Service members who become injured or ill while serving may seek treatment for their medical condition(s). If the medical condition(s) continues to interfere with their ability to perform their military duties, their provider may refer them for disability evaluation to determine whether they are fit to continue serving. Most of these service members are medically discharged and are awarded disability benefits for conditions that make them unfit to serve. These injuries and illnesses are often the result of the physical nature of a military career, but may be more likely to occur during periods of frequent deployment. Since October, 2001, when Operation Enduring Freedom began in Afghanistan, approximately 3 million service members have been deployed, resulting in 4,094 service members killed in action and 52,737 wounded in action (DeBruyne, 2018).1 Over that same time period, tens of thousands of service members have been diagnosed with posttraumatic stress disorder (PTSD) and/or traumatic brain injury (TBI).
PTSD is a mental health condition that some people experience after a terrifying or life-threatening event, such as combat. People with PTSD often experience nightmares, flashbacks, and intense anxiety, with symptoms lasting for months (U.S. Department of Veterans Affairs, 2019), or for some patients, throughout their lives. There are effective treatments for PTSD, including medication and psychotherapy, so ensuring that service members with PTSD are identified and treated is a priority for the military health system. TBI is a serious head injury that causes temporary or permanent damage to the brain. A TBI can be mild, moderate, or severe. Mild TBIs, also known as concussions, are the most common type of TBIs (Centers for Disease Control and Prevention, 2019). Recovery from TBI depends greatly on the severity of the injury; most of those with a mild TBI have a complete recovery, though identification and appropriate treatment are necessary.
At the same time that the U.S. military has been engaged in combat operations in Afghanistan and Iraq, the process by which service members are evaluated for disability has evolved significantly, including a complete overhaul of the Disability Evaluation System (DES) beginning in 2007. Simultaneously, DoD and the services have made policy changes and initiated other efforts to improve screening for PTSD and TBI, encourage service members to seek treatment, improve quality of care, and reduce the stigma associated with treatment for these conditions.
In this study, we conducted an empirical analysis of trends in diagnosis, treatment, and disability evaluation for PTSD and TBI. In a separate, companion report (Simmons et al., 2021), we identified and described changes that have been made to the DES between 2001 and 2018, as well as policy changes associated with identifying and treating PTSD and TBI.
Trends in Diagnosis and Disability Evaluation
Overview of Methods
To analyze trends in diagnosis and disability outcomes, we constructed a person-year file representing each year of active component (AC) service for anyone who served between FY 2002 and FY 2017. We included administrative data containing demographic and service characteristics, records of medical encounters, health assessments following deployments, and disability evaluation data. We defined diagnosis cohorts based on the first observation of a PTSD and/or TBI diagnosis during a medical encounter, and we created disability cohorts based on the presence of a disability rating for an unfitting condition. We followed diagnosis cohorts forward, and looked back in time at disability cohorts, for three years to document disability evaluation outcomes and service experiences, respectively. We also took a multivariate approach that did not restrict the analysis to any length of time and followed the diagnosis cohorts to determine what characteristics were associated with faster time to disability evaluation following assignment to a diagnosis cohort.
In addition to analyzing trends in diagnoses, disability outcomes, and service member characteristics for those diagnosed with or disability-rated for PTSD or TBI, as well as those with both conditions (PTSD+TBI), we also selected three comparison conditions to provide context. For example, if we see an increase in the number of service members who are medically discharged with a disability rating for one or both of these conditions, we might expect a similar pattern for all conditions if the underlying drivers of disability evaluation affect all conditions or service members in the same way. Or, there might be factors associated with one or both of these conditions, but not with other conditions, that would cause the patterns we see to be unique to PTSD and/or TBI. We selected sleep apnea, major depressive disorder, and back pain as our comparison conditions.
Diagnosis and Disability Cohorts
Figure 1 shows the number of AC service members who were assigned to PTSD and TBI diagnosis and disability cohorts. The size of the cohorts grew over time, with the growth in disability cohorts lagging diagnosis cohorts because service members typically first receive treatment, and then disability evaluation, for those referred, may take a year or more. TBI diagnosis cohorts were consistently larger than PTSD cohorts, but over this time period, the number of service members with a PTSD disability rating (the size of the disability cohorts) was larger than the number with a TBI disability rating.
