mean total disability ratings at Time 3 than did their counterparts with a chart diagnosis.
Table 5.
Total disability ratings (combined degree of service connection) for Veterans with discontinued PTSD service connection at Time 3 by statistically significant predictors
| Predictors | N | Inception Mean (SD) | Time 2 Mean (SD) | Time 3 Mean (SD) |
|---|---|---|---|---|
| In Men: | 22 | |||
| Service Era | ||||
| During Vietnam Conflict | 14 | 53.6 (31.8) | 46.4 (36.3) | 74.3 (29.5) |
| After Vietnam Conflict | 8 | 33.8 (27.2) | 40.0 (32.5) | 48.8 (32.7) |
| Chart diagnosis of PTSD | ||||
| Yes | 15 | 49.3 (34.9) | 46.7 (40.1) | 72.0 (33.6) |
| No | 7 | 40.0 (21.6) | 38.6 (17.7) | 50.0 (25.8) |
| In Women: | 157 | |||
| Chart diagnosis of PTSD | ||||
| Yes | 104 | 39.7 (30.5)*** | 54.5 (33.7)** | 67.3 (31.4)*** |
| No | 53 | 19.1 (20.2) | 37.7 (30.5) | 47.4 (35.8) |
| Race | ||||
| African American or black | 35 | 26.9 (24.2) | 49.1 (33.5) | 75.1 (25.7)** |
| Other race | 122 | 34.4 (30.2) | 48.4 (33.7) | 56.4 (35.2) |
| Ethnicity | ||||
| Hispanic ethnicity | 12 | 45.0 (33.7) | 42.5 (33.3) | 64.2 (35.3) |
| Not of Hispanic ethnicity | 145 | 31.7 (28.6) | 49.4 (33.6) | 60.3 (34.1) |
| Trauma history scale | ||||
| Comat and military sexual assault | 14 | 34.3 (32.0) | 43.6 (35.2) | 55.0 (36.1) |
| Combat or military sexual assault | 98 | 35.4 (29.5) | 49.4 (32.9) | 60.6 (34.0) |
| Neither | 45 | 26.4 (26.8) | 49.3 (35.1) | 62.2 (34.4) |
| Persistent serious mental illness | ||||
| Yes | 42 | 41.7 (32.2)* | 67.4 (31.6)*** | 81.7 (23.3)*** |
| No | 115 | 29.5 (27.3) | 42.1 (31.7) | 52.9 (34.3) |
Veterans who lost all service connection for all disorders are given a zero value for their total disability rating (combined degree of service connection). p-value tests for mean differences in total disability ratings (combined degree of service connection) by predictor category at each time point.
p ≤ 0.05,
p ≤0.01,
p≤0.001
Discussion
Discontinuing PTSD service connection should be predicated on highly desirable events, such as losing one’s PTSD diagnosis or recovering enough functioning to no longer meet criteria for PTSD service connection. The low final discontinuation rate in this cohort suggests that such events were rare. Over this 18-year period, 3.5% men and 16.2% of women had their PTSD service connection discontinued; only 1 man in 620 and 8 women in 970 had service connection discontinued for all disorders. Less than 3% of Veterans had their PTSD disability ratings reduced. Furthermore, among Veterans with reduced ratings, nearly half were eventually restored, and almost all experienced increases in their other ratings to compensate for the loss.
Individuals who receive disability support, whether through VA or elsewhere, tend to be more symptomatic than individuals who do not receive support.18–21 The low rates of discontinued or reduced PTSD service connection in this sample suggests that the initial claims process identified Veterans with durable disability. In this setting, routine re-evaluations based on age and not on actual clinical improvement are likely to be of low yield and to carry high opportunity costs. Here, for example, at least 28 male and 6 female re-evaluations were required to find 1 man and 1 woman with a Time 3 discontinued PTSD service connection. Although higher proportions of women than men discontinued their PTSD service connection, from a system’s standpoint, women currently account for just 12–14% of all Veterans service connected for PTSD.22 Furthermore, the high reinstatement and restoration rates, ranging from 34.3% – 57.1% for women and 50.0% for the men, suggest that the current re-evaluation process carries a high error rate, regardless of gender. Errors further encumber the system: Wrongly losing PTSD service connection limits Veterans’ access to free PTSD treatments, potentially reduces their income, and imposes administrative burdens to reverse the error. The onus is on Veterans, for example, to file any appeals, while the VA must review and process those appeals.
