The Paradox of Disclosure

From Sunita Sah (NYT):

Disclosure can also cause perverse effects even when biases are unavoidable. For example, surgeons are more likely to recommend surgery than non-surgeons. Radiation-oncologists recommend radiation more than other physicians. This is known as specialty bias. Perhaps in an attempt to be transparent, some doctors spontaneously disclose their specialty bias. That is, surgeons may inform their patients that as surgeons, they are biased toward recommending surgery.

My latest research, published last month in the Proceedings of the National Academy of Sciences, reveals that patients with localized prostate cancer (a condition that has multiple effective treatment options) who heard their surgeon disclose his or her specialty bias were nearly three times more likely to have surgery than those patients who did not hear their surgeon reveal such a bias. Rather than discounting the surgeon’s recommendation, patients reported increased trust in physicians who disclosed their specialty bias.

Remarkably, I found that surgeons who disclosed their bias also behaved differently. They were more biased, not less. These surgeons gave stronger recommendations to have surgery, perhaps in an attempt to overcome any potential discounting they feared their patient would make on the recommendation as a result of the disclosure.

Surgeons also gave stronger recommendations to have surgery if they discussed the opportunity for the patient to meet with a radiation oncologist. This aligns with my previous research from randomized experiments, which showed that primary advisers gave more biased advice and felt it was more ethical to do so when they knew that their advisee might seek a second opinion.

The piece is…self-recommending!


When I was diagnosed with non-Hodgkins lymphatic cancer 20 years ago, I hired a general oncologist to be my adviser in figuring out which of Chicagoland's three lymphoma specialists to go with. He helped me choose the expert who had a clinical trial opening up for a breakthrough drug called rituximab, which went on to rack up 8 figures in revenue in the 21st century.

And here I am.

I don't know why more patients don't hire a generalist doctor to help them pick the right specialist. It can be a life or death decision, and you probably don't know anything at all about the topic you are suddenly forced to learn about ... while you are sick.

Do doctors often consider it unethical to be paid to evaluate other doctors? Is that considered a violation of Professional Ethics?

Wow, that's a nice feel-good story. I heard the author of Cancer Ward (, Alexsandr Solzhenitsyn, had cancer but lived to the ripe old age of 89.

In Canada, you can't get a specialist unless the generalist sends you their way. It adds a little time, but presumably gets you to the right people and prevents people from using up a lot of specialist resources in an unnecessary manner.

At the extreme, probably a lot of hypochondriacs use of a lot of specialist expertise in the USA (well, since they pay for it, who's to complain, but isn't that a waste of a scarce resource?), whereas in Canada you wouldn't get past the gatekeeper (the generalist) unless initial tests suggested the specialist was needed. However, this introduces a problem where, perhaps, the patient really really needs a specialist but can't find a generalist with sufficiently specialized training/experience to understand that this is the case.

That's the way HMO's and PPOs work in this country. Even 25 years ago most people did not have indemnity plans which allowed you to see a specialist without a referral (head shrinkers, dentists, audiologists, and optometrists excepted). I think they've just about disappeared. IIRC, the woman who blogs as 'neo-neocon' had one as recently as about 10 years ago.

>> well, since they pay for it, who’s to complain, but isn’t that a waste of a scarce resource

They pay for it. Neither employer based nor private health insurance companies are allowed to underwrite and charge more to these loons. And certainly medicare and medicaid don't.

Doing just that is part of what we primary care docs do. Over time we get to know our local specialists and make a personal/professional determination as to their scope of abilities. And then make recommendations to our patients based on that. We get paid based on the nature, time and complexity of an office visit, not specifically paid simply to evaluate another doc. Although rendering a 2nd opinion might satisfy much of the requirements of an office visit and sensible reimbursement.

Similar to your story my wife had the same experience in 2004. Long story short, our local oncologist I did not recommend. I sent her to his program director at the University where she got the same miracle rituxin. His program director, who also just happened to be my chief resident in training 30 years before, and never recommended our local guy in the first place!

Here in the UK we have State-controlled healthcare, you either do what they say or you die

Much like K-12 education in the U.S. If you don't like your taxpayer-funded government school, you can pay to send your children to a non-government school if you can afford it (e.g., virtually every major Democratic politician). Otherwise, you're stuck with your crappy government school.

Out on the West coast , we do have some great Public schools. Those passing out from these do compete well with those coming out of the most expensive private schools. Where are you based?

