The baseline without an exposure record

I spent my Army career on the counter-IED and electronic warfare side. That put me around demolitions, breaching, and range work often enough that the population this report is about is not abstract to me. Explosive ordnance disposal is on the Pentagon's own list of high-risk occupational specialties, and so are the jobs of most of the people I worked next to.

So when the Government Accountability Office published its review of DOD's blast overpressure program on 18 August 2026, I read it as an operator first. Then I read it again as somebody who spends his working life on measurement, and the second reading is the one that produced this report.

The coverage all landed on the same sentence: the Pentagon needs more resources. That is true, and GAO says it plainly. But it is the proximate cause, not the interesting part. The interesting part is what happens to a measurement when you build one half of it well and the other half not at all.

The design is a pair, and GAO says so

DOD's August 2024 blast overpressure memorandum created two requirements that only make sense together.

The first is the cognitive baseline. GAO describes its purpose in one sentence:

Health care providers can compare subsequent cognitive assessments to a service member's baseline assessment to detect cognitive changes.

The second is exposure tracking, and GAO describes that purpose the same way:

Another requirement calls for the tracking of service members' exposures to blast overpressure so that they have a longitudinal record of their hazard exposures.

Put those side by side and you have a study design. The cognitive score is the dependent variable, measured before and after. The exposure history is the independent variable. Without the second, a change in the first is real but unattributable: you can see that a person's score moved, and you cannot say what moved it. Age, sleep, a concussion from a vehicle rollover, a bad week, or eleven years of standing behind a breaching charge all produce the same kind of drift in the record.

This is not a criticism of anyone's intent. It is what the program's own documents say it is trying to do.

Half one, which worked

The baseline half is close to done, and it was the harder half to schedule.

In fiscal year 2025, 198,161 service members received cognitive assessments during initial military training: 97,585 in the Army, 57,265 in the Navy, 43,311 in the Air Force. For the high-risk active-duty population, the deadline was 30 September 2025, and DOD data show the military departments had tested 86 percent by the end of that fiscal year. GAO requested updated numbers, and as of March 2026, 96 percent of high-risk active-duty service members had completed baseline assessments.

One honest caveat, and it is GAO's, not mine. The report notes that the 96 percent figure "is due in part to these exclusions, which resulted in a smaller number of service members who required testing." Some of the improvement between 86 and 96 percent is a shrinking denominator rather than a growing numerator. That does not erase the achievement, and 198,161 assessments in a year is not a rounding error, but a completion rate that moves partly because the population was redefined deserves the asterisk.

Half two, which did not

Now the exposure record. GAO's finding is short: "As of the end of fiscal year 2025, the Army and Navy had started, but had not completed, exposure tracking requirements."

The number underneath that sentence is the one worth carrying away. The Army is the department with the largest high-risk population, and in its quarterly implementation survey it reported that as of the end of fiscal year 2025 it "had started exposure tracking for less than 2 percent of the 142,000 potentially exposed service members."

Not completed for 2 percent. Started for less than 2 percent.

The stated reason is staffing, and specifically a shortage of a job most people outside the services have never heard of. Exposure tracking runs through industrial hygienists, who use the Defense Occupational and Environmental Health Readiness System-Industrial Hygiene to collect and analyze exposures. GAO reports flatly that "DOD component officials reported a lack of industrial hygiene personnel to complete exposure tracking requirements," and that the Defense Health Agency "stated that there were not enough industrial hygiene personnel to implement DOD's blast overpressure requirements." An Army audit found that as of November 2024 the difficulty was made worse when DHA took over industrial hygiene responsibility in October 2022, partly because of insufficient staff.

There is a structural asymmetry hiding here that the resources framing obscures. The baseline half is a clinical function with an existing delivery channel: recruits already sit down in front of a screen during initial training, so you add one more assessment to a queue that already exists. The exposure half has no equivalent channel. Somebody has to attribute a specific service member to a specific weapon event, then record it. Those are different organizations with different capacity, and only one of them had a place to put the new work.

The sensor route, and why it hasn't rescued this

The obvious fix is to stop doing this by hand. Put a gauge on the person, let it record, and the exposure half becomes automatic. DOD is working that problem: officials told GAO the department began a study in 2025 on aspects of commercially available blast overpressure sensors, including durability and longevity, and that sensors "could help indicate if blast overpressure exposure has exceeded levels considered to be safe with real-time information."

It has not arrived, and the reason is the part I recognize from instrument work. From GAO:

As of September 2025, DHA officials said the algorithm to support this approach was not performing consistently and therefore further review would be needed before sensor data could be consistently translated into exposure measurements.