Whether the gender difference in discontinued PTSD service connection represents a more favorable clinical outcome for women, an unfair disparity, or confounding by age is unknown. In the present analysis, men’s discontinuation rate was so low and their characteristics differed from women’s so extensively, the two groups’ discontinuation rates could not be validly compared using logistic regression. Women’s much younger age at inception certainly placed them at longer risk for age-based re-evaluations than the men, and service after the Vietnam Conflict, which was strongly correlated with younger age, was one of only two predictors for discontinued PTSD service connection in the men. Had men and women been closer in age, their discontinuation rates might have been more similar. In previous analyses, observed gender differences in discontinued PTSD service connection diminished considerably—but did not eradicate—when age and other sociodemographic characteristics were controlled.23
Ratings substitutions, where a Veteran loses service connection for one disorder but gains a higher rating or new rating for a different disorder, may not yield cost savings to the system and might even generate cost increases. At the individual level, Veterans who lose their 70% PTSD disability rating, say, but gain a 70% disability rating for something else should not see net changes to their monthly compensation. On balance, most Veterans with discontinued or reduced PTSD ratings experienced increases in their total disability ratings. This was also true for most subgroups we examined, suggesting that ratings for other disorders were substituted in and that their total disability burden stayed high. The one subgroup exception, Veterans without a PTSD chart diagnosis, could have been misdiagnosed originally, had milder disorders, or carried a better prognosis than other Veterans.
To our knowledge, this is the first long-term examination of service connection outcomes in a cohort of Veterans under age 55. We assessed service-connected outcomes at three 8 to 9-year intervals, but re-evaluations occur every 2 to 5 years. Thus, we may have substantially undercounted the number of discontinuations, ratings reductions, and intervening reversals in this cohort. This undercount’s net effect would be to underestimate the error rate and the number of evaluations needed to find one man and one women with discontinued PTSD service connection at Time 3. Veterans with discontinued or reduced PTSD service connection at Time 3 might yet reverse their loss, again underestimating the reversal rate. Although unusual, some Veterans may have been exempted from re-evaluation when they first became service connected for PTSD. Including them in the denominator would underestimate the discontinuation and ratings reduction rate. Despite a reasonably large starting sample, too few Veterans had discontinued or reduced PTSD service connection to meaningfully explore potential predictors of discontinued service connection, particularly among the men. Researchers need to consider this low event rate when planning and powering future studies. We also lacked clinical information, such as disease severity, Veterans’ participation in treatment, access to evidence-based treatments, or their response to therapy, to determine whether observed subgroup differences in discontinued PTSD service connection represented inappropriate disparities or justifiable variations in disease outcomes. Although representative of Veterans applying for PTSD service connection between 1994 and 1998, our results may not pertain to Veterans with more recent applications for disability benefits.
Conclusions
Particularly for men, discontinuing or reducing PTSD service connection was rare and often reversed. Regardless of gender, most Veterans with discontinued or reduced PTSD service connection did not experience reductions in their overall, total disability rating. Mandated, age-based re-evaluations of PTSD service connection represent substantial work for the VA and may be disruptive to Veterans. Whether the opportunity costs required to identify a single Veteran with discontinued or reduced PTSD service connection represents the best use of resources should be more fully evaluated. Prospective cost-benefit analyses could clarify these issues. Incorporating research findings into VA’s disability re-evaluation policies could optimize their efficiency.
Supplementary Material
Supplementary Figure. Changes in PTSD service connection over time for men (1a.) and women (1b.)
Acknowledgements:
The authors thank Ann Bangerter and Barb Clothier for database management.
Funding/COI:
This work was supported by the VA Health Services Research and Development Program (IIR-09-359, CIN-13-406). The funder had no role in data analysis, manuscript preparation, or decision to publish.
Footnotes
Ethical Approval: The Minneapolis VA Health Care System’s Human Studies Subcommittee
Publisher's Disclaimer: Disclaimer: Views expressed are solely those of the authors and do not reflect the opinion, views, policies, or position of the Department of Veterans Affairs.
