ROBERT SIEGEL, HOST:
And I'm Robert Siegel. This week were examining risk and how we handle statistics and probability. And today we're at the doctor's office. Many of us get confused by claims of how much the risk of say a heart attack might be reduced by taking medicine for it. And doctors can get confused, too. Karen Sepucha runs the Health Decision Sciences Center at Boston's Massachusetts General Hospital.
KAREN SEPUCHA: We did a survey of our primary care physicians a few years back and we asked them how confident are you in your ability to talk about numbers and probabilities of outcomes with your patients. And what we found surprised us a little bit. Only about 20 percent of the physicians said they were very comfortable using numbers and explaining probabilities to patients.
SIEGEL: Doctors, like Leigh Simmons, typically use words instead of numbers.
LEIGH SIMMONS: As doctors we tend to often use words like very small risk, very unlikely, very rare, very likely, high risk.
SIEGEL: Simmons is an internist. She's part of a group practice of internists who provide primary care at Mass General. Those words, she says, can be unclear to a patient.
SIMMONS: People may hear a small risk and what they hear is very different from what I've got in my mind. Or what's a very small risk to me, its a very big deal to you if it's happened to a family member.
SIEGEL: Doctor Simmons is part of a program called Shared Decision-Making. It gives patients online, written and visual information and one aim is to make risk understandable to bridge the gap between percent, probabilities and words.
SIMMONS: Most doctors would say if they know people have risk factors like diabetes and high blood pressure and they smoke our concern is extremely high. I'm often surprised when I put that into say a calculator and I see that the number is only in fact 25 percent. In my mind, I'm conceptualizing something even higher than that but for most of us that's the threshold that's quite high - one in four.
SIEGEL: One in four is a high number.
SIEGEL: You know, one can look at that and say, hey, the glass is three quarters empty. I mean, its, you know, the odds are in my favor.
SIMMONS: A lot of people will say that.
SIEGEL: So Doctor Simmons and others who use shared decision-making tools are trying to develop better ways to tell patients about risk.
SIMMONS: Have a seat.
JOE BIONCO: OK.
SIMMONS: So, Joe, I thought we might talk a little bit about your cholesterol.
SIEGEL: Joe Bionco (ph) of Charlestown, Massachusetts, is 60. And rather than just tell him how much at risk he is of a heart attack, rather than give him a number, Dr. Simmons makes it visual. She shows him a computer screen displaying a visual calculator developed by the Mayo Clinic. Imagine a hundred green dots arranged on a 10 x 10 grid, each dot with a little smile symbol on it.
SIMMONS: If we look at each of these dots as one person, it'll show us in the next 10 years how many of these people might be expected to have a heart attack. So we put in your age.
SIEGEL: Once Joe Bionco's (ph) information is entered some of those green dots will turn yellow and those smiles will turn upside down into frowns.
SIMMONS: And then looking back at the most recent cholesterol numbers your good cholesterol is 34 and as we've talked about I'd like to see that a little bit higher.
SIEGEL: When Bionco's profile is submitted and the screen represents 100 men like him, 12 dots turn yellow, 88 remain green. Twelve percent of men like him will have a heart attack. He's seeing a graphic representation of that probability. And those unhappy yellow dots are literally marginal. And little more than one vertical row of them running up the right side, greatly outnumbered by the still smiling green dots.
BIONCO: It looks like my chances are pretty slim. What's that? Ten, 12 percent.
SIMMONS: Twelve percent.
SIEGEL: Bionco had decided earlier not to take a statin medicine to lower the risk of a heart attack. And the 12 frowning yellow dots out of 100 don't change his mind.
BIONCO: Twelve percent isn't bad. If it was 25 percent - but I don't - that doesn't look like threatening news to me.
SIEGEL: Then Dr. Leigh Simmons enters another factor into the Mayo Clincs online computer. What if all 100 men fitting Joe Bionco's profile take a statin drug every day for the next 10 years? The number of yellow dots on the screen, the percent who will have a heart attack, changes for the better.
SIMMONS: That number goes down to seven. So five people are saved from a heart attack by taking a medication. Some people look at that and say, wow, that's almost cutting the risk in half. Other people say, well, I don't know, it's still 88 people who didn't benefit either way and only five who had a benefit. How does that ring to you?