That sentence is the whole difficulty of the field compressed into one line. A pressure gauge on a vest is not a dosimeter. It records a waveform at one point on one body in one orientation, and turning that into "this person received this exposure" requires a model of how the wave interacted with the person wearing it: standing or prone, inside a structure or in the open, facing the charge or turned away, and how reflections off a wall added to the free-field value. That translation is the measurement. The hardware is comparatively easy. This is the same class of problem I have written about here with metal detectors and with radar: the sensor returns a signal, and the hard, contested, career-consuming work is the inference from signal to fact.

What 4 psi is, and what it is not

Almost every article about this program cites a threshold of 4 pounds per square inch, usually with the phrasing of a safety limit. It is worth being precise about what that number is, because DOD itself is.

GAO records the origin: "Specifically, in 2022, in interim guidance, DOD established a threshold of blast overpressure exposure at 4 pounds per square inch," and elsewhere that "the interim guidance established an interim threshold of blast overpressure exposure at 4 pounds per square inch based on reported health effects above this level." In 2024, DOD replaced the interim guidance with the blast overpressure memorandum.

The memorandum's own characterization of the number, quoted by GAO, is careful in a way the coverage is not. Weapons on DOD's list "have been shown to produce blast overpressure exposures exceeding 4 pounds per square inch and this threshold 'was identified as a health-based safety guideline informed by evolving medical science.'" GAO adds that the memorandum "noted that the weapons list is not exhaustive, meaning that there could be other weapon systems that are associated with higher levels."

Read that as written. Four psi is an administrative trigger derived from a body of evidence that DOD describes as still evolving, attached to a weapons list DOD says is incomplete. It is a reasonable place to start managing a hazard. It is not a biological threshold below which nothing happens, and DOD has not claimed it is one. Anyone treating 4 psi as a line between safe and unsafe is reading more into it than the document supports.

None of that is a reason to dismiss the program. It is a reason to expect the number to move, and to want the exposure record built now so that there is data to revise it against later.

What GAO actually recommended

Two recommendations, both open, both about resourcing rather than redesign. GAO recommends that the Secretary of Defense complete a comprehensive assessment of the resources needed for the cognitive assessment requirements, including an evaluation of the new pilot, and a second comprehensive assessment covering industrial hygiene staffing for exposure tracking. DOD agreed with both and identified actions it plans to take.

The pilot is worth watching: a new cognitive assessment tool begins piloting in summer 2026, with a self-administered web application intended to reduce the staffing load. That addresses the half that is already at 96 percent. I have not seen an equivalent lever proposed for the half that is at 2 percent, which is the half where a missing year is unrecoverable. A cognitive baseline can be taken late. An exposure that happened on a range in 2024 and was never written down cannot be reconstructed in 2031 from anything better than memory.

If you are in that population

This is a reporting piece about a measurement gap, not advice about anyone's health, and I am not the person to give the latter. What I will say is the records point, because it is the one that follows directly from the finding above: the occupational exposure record is a document that either exists or does not, and right now, for most of the Army's high-risk population, it does not. Asking what is in yours, and keeping your own dated notes on significant exposure events, costs nothing today and is worth considerably more than a reconstruction attempted a decade from now. Anything beyond documentation belongs with your own medical channel.

What I did not verify

The primary source here is the GAO report itself, which I downloaded in full and read directly. It is an oversight report, not a clinical study, and this piece stays inside what an oversight report can establish: what DOD required, what DOD has completed, and what DOD's own officials told GAO about why.

I did not read DOD's 8 August 2024 blast overpressure memorandum or the 2022 interim guidance as primary documents. Every quotation from those, including the "health-based safety guideline informed by evolving medical science" language and the note that the weapons list is not exhaustive, is quoted as GAO reproduces it. If the underlying memoranda say something GAO did not capture, this report inherits that gap.

I have not independently assessed the science linking repeated low-level blast exposure to brain health outcomes. GAO characterizes it as growing evidence and describes associated effects; I have not gone behind that characterization to the underlying literature, and nothing in this report depends on the strength of that link. The argument here holds whether the effect turns out large or small: if it is small, you still need the exposure record to demonstrate that.

Finally, the figures are snapshots. The completion rates are as of the end of fiscal year 2025 and March 2026, and the sensor algorithm status is as of September 2025. DOD estimated it would complete action on both recommendations by December 2026. If those numbers move, this report needs a dated update, and it will get one.