Read an Excerpt
Demanding Medical Excellence
Doctors and Accountability in the Information Age With a New Afterword
By Michael L. Millenson The University of Chicago Press
Copyright © 1999 The University of Chicago
All rights reserved.
ISBN: 978-0-226-52588-4
CHAPTER 1
Some Do, Some Don't
Science is the knowledge of consequences, and dependence of one fact upon another.
— Thomas Hobbes, 1651
There is an unsettling, if little known, truth about the practice of medicine. Even the best-trained doctors go about their work with an astonishingly shallow base of knowledge concerning the link between what they do and how it affects a patient's health. Whether the setting is a small town, a suburb, or the heart of a big-city medical center, study after study shows that few physicians systematically apply to everyday treatment the scientific evidence about what works best.
Why don't patients typically get the best possible care? And how should the practice of medicine change so that they do? The problem begins with the fact that physicians generally know little about the practices of doctors outside their immediate professional circle. Busy with their own patients, they mostly assume that similarly trained colleagues provide the same kind of care and get the same results. Rapid changes in the technology and economics of medicine, however, have persuaded some doctors and hospitals to take a fresh look at rituals of care that are often taken for granted. In Maine, eight hospitals decided to come together and track their treatment of patients with pneumonia or heart attack, two very common conditions that are also life threatening. The hospitals would then compare notes on what they found. The results in this one small state showed just how deceptive the surface similarities of American medicine can be.
On a dark and drab evening in late fall, an internist from one of the hospitals drove about an hour down the Maine coast to an old inn where doctors and pharmacists from the other institutions were gathered. Richard Kahn had spent twenty-two years as a family doctor in the small town of Rockport, but he still stood out from the crowd. Kahn's bushy gray beard made him physically unmistakable, while his animated manner cut through ingrained local reticence. Chatty rather than taciturn, opinionated rather than reserved, the fifty-four-year-old internist occasionally liked to create a small stir. When a patient exhibited a bit too much faith in his healing powers, Kahn might take out a nineteenth-century brass fleam that he carried in his pocket. He would unfold the knifelike object and flourish the razor-sharp thumbnails on each blade tip that were designed to slice open a vein so the physician could practice therapeutic bloodletting. "This was considered good medicine," the amateur medical historian would explain as the patient eyed the blades warily. Or Kahn might pass on a former professor's sardonic attitude about the latest pharmaceutical miracle: "Use a new drug when it comes out, before it develops side effects."
Kahn was a veteran of medical committees; he served on a national committee of internists interested in medical ethics. He quickly signed up to represent Rockport's ninety-five-bed Penobscot Bay Medical Center in the examination of pneumonia and heart attacks. The first study results were to be unveiled at this evening's meeting. Each institution was identified by a code number to encourage candor, but the open display of sensitive clinical information still seemed a bit like undressing in public.
It didn't take long for the first provocative nugget of information to emerge. All eight of the hospitals treated pneumonia with a class of antibiotics known as cephalosporins, which are similar in structure to penicillin. But though the hospitals treated the same kind of patients, the strategies their staffs chose for using these powerful drugs seemed like a medical version of "Goldilocks and the Three Bears." Some liked expensive and sophisticated drugs; some liked inexpensive, less-sophisticated drugs; and some liked a strategy somewhere in between. Unfortunately for patients, it was unclear which therapy was "just right."
At one end of the spectrum, the doctors at Hospital 3 relied heavily on the massive firepower of sophisticated third-generation cephalosporins. These broad-spectrum antibiotics are the assault weapons of the drug world, destroying almost everything in their path. That's reassuring for doctors who want to prevent the escape of whatever bacteria caused the pneumonia. But bacteria that do survive broad-spectrum antibiotics tend to become "bulletproof" — resistant to almost any antibiotic therapy at all. To prevent resistant bacteria from spreading throughout the population, medical experts strongly counsel judicious use of the most sophisticated drugs.
Despite this advice, the medical staffs at the eight hospitals took very different approaches in prescribing antibiotics. A pneumonia patient entering Hospital 3 for treatment was about twelve times more likely to receive third-generation antibiotics than a patient at Hospital 5, who was the most likely to get a somewhat older second-generation antibiotic. A patient who entered Hospital 8 was the most likely to receive first-generation cephalosporins. Yet at Hospital 1, pneumonia patients received no first-generation cephalosporins at all! Researchers looked in vain for any consistent pattern.
