(Brett Ryder for The Washington Post)

The sun rose as the overnight shift at my emergency department (ED) came to an end. Mr. P had been with us all evening. He had arrived in an ambulance hours earlier after wandering into a homeless shelter, intoxicated. According to the information the ambulance staff delivered to us with Mr. P, he had been disruptive and argumentative at the shelter. According to Mr. P, the shelter staff members had been unreasonable, and he had just been minding his own business.

I performed a brief physical exam, describing in my notes a frail man, 56 years old and probably Caucasian, who had a thick gray beard and blue eyes dulled from alcohol, with subtle icterus (a yellowing of the eyes). His nose had been flattened from previous falls, and his skin was deeply sunburned, almost purple. The city’s dirt had merged with his epithelial layer, giving it a rough, Velcro-like appearance.

He had unremarkable vital signs, with no new trauma and no specific medical complaints. He blew a 310 on the breathalyzer. The textbooks would tell us that at that level, Mr. P — whom I’m identifying only by his initial for privacy reasons — was in serious danger of alcohol poisoning, but as we already knew, he lived at this rarefied level and stumbled and slurred only moderately.

From that point on, we both knew our roles. I would ignore him until the morning, and he would be given a sandwich and have the opportunity to sleep in a room with three other people in similar conditions. Before the development of our modern health-care system, churches and shelters would have provided similar services.

The next morning, Mr. P wanted to leave the ED — and so did I. I printed out his discharge papers, which included a list of detoxification centers, and watched as he tossed them into the trash. As he walked out, he said a pleasant goodbye to the security guards and nurses, calling them by their first names. We knew we would see him again later that day.

He’d be back

Providence, R.I., where I work, has adopted a medical response to public intoxication like Mr. P’s that involves an ambulance ride and a stay at the hospital. State laws specifically mandate this policy. If he were in another state, Mr. P might be arrested and taken to jail for the same behavior. But if he were in the same condition behind closed doors, he would be neither a patient nor a criminal. How the care for patients such as Mr. P varies from community to community exposes the paradoxical policies and philosophies behind the way alcoholism and public intoxication is treated in the United States — if it is treated at all.

At times I felt used by Mr. P. Last year, after we discharged him late one afternoon, he complained that we were not letting him spend the night. (Mr. P did not use our hospital’s formal system for lodging complaints; instead he preferred the expletive-laden diatribe.) I explained to him that he was no longer intoxicated, and I had to discharge him. He promised that he would return later that night, drunker, and I’d be forced to keep him until the morning. He kept his word.

Mr. P’s care in our ED had been routinized to a large extent. We would use a history and physical exam to screen him for acute medical and traumatic issues. Given how frequently he came to the ED, we did not routinely document his chronic liver disease, gastritis, poorly controlled seizures and chronic obstructive pulmonary disease, and blood work for these chronic conditions was not indicated.

Acute issues, such as when Mr. P fell, was assaulted or had low oxygen levels, required a more thorough work-up. We had ordered hundreds of computed tomography (CT) scans of his brain (almost always with normal results) and had often admitted him to the hospital for pneumonia and other infections. But most often, his diagnosis was what we call uncomplicated intoxication, and our ED was there to ensure his safety when he could not.

Despite the challenges of caring for him, including his intoxicated outbursts, Mr. P had become beloved by most of our staff members. The nurses and technicians had spent hours dressing his wounds, caring for his hygiene, helping him to the bathroom and listening to his stories.

He could be remarkably kind when he was not too drunk. He often asked the nurses about their families. Occasionally he gave out hugs and thanked us earnestly for caring for him. He could also be deftly funny. Like alcohol, his humor was warm and soothing, blunting the sharp edges of his personality and alleviating our frustrations in caring for him.

We learned from him, too. Dark humor became our own preferred elixir to quell the unpleasantness of Mr. P’s inevitable health trajectory and our complicity in it. When he left the ED, we often said to each other, “We’ll see him later today, unless he dies.” It was a fatalistic tonic for our guilt.

On days when Mr. P didn’t show up at the ED, we were concerned. He had become our neighbor, a sort of distant relative, subject to the same sort of worry and gossip that we would bestow upon such people. Yet for all of our worrying, we repeatedly discharged Mr. P to the street in the morning, teetering on the edge of alcohol withdrawal, a condition that could worsen and lead to tremors, hallucinations, seizures or death. Luckily, the corner liquor store opened at 7 a.m.

When is someone like Mr. P no longer a patient? When his blood alcohol concentration is zero, or when it reaches the legal driving limit of 0.08 percent? When he can walk steadily? When he decides to leave? What if it is cold outside? What if someone shows up and wants to take him home?