In US News ranking, public and private schools alternate in the "top ten high schools in the US," which is probably the way it should be.

But then why would you care?

The Top 10 golfers make a lot of money. Should I quit my job and play golf?

We are branching off the preposterous position that "government schools" are "crappy." If they are in the top ten, that certainly documents that "some" are not.

In terms of what you should do. Mindful meditation might help.

The US News ranking includes far more public schools than private schools, and private schools are over-represented in the top rankings. They have a higher rate of being ranked "gold" than public schools.

Plus, "expensive private schools" are not part of the rankings--the only private schools included are ones like charter schools that receive public funding.

In addition, a good portion of the ranking is based on the school's relative performance within its state. In some states there are few or no charter schools, so there isn't any competition with private schools that goes into that portion of the ranking.

Overall, this data doesn't disprove the notion that government schools perform poorly.

So now "crap" just means not all in the top 10.

Do you think you meaningfully added to this discussion?

You can read raw differences and adjusted for students results here:

@anon I think the serious debate is about the overall quality of public schools, not whether literally every single one can justly be called "crappy". Clearly the latter is not true. At least some public schools are good. But that is a trivial point that doesn't disprove the broader claim that government-run schools perform worse than private schools.

except, of course, if you have private health insurance, or elect to pay out of your own pocket, or simply ask your NHS GP to find you a different specialist for a second opinion, or even discuss with your GP which specialist to go to first. The latter is routine at my GP's practice.

Sounds a little hysterical.

While they are not required to do what you say, neither are you required to do what they say (psychiatric incarceration excepted, and I'm aware of at least a few cases of some pretty severe violations of rights under such pretenses for the specific case of the UK). Presumably most of these decisions are made on the basis of diagnostic criteria which continue to be debated for matters of best practice.

Or you can fly to New York and see any doctor you want.

When you are carrying a hammer, everything looks like a nail.

My general comment is that I'm a troll, but I also try and be informative as well as entertaining at times.

It's so subtle that it might actually work!

That was so incredibly self-insightful that I wonder if it was actually someone else posting under your name.

No, the clinical cases aren't fixated on him.

On the other hand, I make no attempt at either.

It is hard to talk about bias without knowing what the "correct" value is. For which patients is surgery actually likely to pass a cost-benefit test, and how often do surgeons and non-surgeons get this right?

We have two groups of people, surgeons and non-surgeons, with different rates of recommendation, and the assumption seems to be that the surgeons are the "biased" group. On what basis is this assumption made? This implies that the surgeons are recommending too many procedures; is there evidence for this? I wasn't able to find any in a brief search.

Excellent points.

Just wait until Robin Hanson finds out about this!

What you really need to know is what some relevant sample of doctors recommend for their families or themselves. Many doctors decline the offer of the sorts of end-of-life treatments that keep you hanging on painfully and pathetic, decline several sorts of prostate cancer treatment, and decline routine mammography.

I thought I read that this is a myth and doctors actually choose just as aggressive end of life treatment (although they say they won't before it happens)

There is some evidence, but it's not that strong from what I can see:

The first study suggests doctors are far more likely to have a care plan for end of life, and that makes a difference in their treatment.

However, the second study only showed small differences between doctors and the general public in measures like surgery in the last 6 months of life, or rate of death in hospital. And there was very little difference between doctors and lawyers (a non-health professional group that is fairly similar to doctors). This study is spun as supporting the idea, but it is not really that strong of support for it.

Maybe it differs by country: the end-of-life stuff I've seen stated quite firmly for the UK. For the mammography point see Gegerenzer's book's comments on German cancer specialists:

That may be where I learned about the prostate stuff too.

I have a first degree relation who was saved by a timely mammogram.

A leading mammographer where I grew up was asked to consult on a Canadian study on the efficacy of mammography. She later said that when she was done with those conducting it she'd concluded that the study would demonstrate that mammography was ineffective and would do so because the mammograms taken by this crew were so low quality they could not be reliably read. Sho' nuff, they concluded mammography was ineffective. It got a big splash in the papers.

