Seven years ago, a young woman was driving when her car was rear-ended. She suffered significant brain injury, and was in a coma for weeks. Today, she walks with some difficulty, her memory is flawed, and courts have declared her incompetent.
All of this would be of little note except that the person who rear-ended this young lady was a medical resident who had been on call and awake for 36 hours. The victim's family filed suit against the resident and against the hospital that employed her.
Anyone who has been through residency knows that hours on call exhaust doctors-in-training. It probably should come as no surprise that on-call duties might put the public in danger. Impairment associated with 24 hours of sleep loss is comparable with that associated with a 0.10 percent blood alcohol concentration. (NPR, Feb. 28, 2005, www.npr.org/templates/story.php?storyId-4512366, accessed March 8, 2005.)
In a prospective analysis during the 2002–2003 academic year, every scheduled night shift in a month resulting in a 30-hour stint increased the monthly risk of a motor vehicle crash during the commute from work by 16.2 percent. In months during which interns work five or more nights on call, their odds of falling asleep while driving were 2.39 times greater than baseline. (N Engl J Med 2005;352[2]:125.) Yet it is only relatively recently that any limits to residents' on-call duties have been promulgated, and they were not developed because those residents made the streets unsafe. This is despite the fact that drivers in the United States have been convicted of vehicular homicide for driving while impaired by sleepiness. (Massachusetts v. Salvaggio, N Berkshire County [Mass Dist CT 1994] No. 9428CR000504.) Appeals courts in two states have ruled that an employer's responsibility for fatigue-related crashes can continue after an employee has left work. (Faverty v. MacDonald's Restaurants of Oregon, 892 P. 2d 703 [Oreg 1995].)
Indentured Servitude
One may not expect much sympathy from attending physicians who have been entrenched in academia for decades and have set the rules for years. Long working hours have been a rite of passage between practically living in the hospital for very little pay to a cushy attending job with a six-figure salary. There used to be surgical programs that wouldn't hire residents who were married because that would interfere with their commitment to training, and perhaps they were right. Many have claimed that continuity of care is beneficial to patients and to the educational process, and continuity of care meant continuous oversight of one's patients during their entire hospital stay.
Surgical residents into the 1950s took call in-house three nights out of four, and were paid $50 a month for their time. Those jocks undoubtedly performed more abdominal operations than “girlie men” general surgeons today who have to have a positive CT scan before cutting out someone's appendix, who call interventional radiology to do a procedure that could just as well be accomplished with a scalpel, and who tend to worry a lot about operating where surgical pathology might not exist. The good old days when men were men also doesn't take into account the fact that house officers are paid a middle class wage today.
The Regulators
The Accreditation Council for Graduate Medical Education sets resident hours, which for the past two years have been set at 30 hours maximum at a stretch. In addition, house officers have to have a 24-hour stretch each week without patient care duties, and the weekly work-hour limit has been set at 80 hours, averaged over four weeks. Interestingly, the European Union stipulates a minimum rest time of 11 hours within each 24-hour period, effectively limiting duration of shifts to 13 hours for any physician-in-training. (European Working Time Directive, at www.incomesdata.co.uk/information/worktimedirective.htm#article3.)
The case of Libby Zion was perhaps one of the most influential in all of medical education. She was a young woman who was purported to have received substandard care by fatigued residents. A jury ultimately found the resident negligent and the hospital vicariously liable (JAMA 2004;292[9]:1051), but the case is instructive in many ways, and is worthy of review. The case had a profound impact on emergency care and attending supervision, and represented a frontal assault on graduate medical education as it existed in 1984.
Libby Zion was an 18-year-old woman who had undergone psychiatric treatment for stress. She had been treated with phenelzine, a potent MAO inhibitor. She also had been prescribed Percodan by her dentist, and for a fever and otalgia, erythromycin and chlorpheniramine. At some point in recent months, she also had received imipramine, flurazepam, diazepam, tetracycline, and doxycycline. She presented to New York Hospital with a temperature of 41°C, chills, myalgias, and arthralgias. She was evaluated by a second-year medical resident, who was not supervised by an emergency department attending staff.
She was writhing during her physical examination, and had some orthostatic blood pressure changes. After contact with her referring physician, she was admitted to the medical service, and given acetaminophen. She was agitated and shivering, and for this was prescribed meperidine. Sometime around 4 a.m., she became more agitated and confused. These symptoms were treated with physical restraints and haloperidol. She became more agitated around 6 a.m., and her temperature was noted to be 42°C. Cold compresses and a cooling blanket were ordered. At 6:30 a.m., she went into respiratory arrest and died. The medical examiner found evidence for cocaine by radioimmunoassay.