Liquor store vs. detox

The problem was that Mr. P was intoxicated all the time: His alcohol level never reached zero while he was in our care. And no one ever showed up to take him home. In routine practice, we considered the right time to discharge him to be the “sweet spot” after his period of significant intoxication ended and before withdrawal began. If we missed this sweet spot, we might have to admit him to the hospital for alcohol withdrawal. It was a paradox and a vicious cycle: Discharging Mr. P on the verge of alcohol withdrawal essentially guaranteed that he would start drinking immediately and would end up back in the ED later that day.

But what other options did we have? At 6:30 a.m., the answer was: Not many. In theory, there were alcohol detox programs that we could offer our patients. But these are small programs, run by community nonprofits, with a limited number of “state beds” for the uninsured. Additionally, the patient must be willing to enter the program. Mr. P had been offered detox countless times and only occasionally accepted it.

Even then, getting him into detox at the time of his discharge from the ED was nearly impossible. If he had insurance, authorization was required, and he would have to satisfy eligibility criteria. If he didn’t have insurance, a bed would have to be available for someone without insurance at the very moment he needed it. Finally, most detox programs will not accept patients until their alcohol level approaches zero. Being drunk barred Mr. P from treatment for alcoholism.

I’m sure that Mr. P found it easier to go to the liquor store than to look for a detox program. For me, discharging Mr. P in this sweet spot was also the path of least resistance. Our ED is prepared to deal with critical and complex medical conditions as well as minor ones. But Mr. P didn’t have an acute medical condition: His social condition was the real emergency.

How complicit are we as a community in Mr. P’s predicament? It was easy to blame him for his addiction and his drain on precious resources at the ED. It was harder to see how we had set up a system that took such pains to make sure that he was safe but demanded that he keep drinking to continue a lifestyle that had become comfortable for everyone.

Punitive protection

In 1968, the ability of the states to enforce laws against public intoxication — sometimes known as being drunk and disorderly — was upheld in the U.S. Supreme Court case of Powell v. Texas. Powell was an alcoholic who was often intoxicated in public and frequently arrested for this offense. As a result, he had accumulated a significant amount of fines that he was unable to pay. His lawyers argued that the state was punishing him for a disease. The Supreme Court disagreed; it ruled that making public intoxication a crime did not violate the Eighth Amendment, which prohibits cruel and unusual punishment.

By the early 1970s, the sentiment that alcoholism is a medical condition instead of a crime was taking hold. Alcoholism treatment and research had come under the umbrella of the medical profession and were increasingly being funded through the newly formed National Institute on Alcohol Abuse and Alcoholism.

In 1971, the Uniform Alcoholism and Intoxication Treatment Act was drafted. The legislation urged states to set up continuities of care, which would place a publicly intoxicated person in detox instead of jail and later place the person in a longer-term sober residential treatment facility, though the specifics were to be determined by the states. Most of them enacted laws based on the 1971 act.

Police departments and jails were more than willing to punt homeless alcoholics back to the health-care system for treatment. Soon, many states were struggling to fund the large number of detox beds they needed. At the same time, private hospitals were generally not willing to admit patients for alcoholism.

Large detox programs funded by states or cities were replaced by ones run by smaller nonprofit community organizations. With Providence having no centralized detoxification program and a robust modern ED, it was no surprise that Mr. P wound up with my colleagues and me.

Too late

Mr. P died last autumn, having collapsed at a bus stop not long after being discharged from my ED. He was in cardiac arrest when help arrived, and he was brought to us, intoxicated and dead. It was unclear whether he had aspirated vomit into his lungs, had had a heart attack, or both.

Our attempts to resuscitate him lasted for hours. In the years leading up to this moment, we had been unable to improve his life and health. Now, in cardiac arrest, Mr. P finally had a condition that my colleagues and I had been trained to treat. After multiple defibrillations, rescue medications and chest compressions, we were able to get his heart beating again, albeit briefly. Then his pulse faded, and we started over again. As time progressed, the futility of our efforts became apparent.

Mr. P was pronounced dead in a room full of people who had cared for him nearly every day for 10 years. In an ED that regularly sees death, Mr. P’s still echoes. We were his community, his friends, his family. We came to work expecting to see him, and now he had died in our hands.

As his body lay on the gurney, pale and cold, we gathered to mourn him. We lamented aloud that our final resuscitative efforts had not revived him. The truer story is that we failed him not only on the day when we could not restart his heart but also on the thousands of days when he was our patient with a beating heart and a fatal social disease.

Warren is an attending physician at University Emergency Medicine Foundation in Providence, R.I. This article was excerpted from the November issue of Health Affairs and can be read in its entirety at healthaffairs.org.