Medical Examiner

The Nightmare of Night Float

Is an ignorant doctor really better than a tired one?

Are limits on doctors’ work hours good for patients?

On my first evening as a night-float intern at Memorial, the world-famous cancer hospital in Manhattan, an intern handed me a list of her patients with their major medical problems, allergies, and a short summary of their hospital course. “There is one patient I have to tell you about,” she said almost parenthetically. A patient with colon cancer had been hallucinating all afternoon. “He’s quiet now, so he shouldn’t give you any trouble,” she quickly added. “But if he does, just snow him with more Haldol and Ativan.” Then she left.

Night float was the scene of my worst moments as an intern. Your shift began at 5 p.m., when the other interns departed for the day, and ended at 7 a.m., when they returned. Meanwhile, you had to make critical decisions for other doctors’ patients, about whom you knew next to nothing. You’re an inexperienced intern, tackling potentially serious problems with not enough information.

That first night at Memorial, within minutes, my beeper went off. The patient whose care I had just assumed responsibility for was delirious, and his blood-oxygen saturation was dropping. When I went to his room, he was sprawled in bed, his arms and legs tied to the rails. He apparently did not speak English—apart from obscenities—because a German translator was there, grinning nervously. “He says that things are coming down at him,” the translator said. “He feels that things are crawling on his skin.”

A nurse asked me what I wanted to do. I had no clue. About the only thing my colleague had said to me before leaving was that this patient wasn’t going to give me any trouble. I asked about his base-line mental state. The nurse shrugged. “I’m just a float,” she said, meaning that she worked only per diem shifts. “I’m meeting him for the first time, too.”

When I finally called the delirious patient’s family, hoping for a clue, his daughter informed me that he had undergone a brain scan that afternoon. His intern had forgotten to mention it. What were the results? I did not know. So I gave the patient more Haldol and hoped for the best. (Later, I learned the scan showed a vaguely abnormal speck in the brain—a possible metastasis—which could have explained the delirium.)

Night float is the product of reforms in medical education that limit the number of hours that residents and interns—doctors in training—can work. Because they can no longer rely on the same doctor caring for a group of patients day and night, teaching hospitals have had to arrange more cross-coverage when the primary resident is not on duty. Most have created the position of a resident who works the night shift, usually for a few weeks. The upside is that other residents can sleep. The downside is frequent patient handoffs, which can result in the transfer of faulty or inadequate information. The nightmare of night float raises a central question about work limits for interns: Is it better to be cared for by a tired resident who knows your case or a rested resident who does not?

The push to limit interns’ and residents’ work hours gained momentum with the death of a woman named Libby Zion at the emergency room of New York Hospital, after the intern and resident treating her were slow to respond when she reacted adversely to a drug they gave her. If the young doctors had been more rested, soul-searching medical educators asked themselves, would they have been able to save her? In 1987, a special commission proposed a number of changes in residency training in New York state. Residents were prohibited from working more than 24 hours at a stretch or more than 80 hours per week, averaged over four weeks. They also got one day off a week. After intense debate, in 2003 similar changes were instituted at residency programs throughout the country.

At first glance, reducing the number of hours that residents work would seem a no-brainer. In a survey of American medical residents, 41 percent reported fatigue as a cause of their most serious mistakes. Studies have shown that residents after a call night score lower on tests of simple reasoning, response time, concentration, and recall. Indeed, a single night of continuous sleep deprivation has been shown to be roughly equivalent to a blood alcohol level of 0.10 percent—that is, being drunk.

Once, as a sleep-deprived intern, I had to take an elderly woman with severe angina for a CT scan in the middle of the night. When we arrived in the radiology department, I made what seemed like a reasonable decision: I stopped my patient’s IV drips to get her onto the radiology table. Midway through the scan, she started moaning because of severe chest pains. I suddenly realized that I had stopped her nitroglycerin drip, used to treat angina, and that she was in the early stages of a heart attack. I tried to get the drip restarted, but the IV machine just kept beeping, mixing with her groans. Panicking, I raced the stretcher alone back to the cardiac-care unit, getting lost on the way. I finally got her back to the CCU, and experienced nurses took over. My patient ended up fine.

As harrowing as that experience was, it was nothing compared with night float, in which one was operating from a position of ignorance, in the environment of a teaching hospital, which reveres knowledge and competence. That first night at Memorial, I went to see a patient with esophageal cancer and intractable hiccups. Walking into his room, I felt almost relieved. After what I had dealt with so far that night, hiccups seemed almost laughably unserious. But these were no ordinary hiccups. They had been going on for more than 24 hours, leaving the patient sleepless and utterly demoralized.

I didn’t know what caused hiccups, let alone how to treat them. When I asked a nurse, she mentioned that a drug called chlorpromazine was sometimes used, so I wrote an order for it. Walking through the nurses’ station, I casually checked the patient’s chart. There, amid his papers, was a brief note. He had once suffered a severe reaction to this particular drug. It wasn’t documented as an allergy on the sign-out sheet I’d gotten but was scribbled in a progress note. I immediately canceled the order, relieved that I had caught the mistake in time but alarmed at how easily it might have slipped through.

Night float felt worse to me than working when I was exhausted, but is it really worse for patient care? The data are mixed. A study published in 2004 in the New England Journal of Medicine showed that interns working in an intensive-care unit made 36 percent more serious medical errors during a traditional schedule as compared with a schedule that eliminated extended work shifts and reduced the number of hours worked per week from 80 to 63. On the other hand, a study in the Journal of the American Medical Association appeared to indict the cross-coverage hospitals have been relying on to conform with the work limits. It showed that increasing cross-coverage in a large urban hospital caused delays in tests and an increased number of complications that could have been prevented, like drug reactions and infections. Work limits have other troubling consequences as well, including interruption of resident learning, fracturing of traditional hospital teams, and the creation of a kind of shift-work clock-watching mentality among young doctors.

If tired residents hurt patients, but the ignorance of night float and cross-coverage also pose a danger, what should hospitals do? No doctor can work 24 hours a day, seven days a week, so cross-coverage is essential. The optimal system would provide rested night floats with all the information they need. The best way to accomplish this is for teaching hospitals to have standardized, electronic handoff systems. In medicine, as in aviation, most errors occur at transitions: by pilots, during takeoff and landing, and by doctors, after handoffs. Because of work limits, an intern today might be involved in more than 300 handoffs during an average monthlong rotation. Too many hospitals continue to rely on one intern signing out verbally to another, an invitation for error. Less than 5 percent of hospitals have electronic handoff systems in place.

Without better handoff systems, work limits may well weaken medicine more than exhausted residents ever did. As a doctor in training, you have to see a patient’s illness through its course—observe the arc—to get a grip on the dynamics of disease. It is possible to overcorrect for even the most serious of problems. And in trying to get young doctors a bit more rest, we may have come up with a cure that is worse than the disease.