Awful mistakes in the Zion case did not occur in the ED, and would not have been remedied by 24/7 attending coverage
The investigating grand jury did not file any criminal indictments against the hospital or its physicians, but found much fault with the system of residency training that “allowed” such a death to occur. Not surprisingly, because her father was an attorney and a reporter for the New York Times, heads were bound to roll, and not necessarily in a way representing a rational response to the case. The grand jury recommended that the New York State Department of Health (DOH) promulgate regulations mandating that all Level I emergency departments be staffed with physicians with specific training in emergency care and at least three years of post-graduate training, and ensuring contemporaneous oversight of interns and junior residents by attending staff or physicians with at least three years of training. The grand jury also said the DOH should promulgate regulations to limit consecutive working hours for junior residents and interns at teaching hospitals, and conduct a study to determine the feasibility of requiring Level I hospitals to implement a computerized system to check for contraindicated combinations of drugs. N Engl J Med 1988;318[12]:771.) Finally, they said legislation should be enacted to prescribe when a patient in a medical hospital should be physically restrained, and to standardize care and attention for such patients.
Other recommendations were related to limitation of shift length: no more than 12 hours in emergency departments, 16 hours outside of emergency shifts, with a minimum of eight hours in-between. The Committee on Emergency Services called for a maximum work week of 80 hours, with 24/7 supervision of acute care units by experienced physicians.
Supervision for In-Patient Units
It seems obvious to me that the really awful mistakes in the case did not occur in the emergency department, and would not have been remedied by 24/7 attending coverage in every New York emergency department, though such a recommendation is laudable.
It has not transpired that in-house coverage on medical-surgical floors be at the attending level, even though it would be apparent from the Libby Zion case that this setting was precisely the one in which more mature supervision was most desperately needed.
It also begs the question whether attending oversight would have diagnosed and treated her serotonin syndrome. It is still a fact that serotonin syndrome is quite rare, even with the Demerol, cocaine, phenelzine, and who knows what else in Ms. Zion's body from various sources. Having a physician present with three years of post-graduate training on site would hardly be a guarantee that such a physician would recognize the syndrome or know how to treat it. It is incredible to me that all but a small fraction of physicians at the attending level would have ever seen the disorder at all.
Lastly, it is not clear that any of the mistakes in judgment, such as giving meperidine, a drug with serotoninergic action, to an agitated patient in restraints, was the result of sleep deprivation. Sleep is good, and fatigue may very well be a threat to patient safety. It certainly appears to be a threat to the safety of drivers unfortunate enough to be on the road when a resident finally drives home. To all appearances, however, the clinical decisions in the Zion case were made due to poor judgment, not to sleep deprivation. While we may be better doctors at 3 p.m. than 3 a.m., it appears that Libby Zion may very well have received the same care during the afternoon.
The Bottom Lines
We should not have to generate data to show that interns working 30-hour shifts make 36 percent more serious medical errors than those same interns who work no more than 16 hours consecutively. This should be intuitive. The question is how to make up for the inevitable decreased house officer coverage at teaching hospitals. Some obvious answers: longer and less intensive residency programs, enhanced role for physician extenders, attending staff providing more patient care.
Cutting down a work week to 80 hours should theoretically not even be worthy of note. Airplane pilots have been cited here and elsewhere as critical to public safety, similar to medical personnel. Loss of continuity of patient care as well as decreased exposure to procedures during residency may very well negatively affect the quality and nature of physicians turned out.
The Libby Zion case brought up issues that should have had little to do with emergency care, yet affected emergency medicine in a huge way, with its subsequent mandate for supervision at the attending level. The ultimate upshot of the case has probably been positive in many ways, but not because the solutions had anything to do with the problems actually encountered by Libby Zion in 1984.
If it is true that fatigue is bad, and that care is best avoided between the hours of, say, 11 p.m. and 7 a.m. because doctors are not as coordinated, knowledgeable, or quick on their feet at 4 a.m. as 4 p.m., we as a specialty may want to look into ways of discouraging patients with non-life-threatening problems from showing up during those hours. There are many ways to do so such as EMTALA modifications for off-hour screening of nonurgent patients and surcharges for nonurgent care during graveyard shifts.
