Is there a formula for the best medical care? Physicians will tell you that for most maladies the same treatment that works in one case is very likely to work in the next. If it doesn’t, the physician will pull out the next arrow in the medical quiver. In more cases than physicians might care to admit, however, the treatment applied to two patients with the same disorder can vary enormously from region to region, among hospitals and even among clinicians within the same hospital. Researchers investigating quality of care and insurers who pay for enormously different forms of care for the same ailment are beginning to ask why such variations exist.

As a result, hospitals and some clinical researchers are striving to develop standards of care to make sure that, no matter who or where, a patient’s treatment will not deviate from accepted norms—at least not without a good reason. At the same time, many are asking whether such a drive for statistically objective standards of care won’t turn medicine into a computer-driven system that will ultimately undercut the individual and human element at its core.

With the increasing pace of medical advances and the explosion of new information available to both patients and health care professionals, academic health centers are looking ever more closely at health outcomes. The drive to hold the line on costs and to bring the most effective treatments to bear has put medical centers on notice that they must establish care guidelines. The challenge is to develop logical methods for determining what works best without removing physician experience and individual judgment from the art of medicine.

Outcomes research, a relatively recent field of inquiry, hopes to do just that by studying which approaches to patient care make the patient healthiest and happiest when all is said and done. Outcomes research strives to match the best care with the most patients, all the while keeping an eye on costs.

Such research is also one of those relatively uncommon occasions when the interests of academic investigators and practitioners coincide. “Academic research is methodologically rigorous, slower, more nuanced; it results in publications, not action,” says cardiologist Harlan M. Krumholz, M.D., an associate professor of medicine and epidemiology and a leader in the field. He is one of a number of School of Medicine faculty at the Center for Outcomes Research and Evaluation (CORE) at Yale-New Haven Hospital working to evaluate the quality and outcomes of health care from many perspectives—the patient, health care professionals, hospitals, managed care organizations and employers. “Hospital staff,” he notes, “typically don’t have the time or background to take the academic approach. They are responsible for actually taking care of people; they also have to make policy and business decisions to keep the organization running. CORE offers a wonderful synergism between projects to improve care and academic inquiries about finding the best opportunities to do that.”

Beyond Morbidity and Mortality

“Outcomes research is really a catch-all phrase,” explains Dr. Krumholz. “Traditionally, the measures were morbidity and mortality. Now we are looking at newer and sometimes less tangible domains, including cost, satisfaction, functional status and quality of life.”

For clinicians, it means working through guidelines developed to assure that each of their patients gets comprehensive, optimal care. Yale-New Haven Hospital has developed 57 clinical pathways for inpatients, for problems such as coronary artery bypass surgery, hip replacement, pneumonia, strokes and chemotherapy of standard diseases, says Edwin Cadman, M.D., a professor of medicine and chief of staff at YNHH. “We wanted to document that we were improving quality of care and the outcome of care.”

Clinical pathways and outcomes research are ideas brought over from engineering and business management theory. A clinical—sometimes called critical—pathway is a predefined set of instructions that takes a clinician through each day—sometimes through each few hours—of treating a patient with a particular diagnosis. Usually, a physician tracks progress by initialing various boxes on the chart; the pathway typically includes nursing care, medications, nutrition, rehabilitation, education and discharge planning.

One common area of medical care, treatment of heart attack, has already benefited from Dr. Krumholz’s studies of treatment-related outcomes and the development of clinical pathways aimed at improving outcomes. “A heart attack is a heart attack,” says Dr. Cadman. “There is no reason why we shouldn’t as a country expect the same therapy for a routine myocardial infarction whether it happens in Connecticut or California, but there is amazing variability.”

A recent study by Dr. Krumholz and his colleagues, published in the Aug. 19 issue of the Journal of the American Medical Association (JAMA), found that patients who received beta-blocker therapy had a 14 percent lower risk of death at one year after discharge. Despite its documented benefit, only half of 45,000 patients who were ideal candidates for beta-blockers, a possibly life-prolonging treatment, were prescribed the drug when they were released from the hospital. “Given that mortality after acute myocardial infarction is high in the elderly and that beta-blockers reduce mortality in this group, our findings reveal an ample opportunity to improve the care and outcomes for such patients,” says Dr. Krumholz. Not only were large numbers of patients not getting a potentially beneficial drug, patients in certain regions of the country were far less likely to get the drug. Use of beta-blockers ranged from a low of 30 percent in Mississippi to a high of 77 percent in Connecticut.