The treatment of heart attacks raised even more troubling questions. In a typical attack, a clot in the coronary arteries blocks the flow of blood, and the heart muscle, starved of oxygen, immediately starts to die. If doctors can quickly inject a drug to break up the clot, the victim's chances of survival rise dramatically. Similarly, if doctors administer a drug to thin the blood, the odds drop that another clot will form and threaten the heart. For individual medical reasons, not every patient can be treated this way. Nonetheless, the eight hospitals' uneven performance left some physicians and pharmacists shifting uneasily in their seats as they watched the numbers projected on a screen.
At one of the hospitals where the doctors seemed determined to be aggressive, about two-thirds of heart attack victims under age seventy-five got thrombolytics (clot busters) — nearly twice the rate of thrombolytic use at the least aggressive hospital. At the most aggressive hospital, some 94 percent of all heart attack patients received an anticoagulant. At the least aggressive institution, only 59 percent got a blood thinner. Just as with pneumonia, the remaining hospitals showed no discernible pattern.
These hospitals that looked so much alike to patients and staff varied enormously in the treatment they provided. But why? And what was the effect on patients?
The variation could mean that the researchers conducting the study were mistaken and that the hospitals each treated very different kinds of patients. This didn't seem entirely plausible, however, when one hospital was administering clot busters to twice as many heart attack patients or giving sophisticated antibiotics twelve times as frequently.
The variation could reflect the mix of doctors at each institution. A specialist is more likely than a generalist to keep up with the latest research and use the newest drugs and equipment. Some hospitals could have been using cardiologists to treat heart attacks or pulmonologists for pneumonia while other hospitals relied on generalists. The flaw in that theory was that the hospitals didn't neatly divide into the high-use and low-use groups typical of the gap between specialists and other doctors.
The third possible explanation for the variations was the most unnerving. The differences could result from physicians' "practice styles" or "preferences," the polite terms the medical community uses to describe treatment that varies because doctors vary. This constitutes not so much an explanation as a tautology. The words "style" and "preferences" have an innocuous ring. In our consumerist society, it seems sometimes that the Declaration of Independence implicitly blessed each citizen's right to life, liberty, and the pursuit of styles and preferences. The consumer who prefers pasta over pizza, however, affects only her own life; the doctor who favors one antibiotic over another affects the lives of her patients.
In Maine, the small organization coordinating the variation study for the eight hospitals did not have either the time or the money to ask many questions about how treatment differences ultimately affected the health of the patients. The study's entire budget came to $30,000, or about the cost of one hospital's buying one basic ultrasound machine. Put another way, the money spent examining the treatment of nearly one thousand people who were hospitalized with pneumonia or a heart attack amounted to some $30 per patient.
As the discussion in the Maine inn neared its end that evening, Kahn raised his hand to make a final point. The information on variations in treatment should be presented to hospital leaders, he declared. And it should be given to the federally funded "peer review organization" that oversees the cost and quality of the medical treatment provided to Medicare patients. "If you're just doing this" (here Kahn turned and gestured to the group of about a dozen people), "you're not accomplishing anything. The whole idea is to change behavior and make medicine better."
The entire purpose of medical treatment, of course, is to make patients better. If some practices or preferences are more likely to lead to that result, then physicians and hospitals should be pushed to adopt them. If some styles of treatment are less successful, they should be abandoned.
Discovering "what works best" in medicine, however, is slow and difficult work. The answers are not intuitively obvious, and the search is not very glamorous. There are no teams of physicians working frenetically into the night, and there are no dramatic moments of discovery that would look good on a movie screen. The only exciting thing that happens is that some people get well faster and some people who would have died go home to their families alive.
CURING PNEUMONIA
In the middle of 1989, the medical director of Pittsburgh's Forbes Health System received disturbing news. According to an internal report, an unexpectedly large number of Forbes patients were dying.
A hospital, like a bank, depends for its existence on public trust. It's nice if the bank is open for business during convenient hours and the tellers are friendly, but a bank's basic purpose is to ensure that depositors get their money back when they need it. Similarly, though hospitals may sponsor "wellness fairs," their fundamental mission is to heal the sick. Patients joke about bad hospital food; bad care isn't funny.
The community served by Forbes was unaware of any problem. The information came from a private data company, Medi-Qual, that held a contract to analyze the cost and quality of the system's care. The issue was not malpractice; the information Medi-Qual provided dealt with much more subtle medical behavior. By examining several clinical measures indicating how ill Forbes patients were when they entered the hospital, Medi-Qual calculated an expected patient death rate. It then compared the number of expected deaths with the number that actually occurred. In addition, Medi-Qual compiled a national registry of its customers that showed which hospitals did the best job of reducing the ratio of actual to expected deaths.
