Higher VA disability ratings carry sharply higher mortality — and PTSD/mental-health conditions drive a distinct, external-cause pattern that private insurers price differently than physical ratings

Summary

Veterans with a 100% combined VA disability rating have more than double the one-year mortality risk of veterans rated 10-40% (4.5% vs. 1.9%, adjusted odds ratio 2.45), based on a VA study of over 4 million compensated veterans. That excess risk is not uniform across condition types: the 100% tier is disproportionately loaded with PTSD (44.6% of that tier, vs. 9.5% in the low tier) and cardiometabolic conditions (ischemic heart disease, diabetes), while the low tier is dominated by audiological conditions (hearing loss, tinnitus) that carry little independent mortality signal. A companion VA cause-of-death study shows this plays out as a real difference in HOW veterans with high disability ratings die: younger decedents (who skew toward combat/mental-health/TBI-rated conditions) died of external causes -- suicide, accident, assault -- in roughly a third of cases, versus under 2% among the oldest decedents, who died overwhelmingly of cancer and cardiovascular disease.

Private life insurance underwriting does track this same distinction, but through the individual medical condition and its documented severity/treatment stability -- not through the VA's combined percentage rating, which underwriters generally cannot see or use directly. A veteran whose 100% rating is built from PTSD, TBI, or an unstable psychiatric/substance history is medically underwritten on that basis and can face a substandard (table-rated) offer, exclusion riders, or in more severe/unstable cases a decline; a veteran rated 100% mainly from hearing loss, tinnitus, or a well-controlled physical condition typically underwrites close to standard. Because this creates a real risk of veterans with the highest-mortality service-connected conditions being priced out of or declined by the commercial market, the VA runs its own guaranteed-issue programs (VGLI, and now VALife) specifically so that a veteran with any VA service-connected disability rating -- even 0% -- cannot be medically underwritten out of basic life coverage.

1. The mortality gradient by disability rating

Maynard, Nelson & Fihn (Public Health Reports, 2018) followed 4,010,720 living veterans receiving VA disability compensation as of October 1, 2016, using the VETSNET benefits file linked to the VA vital status file, for one year of follow-up.

• High rating (100%): 12.8% of the cohort, 1-year mortality 4.5%
• Medium rating (50-90%): 39.9% of the cohort, 1-year mortality 1.9%
• Low rating (10-40%): 47.2% of the cohort, 1-year mortality 1.9%

After adjusting for age, sex, service branch, paygrade, and VA health-care use, veterans in the 100% tier had 2.45 times the odds of dying within a year compared with the low-rating tier (95% CI 2.40-2.50). The authors describe the 100%-rated group as "a highly select group with increased short-term risk of death due at least in part to their military service."

2. Mortality differs by condition type, not just severity level

The rating tiers are not interchangeable buckets of the same conditions at different severities -- they carry different condition mixes. In the same cohort:

• PTSD: 9.5% of the low tier, 37.1% of the medium tier, 44.6% of the high tier
• Ischemic heart disease: 4.6% low, 8.8% medium, 16.3% high
• Diabetes mellitus: 7.8% low, 14.5% medium, 21.6% high
• Hearing loss: 40.4% low, 28.3% medium, 22.4% high (the opposite gradient)

A companion study by the same VA authors (Military Medicine, 2018) examined 605,344 veteran decedents with service-connected conditions who died 2004-2014 and broke down cause of death by age at death -- age being a strong proxy for condition mix, since combat-era mental-health and TBI ratings concentrate in younger veterans while cardiovascular/cancer-type conditions concentrate in older ones. Among the youngest decedents, external causes (suicide, accident, assault) accounted for roughly a third of deaths and cancer plus cardiovascular disease accounted for another third; among the oldest decedents, external causes fell to under 2% and cancer plus cardiovascular disease rose to roughly half of all deaths. This is the clearest available evidence that veterans rated for PTSD/TBI/mental-health conditions die differently -- via a much larger external-cause (behavioral-risk) component -- than veterans rated primarily for physical/cardiometabolic conditions, even though both groups can carry the same nominal disability percentage.

No dataset found (in this corpus or the published literature) directly cross-tabulates individual cause of death against the SPECIFIC condition a veteran is rated for at the person level; the evidence above is the closest available proxy, and is presented as such rather than as a single clean table.

3. Does life-insurance underwriting reflect this?

Commercial insurers do not underwrite off the VA's combined percentage rating as a number -- they underwrite the individual medical conditions behind it, using their own severity, treatment-history, and stability criteria. Per industry underwriting guidance (e.g., AccuQuote's veteran-focused underwriting guide), applicants disclosing PTSD or TBI are evaluated on symptom severity, treatment/medication history, hospitalization history, and functional stability; a diagnosis alone rarely determines the outcome, but unstable, severe, or recently hospitalized cases are more likely to draw a substandard rate class or, in the most severe cases, a decline. A veteran whose disability rating is built from hearing loss, tinnitus, or a well-controlled physical condition is underwritten on that condition alone and typically receives standard or near-standard treatment, regardless of the percentage on the VA rating letter.

In other words, the underwriting distinction insurers actually draw (condition type, severity, and stability) lines up directionally with the mortality distinction found in the VA data (mental-health/behavioral-risk-associated conditions carrying a different, external-cause-driven mortality signature than physical/cardiometabolic conditions) -- but it is done condition-by-condition, not via the VA's own combined rating.

Because this individual medical underwriting can price out or decline exactly the veterans the VA data show carry the highest mortality risk (heavily mental-health/PTSD-rated, high-percentage veterans), the VA has built its life-insurance programs around guaranteed issue rather than medical underwriting for this population:

• VGLI (Veterans' Group Life Insurance): guaranteed issue with no medical underwriting if applied for within 240 days of separation -- no health questions, cannot be declined.
• S-DVI (Service-Disabled Veterans Life Insurance): the VA's original guaranteed-issue program for veterans with a service-connected disability rating; closed to new applicants after December 31, 2022, existing policyholders retained.
• VALife (launched January 1, 2023): the VA's current program, available to any veteran age 80 or under with ANY VA service-connected disability rating -- even 0% -- with no health questions and guaranteed acceptance.

These guaranteed-issue programs are the direct policy response to the underwriting gap: they exist because, absent them, the private market's condition-by-condition medical underwriting would be expected to disadvantage exactly the highest-mortality-risk rated veterans (heavily PTSD/mental-health/TBI cases) relative to veterans rated for lower-mortality-risk physical conditions like hearing loss.

4. What this corpus (AskAmerica/govdata) does and doesn't cover

search_catalog was checked for VA disability-rating microdata, veteran-specific cause-of-death data, and life-insurance underwriting outcomes, and find_recipe was checked for a relevant analysis pattern, before turning to external sources. The corpus carries census.acs_veterans / acs_veterans_by_period (veteran population counts, no disability-rating or condition detail) and health.cdc_mortality (state-level, all-population cause-of-death, no veteran flag). Neither reaches the veteran-disability-rating-by-cause-of-death grain this question asks about, no cataloged recipe applies, and no life-insurance underwriting dataset exists in the corpus at all. This is a genuine, confirmed data gap rather than an unexplored one -- the entire analysis above rests on external peer-reviewed VA research and VA/industry program documentation.