The Importance of Reading the Evidence
Created August 29, 2012 by Michael Hochman, MD
When I first began working with patients as a third year medical student, I was struck by the variation in how doctors delivered care. One example sticks in my mind: I was evaluating a patient with new-onset depression, and the doctor I was working with started the patient on an antidepressant medication. Later, I was describing the case in our teaching rounds, and a different doctor adamantly disagreed: “In patients like this, I always try psychotherapy first,” he said. I soon realized that depression care is far from standardized. While some doctors favor starting medications immediately in all patients, others typically try psychotherapy first, while a few start both treatments simultaneously.
Which strategy is right? When I finally got around to researching the question a few years later, I discovered that several studies have directly compared the different strategies. These studies clearly show that, in most patients, medications and psychotherapy are equally effective (though recovery may be slightly faster with medications).
Reading these studies made me feel more informed and confident the next time I encountered a patient with depression. But the experience was also unsettling because depression is far from the only disease that is treated differently by different doctors. For example, some doctors commonly recommend surgery for patients with back pain while others rarely do. Some doctors have a lower threshold for starting medications in patients with HIV than others. And some doctors target very low blood sugar levels in patients with diabetes while others use less aggressive targets. How could I possibly have time to review the data on all of these important topics to make informed decisions for my patients?
Though the task seemed daunting, I realized it was important. My experience reading the depression studies had enabled me to treat my patients more effectively, and I suspected the same would be true for many other diseases. So I decided to make the best effort I could to read the key original research on a handful of important topics, recognizing that I never would have time to read everything I would want to.
To my surprise, this effort proved easier and more rewarding than I could have imagined. Though my knowledge of the medical literature remains far from optimal – and many of my colleagues are much better versed than me – I have realized that knowing even a few key studies goes a long way. It turns out that in many areas of medicine, one or two well-designed studies have defined clinical practice. A basic understanding of these key studies can greatly enhance decision-making.
Reading these key studies has also revealed some themes that have enhanced my perspective on patient care. Most notably, I have learned that medical treatments and technologies are often not as effective as many doctors and patients believe them to be. For example, statins – the supposed wonder-drug of 21st century medicine – actually only reduce cardiovascular events slightly. According to one of the best studies on this topic, the JUPITER trial, approximately 200 patients without cardiovascular disease would need to be treated for two years to prevent one death. In addition, the study raised the possibility that statins increase the risk of diabetes. And an analysis summarizing the data on mammography showed that for every 2,000 women offered screening for 10 years, only one will be saved from breast cancer while 10 will be diagnosed with cancer and treated potentially unnecessarily.
In addition, I have learned that many treatments that have proven to be effective were studied in circumstances that may not reflect real-world conditions, and thus the results may not be broadly applicable. For example, a famous trial establishing that medications are superior to psychotherapy for children with Attention-Deficit/Hyperactivity Disorder was conducted using an intensive treatment protocol in which doctors met frequently with the study children and received regular feedback from parents and teachers. Such a strategy may not be realistic in many settings.
The lessons I learned from reading just a few key studies had an important impact on my thinking. I now pay more attention to patient preference and less to medical dogma. For example, when I encounter a patient who prefers not to undergo mammography, I respect the decision because it is very unlikely she will be helped by the test anyway – and might possibly be harmed by the unnecessary discovery of a cancer that otherwise may not have caused problems.
More importantly, making an effort to read original medical research has shown me that when I encounter a clinical question, it is worth the time to review the literature for an answer. Frequently, a good study won’t be available to address the question, but at least by looking it is possible to provide patients with the most evidence-based information. Ultimately, I believe, this will lead to better care than decisions based largely on experience.
Dr. Hochman is a general internist. He recently published 50 Studies Every Doctor Should Know (www.50studies.com)