Angelina Jolie, meet Nate Silver

She may have made the right call for herself. But when weighing risks and benefits, channel your inner geek

Published May 15, 2013 4:59PM (EDT)

Angelina Jolie with her children, July 26, 2010.        (Reuters/Toru Hanai)
Angelina Jolie with her children, July 26, 2010. (Reuters/Toru Hanai)

Angelina Jolie made headlines yesterday with the announcement that she had elected to have a double mastectomy — not because she had breast cancer, but because she had the BRCA1 gene, a genetic mutation that, along with its sister gene, BRCA2, is responsible for 5 to 10 percent of all cases of breast cancer. There is much to applaud in Jolie’s decision — both to have the procedure and to be public about her thought process. But there’s also cause for alarm: Her message is likely to get outsize publicity. In an area as rife with uncertainty and misunderstanding as cancer risk, is her vocal editorial likely to help, by raising awareness of cancer risk and treatment choices – or could it have repercussions that, in the long term, hurt more than they help?

When it comes to assessing personal risk, two things are true: We place an outsize impact on emotion, whether or not we’re aware that we’re doing so, and we fail to look at objective probabilities in an objective light. In something known as the affect heuristic, we let the way we feel about something influence both our perception of risk and the decision we make as a result. If we judge something in a positive light, we tend to rate its risks as low and its benefits as high. The reverse is true if we see it negatively. We then rely on such emotions to determine how and what we’ll decide to do.

Take, for instance, a medical treatment, such as a double mastectomy. If we see such treatment in a positive light — say, when a celebrity we like endorses it — we are more likely to think that its potential benefits outweigh its potential risks. If, on the other hand, we see it in a negative light — we know a teenage girl who committed suicide after a preventive double mastectomy rather than live with its side effects — the opposite becomes the case. Suddenly, the risks seem far weightier than the benefits.

It doesn’t matter what the actual risks and benefits are, as presented by clinical research or our physician. It doesn’t matter what the statistics say (in the case of a double mastectomy, the results are actually mixed: While it does tend to reduce the risk of breast cancer recurrence, in people who have already had cancer, there are mixed findings on whether it helps in those who have never been sick, or if it does, how strong the effect is and how high the remaining likelihood of breast cancer). So strong is the attending emotion that it is incredibly difficult to make an objective judgment; one probability simply feels different from the other, even if the number stays the same.

Consider the less emotionally evocative example of a lottery. If people have the chance of winning $1 by picking a red jellybean out of a bowl, studies show they will in the majority of cases choose to make their selection from a bowl that has the highest number of red jellybeans, not the bowl that presents the higher probability of winning. Even when the odds are twice as good for the larger bowl, 34 percent will still choose to pick their jellybean from the smaller one. Here’s the kicker: They do so even though they know, logically, that the odds are better in the other bowl. They just feel that the odds are better the other way around. They focus on the positive (the number of possible winning beans) and ignore — or rather, choose to ignore — the rest of the information.

It’s not just an ignorance of relative chances of success. The greater the uncertainty of the outcome, the more we tend to exhibit an all-or-none mentality: We think in absolutes instead of relative probabilities. If the emotion that accompanies a possible outcome, be it positive (winning the lottery) or negative (getting breast cancer) is strong enough, we become remarkably insensitive to variations in actual risk. We think, we’re reducing our risk of cancer to zero, instead of the more accurate, we’re lowering our risk by up to 90 percent — but not only is there still a chance I’ll get breast cancer, I have also not in any way altered my chance of getting some other cancer — or dying from some unrelated cause.

In that thought process, several things happen. We experience probability neglect — we focus on the deeply negative (or positive) emotional impact of the outcome, not the probability that it will actually occur — and we ignore context, or base rates, of different possible outcomes to begin with. Would we decide differently, weigh the pros and cons differently, in the two scenarios? Quite likely; but once we’re in a certain mind-set, it’s remarkably difficult to change that.

