Evidence-Based Medicine

It is morning in the clinic. I wash my hands and I put on a fresh white coat to start the day. Earlier, as I drank my coffee at home, I’d read several recent medical journal articles. Some concerned problems I needed to help patients solve; some were to keep me up to date on new developments. I’d added some to the 18,000-plus medical articles and textbook chapters that I keep cataloged in my laptop, and made notes of critical articles in a computer data base I built myself and use for reference in my treatment room. There are well over 20,000 medical abstracts and notes in that database.

Over 40 years of practice have brought me wonderful challenges. Some are memorable. One five-year-old had jammed a bean deep into his nose and, after hours of tweezers work by his parents, wasn’t about to hold still for my forceps. Solution? Let him cry, wait until he’d taken a deep breath, cover his mouth with one hand and push the uninvolved nostril closed with the other, and watch in satisfaction as he cried even harder on the exhale and blew the bean across the room.

More commonly, I work to help people decide how to respond to high blood pressure, cholesterol, arthritis, or other less dramatic problems. During these 40 years, I have always kept up. I’ve read a lot of studies and my memory and computer skills are good. How will I use all these studies as I meet the first patient of the day?

Buzzword

These days evidence-based medicine is a buzzword in medical practice. It means applying proven scientific knowledge – the results of medical studies – to clinical practice. Wikipedia has a long article on its benefits and problems at http://en.wikipedia.org/wiki/Evidence-based_medicine . It might seem an easy call to say yes – stick with the facts! Science rules! However, using evidence-based medicine is more complicated than first appears. It has value but it can also have adverse side effects. At its worst, it can be a lazy substitute for carefully interviewing the patient, obsessive examination, constant reading, and careful thought. At its best, evidence-based medicine saves lives and improves health.

The Value of Evidence-Based Medicine

First let’s talk about the pluses. Carefully-collected evidence can lead us to conclusions that intuition might make us miss. For example, death rate rises during influenza season, but many of these deaths won’t show up on autopsy as due to influenza. This mystified the experts until careful studies provided some answers.

In one hospital, almost all of the healthcare workers were immunized against influenza, whereas in a nearby hospital few were. The patient death rate in the hospital where few workers were immunized was much higher. Most of those deaths were attributed to non-influenza causes, such as infection, cancer, and heart disease. Deaths from those diagnoses were lower in the hospital with immunized health care workers. Why? Here’s what we believe. Unimmunized workers were bringing their subclinical influenza infection into the hospital with them. Influenza is most contagious in the days before symptoms appear. Such workers were unknowingly infecting critically ill patients who then died of their severe heart disease because they were simultaneously dealing with a subclinical case of influenza.

The Drawbacks of Evidence-Based Medicine

The Study May Not Apply

All medical studies are done on groups of people, chosen to be alike in some rational way. The women's health initiative famously indicated that estrogen can increase risk of heart attack and cancer. That estrogen did not provide an unqualified benefit was big news.

Here’s the catch. This study was done on women who were not having menopausal symptoms. So despite the early headlines that estrogen replacement was bad for all women, the study established only that it was not helpful for women who did not have menopausal symptoms. The study did not answer the question as to the risk-benefit equation in women with symptoms of menopause.

So while the diagnosis of menopause is easy, using this study to determine treatment could be tricky. Thoughtful physicians immediately recognized that this study just did not apply to their women who were not sleeping because they had hot flashes. Such women faced increased health risk because of poor sleep and ongoing distress, and the slight downside to estrogen in women who did not need it was probably outweighed by the benefit of good sleep and comfort in women who did.

Individuals Vs Study Group

Another error that overeager proponents of evidenced-based medicine sometimes make is to underestimate how different individuals can be. Studies are done on thousands of people. Of those thousands, some benefit while some experience harm. It can be tough to determine whether the person sitting in your examination room is one who will benefit or one of those who won’t. In many studies of nutrients, those who are deficient benefit, and those who aren’t, don’t. The doctor needs to ask herself whether her patient matches the average in the study group or whether they differ in some way.

Studies May Not Be Available

When antibiotics were found to cure meningitis, no randomized trial was conducted. Almost everyone died without antibiotics, and a gratifying number lived with them. We will not have randomized trials for every situation and health intervention. And when we treat a person whose illness is unusual, we cannot squeeze them into treatment paradigms that do not fit their situation.

So Many Studies, So Little Time

The evidence itself is impossible to keep up with. If one were to read all the articles published regarding general medicine, let alone basic research and the specialties, one would need to examine 19 articles a day each of the 365 days of the year.[1]

Balance Is All

So what do I do – this morning – as I greet the first patient of the day? John comes in because of his blood pressure. He’s 86 years old and been very athletic all his life. He’s has told me many times that he wants to avoid pharmaceuticals. Further, only his systolic blood pressure, the top number, runs high. I’ve not pushed treatment because older people are more likely to become dizzy and fall when medicated. Also, various studies of blood-pressure meds have not shown a compelling decrease in the death rate. I know that blood pressure varies up-and-down in anyone depending on time of day and other factors, but John has been coming in every month for about six months and most of the time his systolic blood pressure has been high and his diastolic pressure has been excellent.

Today I need to talk to him about it. Yesterday I reviewed recent studies indicating probable benefit for John from blood pressure medication. One of the most compelling studies[2] was done in China and Europe, populations different in some ways from ours. However, the almost 4000 elderly people on blood pressure meds suffered about a third fewer strokes and heart attacks than their untreated cohorts and more importantly, died at only 80 percent the rate. In these studies I also reviewed the response of elderly people to the various anti-hypertensives and decided that indapamide would be most appropriate for John. If he balks at the indapamide, I'll offer an alternative that may or may not work and is much less well studied. John will ask, as people often do, "do I have to do this forever?" I reply, as I always do, "no, just until we learn more or can find something better."

No algorithm can summarize all that is known. There are no cookbook answers. It’s only through careful examination and understanding of this unique individual – John who sits in my examining room – that together we can determine the best treatment for him.

Afterthoughts

“Evidence-based medicine” sometimes is used as a buzzword by people pushing medications or procedures of questionable value. I include routine screening colonoscopy, routine screening mammography, and many newer medications in this category. My views on these matters are, as of now, minority views. As the PSA controversy shows, mainstream views are changing and will continue to change[3].

Atul Gawande, one of my favorite physician writers, goes a step beyond evidence-based medicine to the real issue, which is process improvement. In other words, we may know the correct thing to do but are we actually doing it?

In medicine our challenge is to take the best practices and be sure to employ them in every patient encounter. I strongly urge you to read Gawande’s article. The top chain restaurants, believe it or not, are ahead of medicine in critical areas. Be sure not to miss the part about a central command center for intensive care units. This truly does promise to save lives and money, and the idea can be extended to emergency departments as well. Finally, if you have a question about where to have your knee surgery, Gawande’s article can help.

Endnotes

[1] (BMJ 312:71 1996). M

[2] Treatment of hypertension in patients 80 years of age or older. N Engl J Med 2008;358:1887-98.

[3] Evidence, Preferences, Recommendations — Finding the Right Balance in Patient Care. Timothy E. Quill, M.D., and Robert G. Holloway, M.D. The New England Journal of Medicine Volume 366;18 May 3, 2012 Page 1653.

Drs Quill and Holloway outline the importance of tailoring the evidence we have to the individual patient.