I'm fascinated to know just who these doctors are. Among my first and 2d degree relations and good friends, I can name 2 who were told treating them was a waste and they needed palliative care (one within days of his diagnosis, one withing 2 months); 1 who elected to forego chemotherapy because his wife and her family could not handle nursing him; 1 who was taken off a respirator because general surgeons refused to perform a tracheotomy on her, one who bounced back and forth between various levels of intensive care and rehabilitative nursing home care and then was zonked out and had her feeding tubes removed; and one that was sedated and allowed to expire. One of these individuals was 89 years old and just about ready to go. The others might have had some years yet (though two were dreadfully ill). I did not hold the health care proxy on any of them. The granular details of three of these cases made the decisions made by medical staff (to which the proxy holders were buffaloed into accepting) rather dubious. I've seen some wretched things among older relatives. You do get the impression that doctors treating old people with chronic ailments over-prescribe and that some of their problems are iatrogenic. Still, they were miserable because they were demented and ill, not because they were being treated.

Of course, specialty bias isn't limited to medicine, as it affects everyone from lawyers to economists. I'm a lawyer, and I work with specialists ranging from tax to securities to health care to bankruptcy. What's clear is that each specialist views every project from her specialist perspective: for the tax lawyer the tax consequences (i.e., tax avoidance), for the securities lawyer securities law regulatory compliance, for the health care lawyer health care law regulatory compliance, for the bankruptcy lawyer problems of collection or opportunities for collection avoidance. In economics, which is experiencing a surge in specialization, from micro and macro to development economics to econometrics to health care economics to labor economics to behavioral economics. Unfortunately, something is lost with specialization, which is the ability to see the big picture. In medicine, the patient's primary care physician is supposed to see the big picture, taking into account the details provided by the specialists, but that is the ideal not the practice. In law, clients often seek the advice of the generalist with broad-based experience and a reputation for good judgment, who takes into account the advice of the specialists but makes recommendations based on her experience and judgment. In economics, well, I'm not one, but my observation is that the field needs more generalists with broad-based experience who can synthesize the data and make recommendations based on experience and judgment. The burden of the economist is that the field has become enamored of data and math and specialization. Of course, economics is primarily an academic discipline, where career advancement often depends on data and math and specialization, where the big picture is mostly irrelevant, where being wrong has few if any consequences other than embarrassment (unlike in medicine and law where being wrong can be catastrophic). Yesterday I listened to a podcast of David Beckworth interviewing Mark Thoma. What struck me is that Thoma has become a generalist as the result of his blogging (to be the blogger that he is, Thoma must read an enormous volume of material with a broad scope), and that Beckworth understood and appreciated it. Blogging is the path for the economist to see the big picture! Anyway, that was my revelation. Is that why Cowen is devoted to blogging? Does he see the big picture?

Good comment.

also Steve Sailor's @ 1:34 am above

This has been my experience with the Overcoming Bias/Less Wrong community as well. My interpretation is that people who disclose bias are more confident of their objectivity.

It happened to America. When the Media convinced us that neutrality was impossible, got the Fairness Doctrine repealed, they didn't keep trying, knowing their flaws. They ran with the flaws.

On the other hand, I think the story pre-repeal shows that if you know bias is a risk, but strive for the impossible ideal, you can still do better.

These doctors probably need a mental checklist, some tricks, and a target of neutrality.

When the Media convinced us that neutrality was impossible, got the Fairness Doctrine repealed, they didn’t keep trying, knowing their flaws. They ran with the flaws

The media did no such thing. Mark Fowler of the FCC engineered the end of the Fairness Doctrine, with the support of Nat Hentoff (among others), who had been working in radio broadcasting when the Doctrine was imposed. Democratic politicians (e.g. Louise Slaughter) have been scheming since to re-impose it, because free broadcasting means opposition for the Democratic Party, opposition they don't want.

The link above applies here too

Ironic that the Guardian asks "Does the truth matter any more?"

They seem to think they have the copyright on making up their own facts.

We all have feet of clay, but aspiration is important.

Or is it that they feel comfortable that they've abdicated responsibility for objectivity, so with disclosure they feel comfortable pushing their own (disclosed) interest.

"specialty bias" In my day we called this "professional deformation".

How many are ready to believe that doctors are biased, but markets are efficient?

I believe there is some related controversy on "financial advisors" and what we should expect of them.

"How many are ready to believe that doctors are biased, but markets are efficient?"

Saying markets are efficient, doesn't imply they are 100% correct. Just that they are the best method of allocation of resources that mankind has devised. Just like doctors aren't 100% correct, just the best practitioners of medicine mankind has devised.

A fuller history would include that the EMH did include "the price is right" until a few short years ago. Many still defend thus "hard" form in these pages, explicitly or implicitly.

