Framing risks, benefits perilous for physicians and patients

Despite many scientific advances, much of medicine still exists within a gray zone. There is no black or white answer about when to initiate treatment or how to treat. Rather, there are a number of different approaches to treatment, each with its own risks and benefits. For any individual patient, the best choice may be anything but simple or clear.

Our new column, Gray Matters, will explore this dimension of medicine where physicians advise and patients choose among different treatment options in the face of uncertainty. The matching of a treatment to the individual patient is what we view as the “art” of clinical judgment.

As physicians, the first step in this process is to inform our patients about the various therapeutic choices, specifically the pros and cons of each option. Indeed, the Institute of Medicine of the National Academies places informed patient choice at the pinnacle of quality medical care. This conclusion rests on the fact that it is the patient who either enjoys the benefits of a treatment or suffers its risks. But what does it mean to be truly informed?

As part of the research for our recent book, “Your Medical Mind: How to Decide What Is Right for You,” we interviewed scores of patients making medical decisions. We found that information given to patients came from a wide variety of sources, not only from their physicians, but also drug advertising, Internet sites, and relatives and friends.

One of the first patients we spoke to was a 45-year-old female nurse's assistant in Boston. At a checkup with her primary care doctor, her blood test showed an elevated cholesterol level of 240 mg/dL. She did not smoke, was not diabetic or hypertensive, and had no family history of heart disease. She followed a healthy diet and was physically active. Her physician gave her a prescription for a statin medication and scheduled a return visit in one month.

Statins are among the most commonly prescribed drugs in the world; in the United States alone, more than 25 million people take them to reduce their cholesterol levels. But this patient decided not to immediately fill the prescription for the medication. She is hardly atypical in this regard. Surveys reveal that one-quarter to one-half of patients do not fill their prescriptions or stop taking medication on their own.

At the patient's follow-up appointment a month later, her primary care physician, hearing that she had not started treatment yet, told her, “By taking a statin pill, you will reduce your risk of a heart attack over the next 10 years by as much as 30%.” At the same time, the risk of side effects, the doctor said, was quite small. The benefits far outweighed the risks. The patient replied that she would give it serious thought.

Over the ensuing months, the patient sought information about elevated cholesterol and its treatment. While searching the Internet, she noted a government-sponsored link to a “10-year heart attack risk calculator.” The patient entered her age, total cholesterol number of 240 mg/dL, and HDL of 37 md/dL. She also entered that she was not a smoker, did not have hypertension, and was on no other medications. The result: “Risk score: 1%; means 1 of 100 people with this level of risk would have a heart attack in the next 10 years.”

The risk calculator had given this patient a key bit of information. Without any therapy, her risk for a heart attack was one in 100. If one in 100 women has a heart attack in the ensuing decade, that means that three in 300 do. The statin treatment, by reducing the risk by 30% or one-third, will protect one out of the 300 women. The other two women will still have a heart attack, despite taking the drug, and the remaining 297 would not have had a heart attack even without the medication, so they wouldn't have benefitted from it.

This calculation comes as a surprise to many people, even to us as physicians. When we hear that a statin lowers risks by 30%, it sounds as if the patient is at 100% risk of suffering a heart attack if she doesn't take the drug. But once we know the baseline risk, we can calculate the “number needed to treat,” in this case, one woman in 300. If the current patient's blood cholesterol rose to 280 mg/dL as she entered her fifties, her baseline risk for a heart attack would be two out of 100, and the number needed to treat would be one in 150, meaning one woman would be protected by the statin from a heart attack and the remaining 149 would not benefit.

Any potential benefit of a treatment, of course, must be assessed with regard to its risks. There are varying data about how frequently statins cause muscle pain, ranging from 1% to 10%, depending on the dose, the specific statin prescribed, and other variables. This risk may sound quite large, but if we “flip the frame,” and present the risk as the number without any side effects, it is 90 to 99 out of 100, a much more reassuring statistic.

Many patients also are informed by drug advertisements. Such ads frequently include statistics. After she saw her doctor, this patient paid particular attention to ads for statins. They were not hard to find, since they are a staple in magazines and on television.

A 2007 study conducted by researchers at UCLA Medical Center and other hospitals tallied prescription drug ads broadcast on the evening news and prime-time TV shows. The researchers found that a typical American TV viewer sees more than 1,000 prescription drug ads in the course of a year, amounting to some 16 hours. While advocates of direct patient advertising argue that ads fulfill an educational purpose, frequently these ads lack complete information or are presented in a way that makes it hard for patients to clearly understand the actual benefit of the medication.

For example, in a study from the Dartmouth Institute for Health Policy and Clinical Practice, one group was given actual drug ads, while a second group received the same ads with the brief summary at the end of the text replaced by a “drug-facts box.” Within the box, information on risk and benefit was presented in a clear and understandable fashion, similar to the way we performed the calculation on statin therapy for the current patient.

The Dartmouth researchers noted that nearly two-thirds of the group that saw the original ads overestimated the benefits of the treatment, believing it was 10 times more effective than it actually was. On the other hand, nearly three-quarters of the patients who saw the information in the drug-facts box were able to correctly assess the actual benefits of the therapy.

As physicians, we encounter statistics about relative risk not only in advertisements but also in media reports, in abstracts at medical meetings, and in published articles. In a recent study from Switzerland featured in the Aug. 9, 2011 ACP InternistWeekly, researchers mailed more than 1,000 randomized surveys to every doctor in Geneva and to more than 1,000 recently hospitalized patients. Patients and doctors were asked to interpret the results of a hypothetical clinical trial comparing an old drug to a new one. They were randomly assigned to framing formats of absolute survival (new drug 96% vs. old drug 94%), absolute mortality (4% vs. 6%), relative mortality reduction (reduction by a third) or all three.

When assessing the efficacy of the new therapy, doctors proved to be just as susceptible as patients to the misleading effects of framing information as a relative change. Presenting data as a relative change is often done deliberately to yield maximum impact and can lead us astray in our thinking. So, we as physicians need to be alert to the tricks the mind can play in assessing information when it is framed in different ways. We best inform our patients by giving them understandable numbers such as their absolute risk for an outcome and the number needed to treat. This provides an accessible and accurate way of beginning the process of informed patient choice.

In subsequent columns, we will feature cases submitted by readers, exploring other aspects of patient decision making. We look forward to receiving cases where you have interacted with patients over a difficult decision regarding treatment.