The formula Medi-Qual uses to determine the expected deaths is far from exact. Some experts believe it does a poor job of accounting for the complex effect that sickness has on a patient before treatment. Other experts, though, believe the formula provides at least a roughly accurate confirmation that a hospital is doing well or, conversely, a warning that the quality of its care is lagging. Forbes's medical director, Richard McGarvey, was well aware of both viewpoints. McGarvey decided to investigate on his own whether the numbers on his desk were telling the truth. To do that, he went directly to the medical charts of some patients who had died and read the scrawled notes detailing the story of their treatment in the hospital. When he finished, he called together Forbes's key physicians. "Maybe we can explain to the news media and you folks that we don't buy this comparison, but when I look at these charts, we're not doing things quite the way I remember talking about doing them in medical school," McGarvey told the group. "Let's see if we can improve things."
The flagship of the Forbes system is the 353-bed Forbes Regional Hospital. Situated in the rolling hills of Monroeville, a prosperous middle-class suburb fourteen miles east of Pittsburgh, Forbes Regional is a teaching hospital that prides itself on state of the art treatment. But according to the Medi-Qual data, the largest cluster of unexpected deaths occurred on the floors where it treated pneumonia patients.
No hospital likes finding problems with its care. The typical first reaction is outrage and denial. The second reaction is to form up a professional posse to hunt down and "hang" the doctors or nurses responsible for tarnishing the institution's honor. Unfortunately for that approach, the Forbes task force couldn't find any obvious villains. At first glance, pneumonia care seemed to meet all the hospital's standards. Puzzled, the group went back and reexamined the situation in a different light, this time dividing the pneumonia patients into "high-risk" and "lower-risk" groups. Looked at from this new angle, a grim pattern immediately emerged.
The high-risk group included many of the "old-old"; that is, patients over eighty. They typically contracted pneumonia in the hospital or at a nursing home. Frequently they had serious underlying diseases such as lung cancer or an infection of the bloodstream. These high-risk patients were dying about as frequently as expected. The lower-risk patients, on the other hand, generally started out younger and healthier. They usually contracted pneumonia out in the community, then entered the hospital as their condition worsened. It was these lower-risk patients who were dying at nearly twice the expected rate. During one twelve-month period, for example, ten lower-risk pneumonia patients were expected to die. Instead, eighteen did.
An average of one "extra" death every two months could easily elude the notice of busy physicians and nurses caught up in the daily patient-by-patient struggle against disease. Even if it attracted their attention, they might understandably attribute it to an unlucky run of patients who failed to respond to therapy. The task force suppressed the urge to "explain the adverse outcomes on the basis of demographical, clinical and life-style differences," as McGarvey and assistant medical director John J. Harper put it. Instead, they followed McGarvey's example and randomly selected a number of medical records of pneumonia patients who died and an identical number of records of pneumonia patients who survived. What the reviewers found was a vivid example of the link between mundane medical processes and patient survival.
In its most severe form, bacterial pneumonia is a frightening disease. Microbes invade the lung's microscopic air sacs, causing swelling and inflammation. Patients run a high fever, and they endure attacks of coughing while trying to clear their clogged lungs of a fiber-like fluid. Their breathing becomes shallow, rapid, and painful, and their pulse rate soars. As the body's ability to take in oxygen and rid itself of carbon dioxide slowly diminishes, the brain may die or the heart erupt in violent fibrillation and then stop.
Physicians try to prescribe the type of antibiotics most suited to kill whichever bacteria are causing the pneumonia infection. One way of identifying those bacteria is to analyze a sputum sample from deep within the patient's lungs. A blood culture is even better: in addition to confirming the sputum results, a blood sample shows whether the infection in the lungs is spreading dangerously throughout the entire body.
Although the staff at Forbes Regional knew all this, the hospital wasn't well organized to effectively fight the war on pneumonia. Busy nurses seeking a sputum sample left a cup by the bedside and asked patients to fill it "when they could." What often ended up in the cup was useless saliva. Meanwhile, a blood sample might not be collected until the patient had already been in the hospital for a full day. Though blood and sputum tests don't always provide reliable information on the source of infection, the Forbes staff was unnecessarily flying blind. Worst of all, some patients were spending ten to fourteen hours in the hospital without receiving any antibiotic therapy at all.
(Continues...)
Excerpted from Demanding Medical Excellence by Michael L. Millenson. Copyright © 1999 The University of Chicago. Excerpted by permission of The University of Chicago Press.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