The Shiller Nobel should have closed the door on that, but few are ready for his logical follow on discussion of animal spirits in an imperfect world.

Perhaps even pretending that we live in a society where everything is allocated in a market is an EMH hangover. A just-so story.

Posts that are self-recommending, thereby disclosing their bias, are more self-recommending.

Only if they themselves disclosed that it was self-recommending. If we only heard it was self-recommending from someone else, it should increase our suspicion.

Self-recommending post here.

Wait, that's a second hand self-recommendation. Now I'm a tad more suspicious.

Full disclosure: I am biased against your comment. But that made me even more likely to post another dumb comment in this thread.

This post is also self-recommending.

Tangentially related: when I was in law school, I remember a couple papers being presented by what I will call "disclosure libertarian paternalists," or possibly just "disclosure libertarians."

Basically, the idea would be "disclose it all (or a lot of it) and let The People sort it out." This was, if I recall correctly, typically offered as an alternative to a more command-and-control type regulation requiring certain behavior.

My question then was, and still is: if we can reliably predict certain behavioral results from disclosure (as the research described in this post demonstrates), why is requiring that disclosure better than simply requiring that behavior? And I mean "better" in a strictly moral sense: I think most people of a libertarian persuasion would think that requiring disclosure is less of an imposition on human freedom than directly requiring the behavior, and hence better.

But if the predictable end result in both cases is exactly the same - for instance, suppose disclosure of some kind led to surgery 100% of the time, just as a surgery requirement would - why is the disclosure superior? How is it less of an imposition on human freedom? If I know disclosure will impact your behavior in a certain way, isn't that disclosure just as much an imposition on you as if I required you to behave in that way? Again, I mean this purely in a moral sense. I can certainly imagine that psychologically, we might feel better if our behavior is the result of disclosure and our own "free choice" rather than the direct heavy hand of the State.

Someone wittier than I once told me that obtaining truly informed consent would require putting every patient through med school.

"Rather than discounting the surgeon’s recommendation, patients reported increased trust in physicians who disclosed their specialty bias."

This relates to the above media post. The media can regain public trust by letting everyone know how biased they are. Everyone knows where Breitbart lines up, they don't lie, and they're winning.

No one is bias free, and acting as if you are makes you fake. Admitting makes you honest. People like that.

That sounds like abandoning virtue for honesty. Not really a good trade?

There's nothing in it for PBS. Now that Jim Lehrer has retired, his program is difficult to tolerate for anyone who does not share the outlook of the two hosts, who carefully frame every issue and line of questioning to exclude certain alternatives. Their amour-propre and that of their core audience is maintained by pretending they're not doing what they are in fact doing.

It's not just the recommended therapy which is biased. It is the outcome which is defined as success. When my dad had brain cancer (a metastasis of his lung cancer), he was treated with radiation, which is the standard of care. No dispute there.

The cancer was successfully treated, but my dad had the serious side effect of intractable chronic hiccups which was a major impact on his quality of life. I suspect the radiation oncologist errs on the side of too much radiation to ensure that he meets his definition of success. But if he took a broader view of success, perhaps he'd throttle back the radiation a little.

My back surgeon, in an effort to be fair, did not recommend surgery until all other options were exhausted, even though he believed surgery would be the final result. This cost me several extra thousand dollars, several months of additional pain, and additional permanent damage that was unnecessary.

In a fair few types of back problems, surgery is nearly as likely to make it worse as better. It introduces all sorts of risks.

Sorry for the situation, but having read quite a lot on this issue for certain types of back problems, in most cases it sounds like the doc made the right call.

Also, I would wonder if the additional permanent damage might have resulted from the surgery itself (as I pointed out, this is common) rather than the fact of waiting longer for the surgery. Not my business, but there's a possibility that it might be relevant for your consideration.

Perhaps the best surgeons will be most confident in making the recommendation, but are also aware that their confidence and success will lead them to have a bias to always recommend what they are good at?

This certainly wouldn't explain everything. But if they could correct for a measure of quality this would deepen understanding of what's going on here.

Personally, I have more trust when explicitly informed of bias. In such cases I am less inclined to believe that I will become the object of manipulation, and that even in cases of disagreement or diverging interests you're more likely to be dealing with a straight shooter.

More likely to have surgery? Maybe. Three times more likely? Sounds fishy. Get Andrew Gelman on the phone.

PNAS is does not publish only top quality research.

Comments for this post are closed