California Pacific Currents 2005
Evidence-based Medicine Leads the Way to Better Care
Traditionally, physicians selected diagnostic tests or therapeutic approaches favored by their mentors or recommended in text books. In recent years, however, researchers have demonstrated that conventional wisdom can be wrong, leading to poor medical decisions.
Evidence-based medicine (EBM) addresses these problems by making the results from clinical research more accessible to physicians. To find the information necessary to deliver the most optimal care to patients, physicians must learn to search medical literature databases such as PubMed. This is precisely what the EBM course at California Pacific aims to accomplish.
Learning the Ropes
At California Pacific Medical Center, Daniella Zipkin, MD, and Gayatri Khanna, MD, have developed an EBM course that residents take during their second and third years of study. "It's important to make sure that our diagnostic and therapeutic practices are in line with the best evidence," Zipkin explains. "I want the residents to have the necessary skills to find and interpret information to answer clinical questions."
Zipkin learned about EBM via a six-week clinical epidemiology course in her residency at New York University; after NYU she accepted a fellowship at UCSF before coming to California Pacific. Setting up the EBM courses has presented challenges, she acknowledges. "Physicians in general, and residents in particular, have a hard time incorporating the process of searching the medical literature into patient care," she explains. "But evidence-based medicine can be done quickly and easily if you have the vocabulary and know what to look for." Zipkin can now do a basic search-forming a question, searching PubMed, pulling down an article, and then giving it a quick read to evaluate its relevance and strength-in about 10 minutes. Her biggest challenge is teaching residents these skills in just four sessions of one or two hours each.
Residents begin the process by learning to ask a well-formed clinical question about a patient in their care that has them puzzled. The question should include information about the patient's demographic characteristics, risk factors, medical history, possible diagnosis, and the nature of the proposed intervention or diagnostic tests. When they have learned how to frame clinical questions, residents are then schooled in assessing different research study designs, together with their strengths and weaknesses. These include randomized clinical trials, cohort studies, cross-sectional studies, and case-control studies, as well as variations on the standard research designs. A technique called meta-analysis, which evaluates groups of other studies and draws conclusions based on their results, may be particularly helpful. "Once you have your question, you have to know the optimal study designs for answering it," Zipkin says. "That's crucial; you can't do EBM without it."
To diagnose a problem, residents generally must order laboratory tests. Residents are taught how to evaluate the results of these tests using a concept known as a likelihood ratio. "This is the key concept of the diagnostic testing section," Zipkin says. "It's the ratio of the likelihood of getting a particular test result in a patient who has that condition compared with a patient who doesn't. The higher the likelihood ratio, the closer I get to a diagnosis when the test is positive." More common terms such as "sensitivity" and "specificity" also refer to test properties, Zipkin adds, but aren't easy to apply clinically.
Although this approach provides clinicians with useful information in making decisions about diagnosis, treatment, and prognosis, Zipkin stresses that EBM is a tool, not dogma. Factors such as socioeconomic status, pathophysiological differences, and individual wishes and values, determine much of clinical decision-making, and physicians must remain open-minded. "Knowing the best evidence doesn't necessarily mean you have to use it in a given way," Zipkin says. "You may have a statistically significant result that isn't clinically useful because it doesn't fit the context of the individual patient." For example, Zipkin recently evaluated a woman in her early thirties who complained of chest pain that had been increasing over the previous two weeks. The patient was healthy and had no history of cardiac or pulmonary disease, and Zipkin felt 98 percent certain that the problem was not related to the woman's heart. She ordered an EKG anyway. 'That's not evidence-based from the point of view of making a diagnosis," she explains. "I did it because I thought a normal result would make her feel better, and it did." The EKG was normal, and Zipkin noted that because the woman's discomfort ultimately turned out to be due to psychological stress, the reassurance provided by the test was actually therapeutic.
The next big step, Zipkin believes, would be tracking outcomes to determine how education in EBM affects residents' use of diagnostic tests and selection of various treatments. Such a program has yet to be designed for California Pacific, however, and to Zipkin's knowledge few programs have systematic EBM-related outcome measures in place. “The Holy Grail of any educational curriculum is its effect on skills and behavior," she says.
Many residents think the course is already proving its value without needing formal measures. "It's daunting to appraise all the information that's available," remarks Tim Laio, MD, a third year resident in the California Pacific program. "So using these evidence-based tools, we're better able to determine whether or not the studies are relevant to our practice. Not only do we use them ourselves, but we can also present that information to our patients for their response and let that help make our clinical decisions."
EBM has provided freedom for Lina Ilic, MD, a senior resident who, as part of her duties, is responsible for educating not only herself but other residents, interns, and medical students. "It's made me more independent in terms of forming questions, doing literature searches, and finding answers," she says. "On a daily basis I'll take that back to my team and we'll discover good information. We'll be admitting patients from the ER, and there will be something we're not clear about-medication interactions, side effects, issues like that-and we can look up those things together."
And that of course, is the point. Daniella Zipin believes that EBM, for the most part, is all about day-to-day patient care. "All this information has to be put into clinical context, and everything has to begin and end with patients," she says.