Every emergency physician can describe the patient without opening a chart: the one who has been in before, more than once, for reasons that were never quite medical to begin with, and whose history, somewhere back in the paperwork, includes a night in a holding cell. The overlap is the kind of thing clinicians trade at the nurses’ station and almost never bother to count. On July 8, a team led by Vidya Eswaran, an emergency physician at Washington University in St. Louis, went and counted it, and found that among surveyed adults who came to the ER for substance use, 44 percent had a booking in their past. The jail and the emergency room, it turns out, are working off much of the same roster.

The numbers come from the 2021–2023 National Survey on Drug Use and Health, a nationally representative sample of 139,524 civilian adults. The question Eswaran’s team put to it was blunt. Among people who had ever been arrested and booked, setting aside minor traffic stops, how often did the emergency room turn up in the past year? Adjusted for the obvious confounders, the answer was more often, and the gap widened exactly where you would expect it to.

For an ER visit of any kind, the association was modest, a 20 percent bump in the odds (adjusted odds ratio 1.20, 95% confidence interval 1.12 to 1.29). The sharper signals sat further down. For visits driven by substance use, the odds ran 82 percent higher (aOR 1.82, 1.36 to 2.43); for mental-health visits, 54 percent higher (aOR 1.54, 1.19 to 1.99). The pattern is not subtle. The closer the visit sits to addiction or psychiatric crisis, the more the booking history predicts it.

Turn it around and the same fact reads louder. Of the adults who reported an ER visit for substance use, 44 percent had been booked. Among those who came in during a mental-health crisis, 29 percent had. These are not the country’s outliers. They are a large slice of the people passing through the one medical door that never locks.

There are things worth saying in this study’s favor, and they are not the things a press release leads with. No one funded it. The authors declared no competing interests. In a field where I have read a great many papers underwritten by someone with a product to move, a study with nothing to sell makes for a cleaner measurement, which is not the same as a profound one. What you get here is an honest count, and honest counts are rarer than they should be.

What the count cannot do is tell a story about cause. A cross-sectional survey freezes one moment and asks two questions at once; it can show that a booking history and heavy ER use travel together, but not which one is the passenger. The likeliest reading is that neither is driving. The arrest did not send anyone to the emergency room. The same untreated addiction and mental illness sent them to both.

That reading gets firmer when you notice who the survey leaves out. NSDUH does not interview people held in jails and prisons, and it does not reach adults with no fixed address, which is to say it misses two of the groups whose lives run straight through the booking desk and the ER bay. The authors say as much. Whatever the real overlap is, this is the floor, not the ceiling.

Where the framing starts to outrun the data is in the prescription. Eswaran’s team recommends that criminal-legal involvement “be recognized as a social determinant of health,” and pairs it with a familiar stack of fixes bolted onto the emergency department: universal screening protocols, trauma-informed care training, medical-legal partnerships, warm handoffs to clinics, buprenorphine started in the ER for opioid use disorder. Some of that has real evidence under it. ED-initiated buprenorphine, for one, was tested in a randomized trial a decade ago and roughly doubled the share of patients still in treatment a month later, which is more than most things the ED bolts on can claim. But the reflex is worth naming. The proposal is to add more machinery to the one institution that already cannot turn anyone away, on the strength of a 1.20 odds ratio, as though the emergency department’s problem were a shortage of screening forms.

The plainer account is the one the survey keeps circling without saying. Emergency medicine has already started teaching its residents how to care for patients who arrive with the criminal-legal system attached to them, because those patients keep arriving. The country spent forty years emptying the institutions that once held people with serious addiction and serious mental illness, and never built the replacements. Two doors were left standing to absorb the overflow. This study is a measurement of how much they now share.

It is a quiet finding dressed in policy language, and the quiet part is the part the emergency department lives with every night: it has become the place where the untreated go when the jail is done with them, and where the jail’s next booking is warming up. One door locks at ten. The other never closes.

Sources

  1. PLOS ONE – Eswaran et al., “Prevalence of criminal legal involvement among emergency department patients: Insights from the National Survey on Drug Use and Health 2021–2023” (July 8, 2026)
  2. News-Medical – “Criminal legal involvement history linked to higher emergency department visits” (July 8, 2026)
  3. JAMA (D’Onofrio et al.) – “Emergency Department–Initiated Buprenorphine/Naloxone Treatment for Opioid Dependence: A Randomized Clinical Trial” (2015)
  4. Annals of Emergency Medicine – “Resident Education and the Care of Patients With Criminal-Legal System Involvement in the Emergency Department” (2025)