The Hospital Treats the Emergency. Then You’re Discharged Alone With the Disease That Caused It.

Somewhere in an American hospital today, a physician will treat someone for a GI bleed, a seizure, a fall, or acute pancreatitis — and quietly note, in the chart, that alcohol caused it. The bleed gets fixed. The pancreatitis gets managed. The disease that put the patient in that bed in the first place gets a pamphlet, maybe a phone number, and a discharge order. What happens after that is, according to a systematic review published this week in Annals of Internal Medicine, mostly a guess.

We know how to stop the bleeding. We do not reliably know how to stop the next admission.

Researchers at the University of Colorado Anschutz Medical Campus reviewed 17 studies covering roughly 12,400 hospitalized patients with alcohol use disorder, evaluating 21 different interventions designed to connect them to outpatient care after discharge — everything from bedside counseling to warm handoffs to addiction-medicine consults. Only 8 of the 21 showed any measurable benefit. The overall certainty of the evidence, in the researchers’ own language, was “very low.” Ten of the twelve randomized trials in the review carried a high risk of bias. And the interventions that did seem to work varied wildly in how much they helped — anywhere from an 8-percentage-point drop in follow-up engagement to a 30-point gain, depending on the trial, with no clear pattern explaining why.

“It can feel like a revolving door,” said Dr. Eden Bernstein, the study’s lead author and an assistant professor of hospital medicine at CU Anschutz. It’s a plain sentence for something that has a name in health-policy circles — care-transition failure — and a much plainer name for the person living it: you get sick enough to end up in a bed, you get well enough to leave, and then you’re on your own with the same disease, at the same address, with the same triggers, and nothing waiting for you but whatever you can find yourself. Dr. Vineet Chopra, the study’s senior author, put it more bluntly: “We need better-designed studies to determine which components are actually necessary” — an admission that, after years of hospitals rolling out discharge-planning programs, nobody has rigorously confirmed which parts of them actually keep people out of the next bed.

The layman version of why “very low certainty” evidence is the whole story

Here’s what “very low certainty evidence” actually means in a sentence a patient can use: it means doctors have been guessing, in good faith, at what helps — and the guess has mostly gone untested rigorously enough to know if it’s right. That’s not a knock on any individual clinician. It’s an indictment of a system that treats the medical emergency alcohol causes as the whole job, and treats what happens to the disease itself, after the emergency passes, as somebody else’s problem — usually nobody’s problem at all. Fewer than half of hospitalized AUD patients get timely outpatient follow-up of any kind. Only about 7% of U.S. adults with alcohol use disorder ever receive evidence-based treatment in a given year, hospitalized or not. Alcohol contributes to roughly 180,000 American deaths annually — more than every other drug most newsrooms cover combined, and it gets a fraction of the treatment infrastructure.

If you’ve ever been discharged from a hospital with a drinking problem you didn’t get a real plan for, this study is confirming something you already knew from the inside: the system’s confidence in what happens next was never as solid as it sounded at the bedside. That’s not your failure to follow up. It’s a genuine, documented gap in what medicine currently knows how to build.

The other half of the story: the tools are finally catching up to the disease

The follow-up-care gap is a systems failure. But it’s colliding, this year, with the first real expansion of what there is to follow up to. A randomized trial published in The Lancet in May found that semaglutide — the GLP-1 drug behind Ozempic and Wegovy — cut heavy-drinking days by nearly 14 percentage points more than placebo in a 108-person trial of adults with both obesity and alcohol use disorder, alongside standard therapy. It’s not an abstinence drug; it’s a craving-and-consumption drug, which is precisely the kind of tool harm reduction has been asking pharmaceutical research to build for years. Dr. Mark Duncan, the addiction psychiatrist leading a related UW Medicine trial of a second GLP-1 drug, brenipatide, called the early signal from this drug class “the potential to provide a unique and powerful new medication treatment for alcohol use disorder unlike anything in the past 20 years.” His co-investigator, Mary Hatch, framed the population plainly: “people whose alcohol use is, in many cases, impacting multiple areas of their life, like health, relationships or work” — which is to say, exactly the population leaving hospitals every day with nowhere reliable to land.

But it’s colliding, this year, with the first real expansion of what there is to follow up to.

The part of this story that runs deeper than one study

This isn’t the first time American medicine has discharged people with alcohol use disorder into a void and called it a plan. Alcoholics Anonymous exists, in its original form, because of exactly this gap: in the 1930s, a stockbroker named Bill Wilson and a surgeon named Bob Smith found each other because the hospitals and sanitariums of their era had nothing structured to offer a discharged patient except the advice to try harder. AA was built, literally, as the follow-up care the medical system of its time refused to provide — peer to peer, because there was no other kind on the table. Ninety years later, a 12,400-patient review is confirming that the medical side of that follow-up gap never fully closed. The peer-support model filled part of the space AA’s founders found empty in 1935. The clinical-transitions half of that same space — the warm handoff, the addiction-medicine consult, the guaranteed next appointment — is still mostly improvised, hospital by hospital, discharge by discharge.

There is real money moving into behavioral health right now — HHS announced $383.4 million in new grants this week alone, spanning crisis response and substance use treatment infrastructure. Money is not the same as a tested protocol, but it is the raw material one gets built from, and it’s real, and it’s moving now.

If you’re reading this after your own hospital stay, or a family member’s, and nobody handed you a follow-up plan that felt solid: you weren’t imagining the gap. The researchers who study this for a living just confirmed it in print. What comes next — for you, for the next person in that bed — is a plan getting built in public, in real time, out of the same uncomfortable data this study just put on the record. That’s not nothing. It’s the first honest step toward a system that stops treating the emergency as the whole job.