If the diagnosis cohort counts are scaled to the size of the total active force, the number of service members assigned to the TBI diagnosis cohort represents approximately 0.8 percent of all AC service members present in the first year of our data (FY 2002). In peak years (2008–2011), approximately 1.6 percent of the total active force was first diagnosed with TBI (and therefore assigned to the diagnosis cohort). The share of the total active force assigned to a PTSD cohort in a given year was consistently smaller than the share assigned to TBI cohorts: It started out around 0.2 percent of the total active force in 2002 and peaked at 1.3 percent in 2012 before beginning to decline.
Approximately 2 percent of service members who were medically discharged in 2002 had either a PTSD and/or TBI disability rating. The share of medical discharges with a PTSD disability rating grew quickly after that, to approximately 20 percent between 2009 and 2014. The proportion of medical discharges with a TBI rating was significantly lower throughout the time period, peaking around 7 percent in 2009 and 2010 and remaining steady at around 5 percent for the rest of the observation period.
Prospective Analysis of Diagnosis Cohorts
Following diagnosis cohorts for three years, we found that the share of service members in a PTSD diagnosis cohort who were evaluated for disability more than doubled—from 16 percent in 2002 to 34 percent in 2015 (Figure 2). Service members diagnosed with TBI experienced slower growth in the rate of disability evaluation over the time period. Since 2008, approximately 80 percent of service members in the PTSD diagnosis cohorts who had a disability evaluation were medically retired (disability rating of 30 percent or higher). The percentage of service members in the PTSD diagnosis cohorts who were evaluated for disability and received a disability rating for PTSD ranged from 40 to 60 percent. For TBI diagnosis cohorts, the share with a TBI disability rating at the conclusion of DES was even lower: Until 2008, 30 percent of the TBI diagnosis cohort also had a TBI disability rating, which declined to just more than 10 percent in the latest cohorts.
Using the same diagnosis cohorts, we conducted hazard analyses of time to disability evaluation following diagnosis. Service members diagnosed with PTSD or PTSD+ TBI were evaluated for disability sooner after their diagnosis than those diagnosed for some of the comparison conditions we analyzed. In addition, service members in more recent diagnosis cohorts were evaluated for disability sooner than service members diagnosed in earlier years, particularly those in the PTSD and PTSD+TBI cohorts.
Finally, we relaxed the three-year window of observation for following the diagnosis cohorts and characterized their disposition at the end of our data period in FY 2017 (Figure 3). In FY 2002, 21 percent of the PTSD cohort was medically discharged through the DES; that rate was near double (35–39 percent) for the FY 2008–2014 cohorts. The proportion of the TBI cohorts that was medically discharged was consistently lower, peaking around 25 percent for the FY 2007–2013 cohorts. The combination of Expiration Term of Service (ETS) and retirement account for 30–40 percent of exits in the PTSD and TBI diagnosis cohorts, and less than 2 percent of service members in all disability cohorts had died by the end of the analysis period. The remainder of service members in the cohorts were administratively separated or still serving in FY 2017, and a small fraction had unknown dispositions.
Retrospective Analysis of Disability Cohorts
Using the disability cohorts, we found that approximately 90 percent of service members with a disability rating for PTSD and/or TBI had a diagnosis for that condition in the previous three years. Among those with a disability rating for PTSD, 60–70 percent had a positive screen for PTSD on the post-deployment health assessment (PDHA) or post-deployment health reassessment (PDHRA). Service members who were medically discharged with a TBI disability rating between 2009 and 2012 were more likely to have a positive screen for TBI on the PDHA or PDHRA than those discharged in earlier or later years. A 2008 policy mandating that service members with a mental disorder from traumatic stress (such as PTSD) were to be awarded a minimum 50-percent disability rating was borne out in the data; in 2008, 57 percent of the PTSD cohort had a total disability rating of 0–40 percent, but by 2009, almost every member of the cohort had a rating of 50 percent or higher.
A multivariate analysis of the probability of having a PTSD or TBI disability rating at the conclusion of DES showed that deployments, occupation, service, and fiscal year of discharge were the main drivers. The effects of all variables on the probability of receiving a PTSD rating were larger than on the probability of receiving a TBI rating.