On Jan. 28 The New England Journal of Medicine published a follow-up study by Dr. Krumholz and colleagues including CORE’s deputy director, Martha J. Radford, M.D., and medical student Jersey Chen. They showed that much of the better performance by top U.S. hospitals in acute myocardial infarction could be explained by their greater use of two inexpensive but highly effective drugs, beta blockers and aspirin.

Yale’s outcomes researchers are helping to establish benchmarks for care, thereby decreasing such wide variability. “Harlan’s work on congestive heart failure in the elderly has changed our thinking about how to work up and manage a group of patients. It’s having a national impact,” says Ralph Horwitz, M.D., chair of the Department of Internal Medicine.

Dr. Krumholz has also published research that links emotional support to recovery for coronary patients, demonstrating that those with social networks are far more likely to recover quickly. This has led to more direct efforts by hospitals to ensure that patients, who may not have family or friends to assist them after discharge, receive more than a prescription upon leaving the hospital.

Defining a Good Outcome

While cost efficiency and the effort to develop state-of-the-art care motivate outcomes research, the goal of physicians lies elsewhere. “The system should be for patients,” says Dr. Krumholz. “The question is how to configure it in a way we can afford.”

To be sure they are getting accurate and complete information, then, is a challenge for CORE staff. “We try to use a combination of standardized, validated surveys that have been used in the field,” says Jennifer Mattera, M.P.H. ’95, CORE’s assistant director. “And we include a few open-ended questions, what I would call common-sense questions, like ‘How do you feel? Do you feel you’ve recovered? Do you feel better now than before you came to the hospital?’ Both are useful in different ways.”

What complicates matters is that patients may not value the same things as their doctors. What, for example, determines if a hip replacement procedure is a success? A surgeon may focus on technical achievement: Did the patient leave the hospital alive and well? But this explains only a small part of the outcome. Researchers want to know: What percentage of patients who had hip replacement feel they are better off as a result six months later? And how do they measure “better off,” which can vary from returning to the game of golf they love to just getting out of bed in the morning without pain? Linking perceptions of benefit with real cost reductions can help physicians formulate the optimal approaches to treatment and follow-up care.

Another question researchers are asking is why some people do better than others, even when the severity of their medical conditions is the same. “It may have to do with whether and where they had rehabilitation or if they have a family member invested in their recovery,” says Dr. Krumholz. “If we can isolate these factors, we can then identify people at risk and develop interventions for them.”

The YNHH Congestive Heart Failure Continuity of Care (C-Care) Program, a study developed by Dr. Krumholz and Joan Amatruda, R.N., B.S.N., along with a team of nurses, physicians, epidemiologists and social workers, exemplifies this approach. C-Care is designed to improve outcomes for people with heart failure, thereby decreasing their hospital readmission rates. “Our idea is to design a strategy that will improve patients’ quality of life and allow them to get control over their illness,” says Ms. Amatruda, C-Care’s director of clinical nursing services.

For the study, patients have been randomly assigned to two groups. On leaving the hospital, one group gets the standard care and a set of written instructions for what to expect and how to respond to problems. With the second or “intervention” group, Ms. Amatruda and other staff go to extra lengths to follow up and educate the patients. During home visits, Ms. Amatruda teaches patients about their illness, medications, and lifestyle and health behaviors that play a significant role in the condition. She also teaches them the early signs and signals of a problem and makes sure that patients know how to get the help they need. And she gives patients a book she wrote which summarizes the necessary information and lists community resources and how to get access to them.

With one patient in the C-Care program, part of the intervention was teaching him to negotiate the system. When heart attack-sufferer George Sholes was discharged from the hospital for a second time, he recalls that his attending physician told him, “‘When you notice or suspect you’ve got this, that or the other thing, call me.’” Mr. Sholes wasn’t clear just what that meant in practical terms. “I was pretty sure he didn’t mean simply, get in touch with me, since that’s what I thought I had been trying to do by going through the usual chain-of-command from my primary care person at my HMO.” Nor did he think he should page the physician, as a ward nurse suggested. He was feeling adrift as he was preparing to go home until Ms. Amatruda made sense of the “cultural fog.” She told him that, in the event of a problem, he was to call his attending physician’s nurses, who would contact the doctor if it were necessary. Moreover, she supplied the necessary telephone numbers.

“It’s no magic formula but absolute common sense,” explains Ms. Amatruda. Researchers will follow both groups and compare their readmission rates. If the program is successful, it can be adapted to any number of other diseases.

