When called to a death on the wards, you sense upon entering that the air in the room has changed. You stand momentarily in the presence of something, then hurry away before it swallows you. Pagers chirp, monitors ping, and you re-enter the hospital’s slipstream.

When my hospitalist group conceived of an abdominal paracentesis clinic to prevent unnecessary admissions, I imagined it would feel like inpatient work: a defined problem, a controlled environment, a low likelihood of attachment.

“Hey, Karen, isn’t somebody missing on our schedule today?” I asked, one hand steadying the catheter that was rapidly shrinking my patient’s abdomen.

“I was meaning to talk to you,” she said. “We’re supposed to have Mr. J, but he no-showed his hepatology appointment last week.” Karen leaned on a doorframe, clipboard in hand. “We left two messages.”

It takes time—often 30 minutes or more—to drain ascites. Time passes with the physician mostly idle and the patient captive but awake. It creates what the Japanese call “yutori”: breathing room, time to look around, spaciousness. Because the space is shared—and perhaps because such space rarely occurs in modern medicine—it tends to foster connection.

The regulars come every two weeks. We start refilling their prescriptions. We check in between visits. Eventually, ours become the only appointments they keep.

Periodically, one of them will vanish.

The first to go was Mr. Martini, a Vietnam vet with slick white hair, dark tan skin, and a voice like a cigarette, who was with us for 2 years until, inexplicably, his ascites stopped accumulating. For a while, he came anyway to talk sports or trade paperbacks.

“Still nothing,” I’d say, pressing the ultrasound probe to his belly.

“Shoot doc, I hope you’re not trying to get rid of me,” said Martini, smirking.

Martini stopped coming to clinic, but we never forgot him. Even after his death, we’d catch each other in the hallways and revisit “Martini’s miracle”—how it defied physiology. Mostly, we just missed our friend.

We became attached to our surviving patients, relied on their presence to keep the synchronous advance of time and chronic illness at bay.

No one did it better than Roxanne, who always drove herself, arriving fully made-up for her 8 am appointments.

“Just look at this silly-ass thing,” she would say, as if the juxtaposition of a pregnant belly on a slender, 60-year-old frame was merely some cosmic prank.

Rox had a way of dismissing the specter of death, which in her case arose most often in the context of critically low potassium. She was so persuasive I began to wonder if I were overstating the problem of mortality.

“Ah, give it a rest, doc,” she’d say.

In the end, Rox drove herself to an ER close to her home with hemoglobin in the 3’s. Death caught up to her, but she didn’t make it easy.

Midway through the procedure, Karen reappeared and asked if I wanted a “welfare check” on our latest missing person.

Mr. J was our most “high-maintenance” patient. He could never get comfortable on the table and would ask the nurses every few minutes to turn up the heat, adjust his blanket, or scratch an itch. It was both endearing and grating how Mr. J treated these appointments like spa days.

Over the years, we learned of a daughter struggling with addiction, cash and credit cards disappearing from his wallet, grandkids crashing on a spare mattress. A promising development in one visit would give way to an even greater disappointment the next. Eventually, Mr. J shut down the subject.

“Can we talk about something good for a change?”

Karen’s welfare check confirmed what we half-knew already. I searched online for Mr. J’s obituary and added it to my private catalog of the deceased. For weeks, I kept imagining that he’d come back to clinic: the faded bulldog tattoo peeking out from one of his sleeves; his dark hair neatly parted, the way a schoolboy looks on picture day.

“Are you serious?” said one of my partners. “He was, like, our number one patient.”

Mr. J had broken the record for greatest volume of ascites on multiple occasions, but that’s not what my colleague meant. Mr. J’s death seemed untimely. Improbable, even.

Somehow, despite rising MELD scores and Child’s C classifications, every death seemed improbable. Physicians generally overestimate the survival of the terminally ill,Footnote 1 and this tendency seems most pronounced with our dearest patients.

A death in the hospital leaves little room for mystery. Perhaps, that is why families often seem eager to ferry the bodies of their loved ones away. But I am beginning to recognize another manifestation of death: a gradual awakening to absence rather than a sudden, definite shift. The kind of absence that develops only after a period of sustained presence, like a photograph reverting to its negative.

Patient names were changed and specific identifiers omitted to protect patient privacy. The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of Veterans Affairs or any agency of the U.S. government.