Medical care often not optimal, study finds
A study published in the New England Journal of Medicine says Americans have a slightly better than 50-50 chance their medical problems will be addressed in an optimal way when they visit a doctor's office or enter a hospital.
By David Brown
Washington Post Staff Writer
Thursday, June 26, 2003; Page A02
Americans have a slightly better than 50-50 chance their medical problems will be addressed in an optimal way when they visit a doctor's office or enter a hospital, according to a new survey.
The failure to do the right thing -- or, more precisely, all the right things -- extends across the spectrum of activities physicians are expected to perform.
Recommended "best practices" were followed about two-thirds of the time in diagnostic testing, prescribing drugs for acute and chronic illnesses, and monitoring patients' long-term health. In the area of counseling and health education, there was a 1-in-5 chance patients would get everything experts say they should.
The quality of treatment differed markedly by disease, with the best performances seen in breast cancer, certain forms of heart disease and low back pain. For pneumonia, bladder infections, diabetes and peptic ulcers, however, fewer than half of the recommended best practices were followed. Most of the time, the problem was that physicians and nurses did not do or ask enough. But with some conditions, such as migraine headaches, patients were overtreated.
The study, being published today in the New England Journal of Medicine, sketches a dark and disturbing portrait of American health care. It adds to the rapidly growing body of research showing a huge gap between what is known by medical scientists and what is done by medical practitioners.
"Everyone is at risk of failing to get care that they need to live a longer and healthier life," said Elizabeth A. McGlynn, a researcher at Rand, a consulting company in California, who headed the study. "It is time to stop having a debate whether we have a problem, and start having a talk about how we can solve the problem."
Although many studies have shown that it takes a long time -- often more than a decade -- for the majority of clinicians to adopt practices of proven worth, several experts not involved with the Rand survey were nevertheless surprised by the low level of performance it found.
"The bad news is just how bad the results are. The good news is that there is a lot of work going on in this area," said Carolyn M. Clancy, head of the Agency for Healthcare Research and Quality in the Department of Health and Human Services.
"If auto repair defect rates were the same as this, we wouldn't be alive today," said Donald M. Berwick, a pediatrician who heads the Institute for Healthcare Improvement, a nonprofit organization in Boston. "This is something that the public ought to be very concerned about. We ought to set a national agenda for dramatic improvement of care."
The study built on a previous survey that asked 20,000 randomly chosen adults in 12 metropolitan areas where and how they received medical care. In this study, they were asked to name their physicians and consent to the release of their medical records for the previous two years. A brief medical history was also taken over the phone. Ultimately, copies of hospital charts and clinic notes from about 40 percent of the people surveyed were sent to Rand researchers.
Twenty nurses then reviewed the records, looking for evidence that specific interventions were done -- or, in some cases, avoided -- in people with particular medical histories, conditions, symptoms, findings on physical exam, habits or laboratory results.
The recommended interventions were chosen by experts, based on strong evidence of value or harm found in scientific studies. In the case of some conditions, there were many -- 37 for coronary artery disease, 27 for high blood pressure, 25 for asthma. In others there were few -- 5 for alcohol dependence, 5 for pneumonia, 3 for arthritis. In all, there were 439 on the list.
The percentage of the time that patients got the recommended treatment for a selection of conditions was: cataracts, 79 percent; breast cancer, 76 percent; prenatal care, 73 percent; low back pain, 69 percent; coronary artery disease, 68 percent; hypertension, 65 percent; congestive heart failure; 64 percent; depression, 58 percent.
Performance was worse in treatment of diabetes, in which patients got 45 percent of recommended care; peptic ulcer, 33 percent; hip fracture, 23 percent; alcohol dependence, 11 percent.
The researchers also looked at performance based on general type of intervention. Medication choices followed recommended practices 69 percent of the time; immunizations, 66 percent; physical examination, 63 percent; and lab testing, 62 percent. However, physicians asked key questions while getting the medical history for the patient 43 percent of the time. Adequate counseling and teaching were done 18 percent of the time.
Some of the oversights found in the survey involve recommended practices with major effects on mortality that have been known for years. For example, 61 percent of people with heart attacks (myocardial infarction) received aspirin, which reduces the risk of death by 15 percent. This has been well established since 1988.
Among people with the abnormal heart rhythm called atrial fibrillation, the study found that 57 percent of those younger than 65, and 43 percent of those older than 65, were prescribed the anticoagulant warfarin. A study published in 1990 showed that giving the drug to people with that condition lowered the annual risk of stroke from 3 percent to 0.4 percent.
Several experts said that blame for poor performance should fall not primarily on practitioners.
"I don't think that this study should be read as an indictment of physicians or nurses. They are simply working in a care system that is incapable of supporting excellence, a system we have designed for failure," Berwick said.
Among the many urgent changes that are needed, he said, is much better use of information technology to track what has been done for patients, remind physicians what needs to be done, provide information about best practices in real time, and measure clinician performance on a regular schedule.