Because physician judgment varies so widely, so do treatment decisions; the same patient can go to different physicians, be told different things and receive different care. When so many physicians have such different beliefs and are doing such different things, it is impossible for every physician to be correct.
Why are so many physicians making inaccurate decisions in their medical practices? It is not because physicians lack competence, sincerity or diligence, but because they must make decisions about tremendously complex problems with very little solid evidence available to back them up. (That situation is gradually changing with the explosion in medical literature. Recent surveys by the Healthcare Information and Management Systems Society (HIMSS) reveal that an increasing number of hospitals and healthcare organizations are adopting technologies to keep up with the flow of research, such as robust, computerized physician-order-entry (CPOE) systems to ensure appropriate drug prescribing.)
Most physicians practice in a virtually data-free environment, devoid of feedback on the correctness of their practice. They know very little about the quality and outcomes of their diagnosis and treatment decisions. And without data indicating that they should change what they're doing, physicians continue doing what they've been doing all along.
Physicians rely heavily on the "art" of medicine, practicing not according to solid research evidence, but rather by how they were trained, by the culture of their own practice environment and by their own experiences with their patients.
For example, consider deep-vein-thrombosis (DVT) prophylaxis, that means therapy to prevent dangerous blood clots in vessels before and after operations in the hospital. Research offers solid, Grade-A evidence about how to prevent DVT in the hospital. But only half of America's hospitals follow these practices. That begs an important question: Why? We have the science for that particular sliver of care. How come we still can't get it right?
The core problem we would like to examine here is that a disturbingly large chunk of medical practice is still "craft" rather than science. As we've noted, relatively little actionable science is available to guide physicians and physicians often ignore proven evidence-based guidelines when they do exist. A guild-like approach to medicine—where every physician does it his or her way—can create inherent complexity, waste, proneness to error and danger for patients.
A great example comes from Peter Pronovost, MD, PhD, a patient-safety expert and a professor of anesthesiology, critical-care medicine and surgery at the Johns Hopkins University School of Medicine. He is co-author of Safe Patients, Smart Hospitals: How One Doctor's Checklist Can Help Us Change Health Care from Inside Out. In a televised interview about his book, Pronovost said that we (that is, physicians) knew that we were killing people with preventable central-line blood-stream infections in hospitals and we accepted it as a routine part, albeit a toxic side-effect, of practice. We were killing more people that way, probably, than those who died of breast cancer. We tolerated it because our practices didn't use available scientific evidence that showed us how to prevent such infections. We ignored the science and patients paid the price with their lives.
Cost is another toxic by-product of care delivery practices that are not based on solid science and the tremendous clinical variation that results from them.
Doing the right thing only half of the time
When we look at how well physicians are really doing, it's scary to see how off the mark they are. Anyone who feels self-assured about receiving the best medical care that science can offer is in for a shock, considering some eye-opening research that shows how misplaced that confidence is. Let's start with how well physicians do when they have available evidence to guide their practices.
The best answer comes from seminal research by the Rand Corporation, a respected research organization known for authoritative and unbiased analyses of complex topics. On average, Americans only receive about half of recommended medical care for common illnesses, according to research led by Elizabeth McGlynn, PhD, director of Rand's Center for Research on Quality in Health Care. That means the average American receives care that fails to meet professional evidence-based standards about half of the time.
McGlynn and her colleagues examined thousands of patient medical records from around the country for physician performance on 439 indicators of quality of care for thirty acute and chronic conditions as well as preventive care, making the Rand study one of the largest of its kind ever undertaken. The researchers examined medical conditions representing the leading causes of illness, death and healthcare service use across all age groups and types of patients. They reviewed national evidence-based practice guidelines that offer physicians specific and proven care processes for screening, diagnosis, treatment and follow-up care. Those guidelines were vetted by several multispecialty expert panels as scientifically grounded and clinically proven to improve patient care.
For example, when a patient walks into the doctor's office, the physician is supposed to ensure that when the patient shows up for hip surgery, he or she will receive drugs to prevent blood clots and then a preventive dose of antibiotics.
Even though clinical guidelines exist for practices like these, McGlynn and her colleagues found something shocking: physicians get it right about 55 percent of the time across all medical conditions. In other words, patients receive recommended care only about 55 percent of the time, on average. It doesn't matter whether that care is acute (to treat current illnesses), chronic (to treat and manage conditions that cause recurring illnesses, like diabetes and asthma) or preventive (to avert acute episodes like heart attack and stroke).
How well physicians did for any particular condition varied substantially, ranging from about 79 percent of recommended care delivered for early-stage cataracts to about 11 percent of recommended care for alcohol dependence. Physicians prescribe the recommended medication about 69 percent of the time, follow appropriate lab-testing recommendations about 62 percent of the time and follow appropriate surgical guidelines 57 percent of the time. Physicians adhere to recommended care guidelines 23 percent of the time for hip fracture, 25 percent of the time for atrial fibrillation, 39 percent for community-acquired pneumonia, 41 percent for urinary-tract infection and 45 percent for diabetes mellitus.
Underuse of recommended services was actually more common than overuse: about 46 percent of patients did not receive recommended care, while about 11 percent of participants received care that was not recommended and was potentially harmful.
Here is disturbing proof that physicians often fail to follow solid scientific evidence of what "quality care" is in providing common care that any of us might need:
• Only one-quarter of diabetes patients received essential blood-sugar tests.
• Patients with hypertension failed to receive one-third the recommended care.
• Coronary-artery-disease patients received only about two-thirds of the recommended care.
• Just under two-thirds of eligible heart-attack patients received aspirin, which is proven to reduce the risk of death and stroke.
• Only about two-thirds of elderly patients had received or been offered a pneumococcal vaccine (to help prevent them from developing pneumonia).
• Scarcely more than one-third of eligible patients had been screened for colorectal cancer.
These findings have shaped the conversation among experts on American healthcare quality by establishing a national baseline for the status quo. That baseline is jarring and disturbing. The gap between what is proven to work and what physicians actually do poses a serious threat to the health and well-being of all of us. That gap persists despite public- and private-sector initiatives to improve care. Physicians need either better access to existing information for clinical decision-making or stronger incentives to use that information.