When a doctor tells us our risk of some cancer or the result of some genetic test, we are never just going to take that in as an objective probability distribution. How he communicates it, whether he puts it in context, whether he frames it in positive or negative terms, even how we feel about the doctor himself: All of these issues will factor into our perception of our own risk – and our thought process about how we should act as a result. Even the label “positive” and “negative” for a test matters. Tell someone they’ve tested positive (or not) for the BRCA1/2 gene and you get an effect that’s quite different from telling someone that you know that they have a certain chance of getting breast or ovarian cancer, minus the positive or negative label. Getting a “positive” result tends to increase perception and worry relative to getting just the risk information (i.e., what a positive result actually means). Give us a positive, and we become more likely to try to act — which is not always the best option, statistically speaking.

Of course, in Jolie’s case, her objective probability of breast and ovarian are quite high: 87 percent and 50 percent, respectively. She likely made the best decision for her, given her risk and the information available to her. But here’s the thing: Her choice is a remarkably personal one, and she may not realize just how much of an impact it will have on others, for whom it may not be nearly as optimal.

In her Op-Ed, Jolie talks only about her risk, her family history and her resulting decision. But she fails to mention a number of other details — details that should affect not only the decision to get a double mastectomy but to get tested for the BRCA1/2 gene to begin with.

Let’s start with the practicality of the test itself. Jolie omits one crucial detail, which we know intuitively but may not apply in this particular instance: She is remarkably wealthy. For her, insurance or premiums are not a factor. But what about people for whom money matters (and it just so happens that many of the highest-risk populations are also less financially secure)? True, she mentions that the cost of the test itself is prohibitive for many. But what of the consequences? If you make the choice to be screened, your results become part of your medical record. If you happen to be positive, one of two things will likely happen: Either your insurance premiums go up, or insurance companies will deny you coverage to begin with. You simply become too high a liability. (While, according to the Genetic Information Nondiscrimination Act (GINA) health insurance providers are not allowed to discriminate based on genetic data, the prohibition does not extend to other types of insurances, such as life, disability, or long-term-care.)

You then have to decide whether for you, given your background, your history and your situation, screening makes sense. Sure, you may think it’s good to know. But what if knowing — and doing something about it — will only increase your long-term survival by 3 to 5 percent? (That number isn’t random; it’s based on a 2007 estimate of BRCA1 and 2 penetrance, and how survival improves with mastectomy at various points in time.) Is it still worth it?

Absent, too, is any discussion of the other cancers that the gene mutation predicts. Breast and ovarian are not the whole list. Assume you get tested for the mutation. Assume you test positive. What does your mastectomy actually accomplish? Yes, it will certainly reduce your risk of breast cancer (though never to zero). But what about ovarian cancer? You need to get an oophorectomy to make a difference there. And that’s not to mention peritoneal cancer, the risk of which also increases dramatically with the mutation. How far will you go to protect yourself?

Moreover, once you do something, how likely are you to remain vigilant instead of becoming complacent? (I’ve eliminated my chance of cancer!) Even if you know, theoretically, that the mastectomy bore no relevance on your other cancer risk — and that it didn’t even bring your breast cancer risk to zero — will you listen?

That’s not to mention factors such as age -- Jolie is incredibly lucky that she waited until 37 and remained cancer-free; the most effective survival tactic with the BRCA mutation is a double mastectomy in your teens, since the gene is mostly associated with early-onset cancers — and other lifestyle choices. Are you a smoker? Do you participate in extreme sports? Do you use your seatbelt when you drive? In other words, where does this risk fit into your overall risk context? High, or low? We tend to evaluate risks as stand-alone things, when in reality, they are always part of a broader picture – and that picture may dictate a different decision from the isolated data point.

Then, there’s the highly counterintuitive aspect of the choice to get screened in the first place. As it turns out, knowing your cancer risk may not actually translate into better decisions. Accurate knowledge of risk can, in turn, make you less vigilant — especially if the news is, relatively speaking, good. Consider the numbers that Jolie mentions: 87 percent risk for breast cancer and 50 percent risk for ovarian. For someone who is largely uninformed about cancer risks (unfortunately, quite a large percentage of the population), or even for someone who is informed but may not remember the precise numbers, Jolie’s personal risk factor becomes the salient baseline of the moment. That risk is high. Almost anything is likely to look good in comparison, with potentially less than good consequences. In one study, patients who were first asked to estimate their own risk of breast cancer and then told that the overall risk in the population was 13 percent experienced relief at that second number: Their own estimates had been much higher (on average, around 46 percent). And women who were just given the 13 percent risk factor without making a prior estimate? Relief was the last thing they felt. They felt that the risk was a substantial one. That initial anchor, 13 percent or 46 percent, made the subsequent “real” risk look quite different. So what if you’re starting with 87 percent? Knowing the facts doesn’t mean you are able to evaluate them effectively.