BIONCO: If the facts support it then I'd take, you know, if it makes my decision valid by looking at the graph. Substantiates it.
SIEGEL: After Bionco left her office, Dr. Simmons told us how that graph has made her job easier.
SIMMONS: In his case he has had a long-standing concern about taking medications. And he has found it helpful to look at things more in that graphic form. So to engage him in that way has been pretty useful for me and not holding out an absolute - you must take a medication - but that I'm going to work with you to improve your overall health has been useful. So, I'm in agreement with him that holding off on a medication is a reasonable choice.
ANDY FREIBERG: About you and this evening?
JIM WESTBURG: Well, I've been having increased hip stiffness and pain.
SIEGEL: We're still a Mass General, but no longer primary care. We're now in the orthopedics department. Jim Westburg of Nashua New Hampshire has come to see surgeon Andy Freiberg. Dr. Freiberg is in charge of hips and knees.
FREIBERG: I just want to rotate your hips. Tell me if anything hurts or is weird. I won't do it hard. Does that hurt to rotate it in?
WESTBURG: A little, right there.
SIEGEL: Westburg is a very active 59. He hikes, swims, skis and rock climbs when he's not at work selling 3D printers. He wants a hip replacement and in the course of examining him, Andy Freiberg cites some numbers.
FREIBERG: Your risk of infection is probably under 1 percent. Probably close to half a percent.
SIEGEL: That less than 1 percent risk does not deter Jim Westburg. Before his appointment he had received a shared decision-making packet, booklet and a DVD all about hip replacement. He said that was helpful. He said he did some other research and just as important anecdotally, people he knew who had the surgery and were thrilled with the outcome. I asked how important it had been to hear numbers - percent probabilities attached, to say, the risk of infection.
WESTBURG: I think they're fairly important because I actually have an engineering degree. I've a technical background. So I - maybe I'm a little bias but numbers do mean something to me and the risk we talked about of a half to 1 percent doesn't really concern me that much.
SIEGEL: If he'd said 15-20 percent risk of infection?
WESTBURG: That would be a concern.
SIEGEL: That would be a serious concern.
SIEGEL: Can words suffice in that context? Is it OK to say the risk is really very low or it's very rare or do you need to hear a number attached to it?
WESTBURG: I like to hear a number attached to it. It do personally think the numbers are important. I mean, how somebody says, well, it's a moderate risk. Well, what does moderate mean? With numbers its much clearer.
SIEGEL: That's Jim Westburg of Nashua, New Hampshire. And so Audie, we've heard about communicating the probability of a successful operation or an illness by words, with images or using numbers. And Mr. Westburg says numbers are much clearer than words. Here is one measure of that that I came across that I want to try out on you, if you don't mind. It's about how common side effects are for certain drugs. The FDA urges drug companies to give numerical values for risk not, and I'm quoting from their website here, not vague terms such as rare or infrequent. But a few years ago the European Union matched a scale of terms for side effects - very common, common, uncommon, rare and very rare with numerical definitions for each of those five levels of frequencies. So what - here's the question to you - what percent of cases do you think qualified for the top level, a very common side effect?
AUDIE CORNISH, HOST:
Well, when you think about like ads for drugs and everything you always hear so many side effects. I don't know. So maybe I'll say 73 percent.
SIEGEL: Well, the E.U. defined a side effect as very common if it occurred in than more than 10 percent of cases.
CORNISH: Oh wow. That's all?
SIEGEL: That's all. And what if a drug label said, according to those E.U. definitions, that a particular side effect was very rare - the least frequent probability they defined - what percent of cases do you think?
CORNISH: Well, very rare - that's one I want to hear. You know, one percent or less.
SIEGEL: Actually in this case a very rare side effect, according to that European Union scale, is something that occurs in fewer than 100th of 1 percent of cases.
CORNISH: Oh, so I'm willing to take on more risk, it sounds like.
SIEGEL: I have asked several colleagues and friends over the past several weeks what these words suggest to them in the way of numbers and the answers are all over the map. Tomorrow in fact we'll hear more about numbers and prescription drugs.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.