We faced analytic and data limitations in this study. First, although we are able to make associations between some of the outcomes of interest and the time period or environment in which the outcomes occurred, we are unable to make causal inferences about the outcomes. This is for of two reasons: (1) Policies are not issued in isolation, and (2) there are many factors that we cannot control for using secondary data, including service member preferences, exposures, and experiences that may lead to different outcomes and could be misattributed to the policy change.
We restricted our analysis to AC service members, because our analyses either depended upon a diagnosis or pertained to treatment for a disabling condition. Most Reserve and National Guard service members do not receive health care through the military health system or on the TRICARE purchased care network, so including them in our analysis would have been incomplete and misleading.
With respect to data, the most important limitation is that the disability data source covering the most recent years may include both fitting and unfitting conditions, whereas files covering earlier years (prior to the integration of VA and DoD in disability evaluation) include only conditions found by DoD to be unfitting (which was our intent). The consequence is that later disability cohorts (beginning in 2007 but especially 2012–2017) may be larger than intended. Additionally, not all files contained condition-specific disability ratings, so we had to use total DoD disability rating, which includes all unfitting conditions.
Across analyses, we found that the number and percentage of service members with a diagnosis or disability rating for PTSD and/or TBI increased over time until recently, although this was not true for all comparison conditions. Given that these two conditions are thought of as signature wounds of the wars in Iraq and Afghanistan, it is perhaps not surprising that as the number of service members deployed increased, so too did the incidence of these conditions. And similarly, now that a relatively small number of service members have been deployed in recent years, there has been a downward trend in the number and percentage being treated or rated for PTSD and/or TBI. We also observed some trends that align with policy. The clearest example of this is the shift toward higher total DoD disability ratings for service members in the PTSD and sleep apnea cohorts following a 2008 policy directing the services to adhere to the U.S. Department of Veterans Affairs Schedule for Rating Disabilities (VASRD) (which resulted in service members with PTSD receiving a minimum 50-percent disability rating and service members with sleep apnea being rated according to clinical criteria rather than civilian earning capacity).
While there is clearly a relationship between deployment patterns and policy and the trends observed in this study, we cannot specifically attribute the findings to deployments or any other factor. We can note that during the time period covered by these analyses, the disability system was restructured, efforts were made to raise awareness about these conditions and encourage people to seek treatment and to reduce stigma for receiving treatment, and there were many policy changes related to disability evaluation and the identification, diagnosis, and treatment of PTSD and TBI. The results in this study are likely a reflection of all of those factors.
While in general we cannot judge whether the trends we have observed or the policies put in place over this period are positive or negative, there are some outcomes that likely improved service member well-being. For example, we observed an increase in the number and percentage of service members diagnosed with PTSD and TBI. Over this time period, there was increased focus on the psychological toll of the wars, and in response, DoD made changes to how it organized psychological health resources and capabilities. Service members were encouraged to seek treatment, and there were efforts to improve screening, diagnosis, and treatment for these conditions (Simmons et al., 2021). An increase in the number of service members treated might reflect the success of these programs and initiatives.
Coinciding with the mandatory 50-percent minimum disability rating for service members with a mental disorder resulting from a traumatic event, our results showed a clear trend toward higher ratings for service members in the PTSD (and to a lesser extent, other conditions) cohorts after the policy change. To the extent that improved financial well-being is tied to health and socioeconomic outcomes, an increase in benefits associated with higher disability ratings may enhance opportunities for medically retired service members to reintegrate into the civilian world and continue to receive necessary treatment.
In the future, DoD and the services will continue to evolve policies and practices to improve system performance and service member health. To the extent possible, the effects of those changes should be evaluated as they happen. This study documents a confluence of changes that occurred over a 16-year period, which makes it difficult to identify the impact any single policy had on an outcome such as fitness to serve or disability. But it may be possible to evaluate such changes within a narrower band of time or the effects of the changes on a subset of service members to ensure that the desired outcome is achieved.
Furthermore, with existing data, it would be possible to more formally assess veteran outcomes following medical discharge. This could be done by linking records from military service (personnel and health data) to VA health care utilization (or other benefits) data, or to Social Security Administration data on earnings. Additional data collection would be required to be able to study long-term health outcomes, as this information is not currently recorded in DoD or VA health care utilization data. These types of studies would enable an assessment of the well-being of service members whose war-related conditions affect life after medical discharge.