Even knowing comparative risk levels changes how we see or evaluate each risk separately. When women were told that they had a 6 percent chance of developing breast cancer but could take a pill (with extensive negative side effects, such as endometrial cancer and stroke) to reduce it to 3 percent, their choice depended on whether they thought the general prevalence in the popular was 3 percent or 12 percent. If they thought it was 12 percent, they were not likely to take the pill; in the 3 percent scenario, interest in the pill increased. Though the objective risk remained the same, the action changed drastically. Logically, each woman should have decided, for her, what the 3 percent reduction would mean. Were the tradeoffs worth it? Someone else’s risk shouldn’t have played any role at all. But that’s not how we think — or how we decide. A statistic is never just a statistic.

Angelina Jolie has made a brave choice, no doubt about it. But does she realize just how much she is coloring the conversation? When we see individual cases instead of statistics, our judgment changes. In one study, clinicians were given information about a hypothetical patient in one of two ways: They were either told the probability that he would act violently (“patients similar to Mr. Jones are estimated to have a 20 percent chance of committing an act of violence”), or, they were given the information as a relative frequency (“20 out of every 100 patients similar to Mr. Jones are estimated to commit an act of violence”). Strictly speaking, the two are absolutely identical. And yet, while in the latter case, only 21 percent refused to discharge him, in the former, fully 41 percent did so. Why? It is far more emotionally powerful to imagine an individual than a statistic — and that power changes not only our perception of risk, but the decision we choose to make as a result. One can only imagine that when the individual in question is Angelina Jolie, the effect is magnified.

Writing for the New Yorker, Rebecca Mead calls Jolie’s choice “bold and brave and admirable. It is what celebrity is for.” I would agree if she were describing the choice of the surgery. In the choice of going public, I’m not so sure. True, it does bring the discourse away from the sex appeal of breasts and to the empowerment of women. But how empowering is it, really, when it comes to the choice itself? It’s hard enough to assess risk without added bias. With the Angelina wow factor in the mix, it becomes well nigh impossible.

Celebrity also comes with responsibility. Does Jolie want to inform women of breast cancer risks and treatment options? Excellent. She can present the issues without making herself — and her choice — the center of the presentation. Jolie focuses on the details of her surgery, true. But she doesn’t focus on the assessment of risk that preceded it, other than mentioning her mother and her genetics. She doesn’t talk about the imprecision of the data or the lack of medical consensus. She doesn’t mention that the most effective and persuasive mastectomy results are in young women—and that she’s past the target age and has been lucky to make it there to begin with. She doesn’t mention the other cancer risks, the trade-offs of screening, or the need for further drastic surgery. She doesn’t mention the risks that remain, that her choice, while the best for her, is necessarily an imperfect one.

I realize her Op-Ed is short and her space, limited. I applaud her for writing it and understand that she can’t be expected to speak with limitless nuance, nor can she expect that a less personal account, should she have chosen to go in that direction, would garner anywhere near the same level of news coverage. And yet, she should know what impact she, personally, has, and that when she presents information selectively, people will read it in absolute terms. Her choice is her choice, and it is indeed brave and admirable. But how many will see it and think, I need to do that, too? I’m not doing enough? Just starting the conversation is an important step — but that conversation needs to be put into strict context.

At the end, medical decisions are intensely personal. Do I screen? Do I prevent? There are myriad ways of being in control, and they are as far from one-size-fits-all as they come. When Angelina sets a hemline trend, it’s one thing. If she sets a mastectomy one, that’s an entirely different ballgame.

By Maria Konnikova

MORE FROM Maria Konnikova

Related Topics ------------------------------------------

Angelina Jolie Breast Cancer Double Mastectomy New York Times Science Statistics