Researchers tracked 1.3 million veterans through 29 drug shortages. Who lost access had almost nothing to do with who they were.
When a drug goes on shortage, you would guess the people who fall through the cracks are the usual ones. Rural. Older. Poorer. VA researchers checked. That is not who fell through. What predicted trouble wasn't printed on the veteran's record. It was printed on the bottle.
The Setup
Tracking every refill through 29 shortages.
Between 2017 and 2020, 29 chronic-condition medications went into shortage inside VA, from blood pressure pills to antidepressants to glaucoma drops. Researchers pulled the refill records of every veteran actively taking one and watched the six months before and after each shortage hit. Did the refills keep coming, or did they stop?
1.3M
Veterans tracked
29
Medications that went into shortage
12 mo
Refill window around each shortage
131K
Users with shortage-pattern disruption
What Happened
Four patterns. Most held. One didn't.
Refill histories sorted cleanly into four groups. The good news is the big green block: for about seven in ten, the shortage never touched them. The orange block is the story: veterans filling like clockwork whose refills started dying right as the shortage was reported, then went to zero.
69.2%
14.1%
8.5%
8.3%
69.2%
High adherence
Refills stayed above 80% coverage straight through the shortage. The system absorbed it. They likely never knew.
14.1%
Moderate adherence
Hovered around half coverage before and after. Imperfect refill habits, but the shortage didn't visibly change them.
8.5%
Shortage-pattern disruption
Steady, reliable refills that collapsed right at the shortage report and fell to zero. Over 131,000 medication users. This is the group that matters.
8.3%
Already falling off
Refills were dying two to three months before the shortage was ever reported. Whatever stopped them, it wasn't this.
The Number That Tells the Story
Count the factories, not the demographics.
Here's the part worth sitting with. Sex, race, income status, rural address: after adjustment, none of it moved the needle much. The strongest predictor of losing your medication was how many manufacturers make it. A drug with one to three makers carried nearly four times the odds of disruption compared to a drug with twenty or more.
Odds of shortage-related disruption, by number of manufacturers
Adjusted odds ratios vs. drugs with 20+ manufacturers. Longer bar means higher risk.
1 to 3 makers
3.64x
4 to 9 makers
2.73x
10 to 19 makers
1.86x
20+ makers
1.0x
Drug type mattered too. Glaucoma eye drops and antipsychotics carried roughly twice the odds of disruption. The one patient factor that stood out: veterans under 45, at about 2.5 times the odds of those 75 and up.
The List
These weren't exotic drugs.
The first question anyone asks is whether their own medication is on the list. These shortages hit the workhorses of the VA formulary, the pills sitting in millions of medicine cabinets right now.
Lisinopril
Blood pressure
745,000 users, nearly half the study
Sertraline
Depression, PTSD
256,000 users
Atenolol
Blood pressure
171,000 users
Rosuvastatin
Cholesterol
105,000 users
Buspirone
Anxiety
72,000 users
Timolol
Glaucoma drops
26,000 users, twice the disruption odds
Also on the list: valsartan, famotidine, hydroxychloroquine, tacrolimus for transplant patients, and several antipsychotics including haloperidol and fluphenazine.
Read It Straight
What this study doesn't say.
The honest edges, before anyone turns this into a headline about veterans abandoned by the system.
Seven in ten veterans sailed through without a visible blip. VA's mail pharmacies track supply problems and quietly reroute prescriptions or trim fills to 30 days so the supply stretches. Many shortages never reach the patient at all. That's the unglamorous machinery doing its job.
The study reads refill records, not minds. A refill that stops at the shortage date looks like the shortage did it, but some of those stops were side effects, switches to a different drug, or a doctor's call. The timing is suggestive, not proof, and the authors say so plainly.
This paper shows refills stopping. Whether those stops led to strokes, hospitalizations, or worse is the next study. Prior work on the valsartan shortage found increased cardiac events, so the question is live.
VA is an integrated system with its own pharmacies and national coordination. If 131,000 users slipped through here, the picture at your corner drugstore, with none of that machinery, is probably not better.
The Takeaway
A veteran can do everything right. Take the pill every morning, refill on time, show up to every appointment. And still find the shelf empty, because of a decision made years earlier in a boardroom about how many factories would make that molecule. The risk isn't in the patient. It's in the supply chain, and it was sitting there before anyone got sick.
Most of us never see the machinery that keeps the bottle full. This study is what it looks like when, for 131,000 people, the machinery missed.