The Subjective Animal

One of the challenges facing outcomes researchers is assessing patient satisfaction and quality of life in a scientifically acceptable way. Some researchers hold that only standardized tools that have been proven reliable and valid when used in large populations are legitimate. But others argue that instruments such as checklists lose sight of the individual and miss important information. Moreover, some physicians view this work as a challenge to their authority and interference with their practice. They are suspicious that managed care and cost considerations are driving the research. Detractors say that medicine is an art and that clinical pathways are “cookbook medicine” that will stifle innovation.

“It’s troublesome that the psychometric academic Mafia think that they can tell you the quality of your life from some sort of checklist,” says Alvan Feinstein, M.D., Sterling Professor of Medicine and Epidemiology. “Creating these checklists for measuring health describes a status, but not quality of life. Suppose we consider someone who has had paralysis in both legs and has to walk with braces and crutches. Does that person have good quality of life? If you say no, you are denying that Itzhak Perlman has a good quality of life. There’s a difference between someone’s health status and their reaction to that status and the only way to measure that in my opinion is to ask someone directly.” To critics who dismiss such a technique as subjective, he replies: “If you want to deal with wholly objective data, study rats; it’s the subjective stuff that distinguishes people as people.”

Ms. Amatruda’s home visits for the C-Care Program demonstrate how misleading any patient-supplied information may potentially be. “One gentleman looked good on his profile,” she says, “but when I visited, he was living alone and was unable to get up to his second-floor bedroom. He had set up an old broken Barcalounger in the kitchen to elevate his legs so that he could sleep. There was nothing in the refrigerator and his stomach was growling.” Ms. Amatruda sent in a social worker; he now has a home health aide several times a week and has accepted Meals on Wheels.

On the other hand, Ms. Amatruda visited a woman whose husband had just died and who, based on her profile, looked like she was in trouble. But the home visit revealed that she was appropriately grieving and had plenty of support.

Putting Data to Work

The rise of outcomes research has been powered in part by the increasing use of computers to collect and maintain data about patient care. Physician-to-physician comparisons can be made, as well as studies of across-the board care of similar conditions. “Every order at our hospital is computerized,” says Dr. Cadman. “You cannot hand write an order and haven’t been able to do that for five years. That sets us apart from probably 95 percent of other medical centers. Because everything is retained in the database we can ask ‘How many times does Physician X order such a test?’ or ‘How many times is this test ordered in heart attack cases?’”

Having such information is invaluable. When hospital administrators learned, for example, that surgeons were ordering a daily chest X-ray for patients undergoing a particular surgical procedure—and that the X-ray almost never showed any change—they were able to report back to the surgeons and eliminate the practice. “In a case like that, there’s no diminution in quality of care and you can reduce costs substantially,” says Dr. Cadman. It can also be used as a teaching aid. For example, cardiac surgeons discuss how they treat patients, debate the merits of various practices and come up with a consensus, which then becomes the “Yale” standard of care. Once developed, the standards are available on the computer system. “It’s a great teaching tool for interns and residents,” says Dr. Cadman.

In addition to individual patients and clinical treatments, CORE staff look at the bigger picture: “How should resources be allocated?” asks Dr. Krumholz. “Where are the gaps? What care ought to be delivered and how?”

Outcomes research can help to fill in holes in medical care. For example, says Dr. Krumholz, “People with a condition called atrial fibrillation can be treated with blood thinners to prevent strokes. But the number of people getting this treatment is small. Why?” The aim of outcomes researchers in this instance is to identify the gap and to increase the number of patients being treated appropriately.

Although Dr. Krumholz and colleagues acknowledge the potential for problems, they argue that the clinical pathways are not static. Rather, they represent the optimal suggested practices based on current knowledge. In the case of Yale-New Haven Hospital, teams of local expert clinicians worked exhaustively over several years to come up with the 47 existing pathways, a process coordinated by Dr. Cadman and colleagues Leo Cooney, M.D., and Victor Morris, M.D. Are those pathways now complete? No, says Dr. Krumholz, and they never will be. “We have to continually reevaluate the best practices and new treatments,” he says.

Besides, a physician is free to depart from the standard. “But if he or she deviates there must be a reason,” says Dr. Cadman. “Maybe that particular patient is sicker or there is another complicating feature.” And because any variation must be documented and justified on the patient’s chart, physicians and the hospital gain a measure of legal protection in the event of a malpractice claim.

While deviations from the standard may need to be justified, they aren’t necessarily bad. “Not offering a blood thinner to someone with atrial fibrillation because no one thought about it is the variation we aim to get rid of,” says Dr. Krumholz. “But tailoring an intervention to an individual patient’s preferences—we never want to extinguish that kind of variation.”