References
- 1.Panangala S Department of Veterans Affairs FY2017 appropriations. Congressional Research Service; September 19, 2017. R44625. [Google Scholar]
- 2.Veterans Benefits Administration. Annual Benefits Report: Fiscal Year 2016: Compensation. Washington DC: US Department of Veterans Affairs; 2016. [Google Scholar]
- 3.Kessler R, Sonnega A, Bromet E, et al. Posttraumatic stress disorder in the National Comorbidity Survey. Arch Gen Psychiatry. 1995;52:1048–1060. [DOI] [PubMed] [Google Scholar]
- 4.Breslau N, Kessler R, Chilcoat H, et al. Trauma and posttraumatic stress disorder in the community: the 1996 Detroit area Survey of Trauma. Arch Gen Psychiatry. 1998;55(7):626–632. [DOI] [PubMed] [Google Scholar]
- 5.Hoskins M, Pearce J, Bethell A, et al. Pharmacotherapy for post-traumatic stress disorder: systematic review and meta-analysis. Br J Psychiatry. 2015;206(2):93–100. [DOI] [PubMed] [Google Scholar]
- 6.Steenkamp M, Litz B, Hoge C, Marmar C. Psychotherapy for military-related PTSD: a review of randomized clinical trials. JAMA. 2015;314(5):489–500. [DOI] [PubMed] [Google Scholar]
- 7. 38 CFR. 2000 §3.327. 2000.
- 8.IOM (Institute of Medicine) and National Research Council (NRC). PTSD Compensation and Military Service. Washington, DC: The National Academies Press; 2007. [Google Scholar]
- 9.Spoont M, Sayer N, Nelson D, et al. Does clinical status change in anticipation of a PTSD disability examination? Psychol Serv. 2008;5(1):49–59. [Google Scholar]
- 10.Sayer N, Spoont M, Nelson D, Nugent S. Development and psychometric properties of the Disability Application Appraisal Inventory. Psychol Assess. 2004;16(2):192–196. [DOI] [PubMed] [Google Scholar]
- 11.Murdoch M, Hodges J, Cowper D, et al. Racial disparities in VA service connection for posttraumatic stress disorder disability. Med Care. 2003;41(4):536–549. [DOI] [PubMed] [Google Scholar]
- 12.Murdoch M, Hodges J, Hunt C, et al. Gender differences in service connection for PTSD. Med Care. 2003;41(8):950–961. [DOI] [PubMed] [Google Scholar]
- 13.Murdoch M, Hodges J, Cowper D, Sayer N. Regional variation and other correlates of Department of Veterans Affairs (VA) disability awards for posttraumatic stress disorder. Med Care. 2005;43(2):112–121. [DOI] [PubMed] [Google Scholar]
- 14.Murdoch M, Spoont M, Kehle-Forbes S, et al. Persistent serious mental illness among former applicants for VA PTSD disability benefits and long-term outcomes: symptoms, functioning, and employment. J Trauma Stress. 2017;30(1):36–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Frueh B, Grubaugh A, Cusack K, Elhai J. Disseminating evidence-based practices for adults with PTSD and severe mental illness in public-sector mental health agencies. Behav Modif. 2009;33(1):66–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Janes G, Goldberg J, Eisen S, True W. Reliability and validity of a combat exposure index for Vietnam Veterans. J Clin Psychol. 1991;47:80–86. [DOI] [PubMed] [Google Scholar]
- 17.Murdoch M, McGovern P. Development and validation of the Sexual Harassment Inventory. Violence Vict. 1998;13(3):203–216. [PubMed] [Google Scholar]
- 18.Solomon Z, Benbenishty R, Waysman M, Beich A. Compensation and psychic trauma: a study of Israeli combat veterans. Am J Orthopsychiatry. 1994;64(1):91–102. [DOI] [PubMed] [Google Scholar]
- 19.Kouzis A, Eaton W. Psychopathology and the initiation of disability payments. Psychiatr Serv. 2000;51(7):908–913. [DOI] [PubMed] [Google Scholar]
- 20.Bilder S, Mechanic D. Navigating the disability process: Persons with mental disorders applying for and receiving disability benefits. Milbank Q. 2003;81(1):75–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Knudsen A, Øverland S, Aakvaag H, et al. Common mental disorders and disability pension award: seven year follow-up of the HUSK study. J Psychosom Res. 2010;69(1):59–67. [DOI] [PubMed] [Google Scholar]
- 22.Veterans Benefits Administration. 2017 Annual Benefits Report: Compensation. Washington DC: US Department of Veterans Affairs; 2018. [Google Scholar]
- 23.Sayer N, Hagel E, Noorbaloochi S, et al. Gender differences in VA disability status for PTSD over time. Psychiatr Serv. 2014;65(5):663–669. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure. Changes in PTSD service connection over time for men (1a.) and women (1b.)