• When is the shared decision-making model appropriate?

    It’s as if Kenny Lin* read my mind about the PSA test.

    The answer is that primary care physicians have already been trying this [shared decision-making] approach for many years, and it does not work. Numerous studies have shown that no matter how much information men are given in any format, the vast majority still choose to receive the [PSA] test, and most of them end up worse off because of that decision. In fact, as a new study concludes in the Archives of Internal Medicine, once prostate cancer is diagnosed via PSA testing, the men most likely to receive aggressive (“curative”) treatments are those who are least likely to need it – men who would have been better off not getting the test in the first place.

    Lin is a physician and he comes to a very strong conclusion that shared decision-making for the PSA test doesn’t work and should be abandoned, that physicians should advise against the test. He seems to know what “better off” means for patients.

    I am not a physician and I am a little less confident (but not much less) than Lin that I can know what makes patients “better off” when it comes to the PSA test. Maybe I shouldn’t be so timid, but it’s my nature. Nevertheless, my main objective is to push back on the idea of shared decision-making as a concept that should be universally applied. I am willing to believe there are some tests, procedures, or medical services that the patient really shouldn’t have much say in. These could be things that are demonstrably overwhelmingly harmful and we just should not do them. There’s no point in educating patients about things so they can possibly choose a dangerous path. (The PSA test may not quite fit this description, but it might be close. We should not let patients hem and haw about whether they should undergo bloodletting though. There, I’m confident.)

    Or there may be therapies that are, essentially, unique. If you arrive at the hospital with a hip fracture and want treatment, your hip will be reconstructed. There’s not a lot of wiggle room. If you have colon cancer and want treatment you’ll get a colectomy. Again, not many choices there. (Both examples are from Jack Wennberg’s book Tracking Medicine, which I am enjoying.)

    Bottom line: Shared-decision making is not the only appropriate style of care delivery. It may be suitable for some care and not for others. That’s something worth discussing and I’m glad Lin raised the issue.

    * I’m inferring Kenny Lin is his name. All I can find is his blog handle which is kennylin. Why do some blogs not have clear by-lines?


    • I agree on the harmful cases, and probably also on most “medically unnecessary” cases.

      But if I’m reading you right (and I may not be), there is an unstated premise in your treatment of what you call here the “unique” cases. You seem to assume that if the patient desires the end goal, she thereby must accept the (standard) means to get to it. This may be true in most cases, but it’ll be false in enough cases to raise questions about what can and cannot be left undiscussed.

      Suppose a devout Jehovah’s Witness (JW) seeks an elective procedure. In the consultation with the specialist, the JW says things like, “I really just want to get back to living normally, which is why I have come to you.” Now one reading of your post suggests that the doctor may then permissibly transfuse the patient if something goes wrong on the operating table. In explaining to the patient what happened after the fact, the doc might say, “There’s not a lot of wiggle room. If you have colon cancer and want treatment you’ll get a colectomy. If you have condition X and you want elective treatment Y, *and* if you have a desire to get back to normal, you’ll get a transfusion.”

      I assume you don’t want to suggest that an expressed desire to “get back to normal” thereby renders unnecessary shared, prospective decision-making about what to do if transfusion is indicated (or whether to ask if the patient has any strong religious views that the care team should know about). Likewise, *I* certainly don’t want to force doctors to ask patients ahead of time if they have conscientious objections to different types of suture techniques. So I want to draw the line too, and I admit I don’t know where exactly to draw it. But the quick treatment you give to the “unique cases” is too quick (again, if I’m reading you right).

      • What’s too quick for a philosopher is often just right for a blog post by an economist. The conversation always needs to be taken to a deeper level. Sometimes that’s an excuse to not draw any lines. I know that’s not what you’re saying, but I am. I’m not disagreeing with you!

    • Kenny Lin is Kenny’s name, and I disagree with him on this one, despite having written an article in the NYTimes Magazine that strongly supported Kenny’s work at the US Preventive Services Task Force, which found that PSA testing does not offer a mortality benefit. Here’s the link for those who are interested:

      Shared decision making is not about getting the patient to do what the doctor wants him to do, which judging from Kenny’s blog he thinks it is (and his opinion about the PSA test is, understandably, “Just don’t do it!” ).

      Telling patients what to do is not the point of shared decision making, and frankly as a patient, I find myself bristling when doctors, even those who are caring and well-informed, insist that they know best about what I want to do with my body. The point of shared decision making is to help patients: a) understand that ELECTIVE decisions mean the patient has a choice; b) understand the tradeoffs involved in each of the choices; and c) come to a decision that is in keeping with their values and preferences.

      What I think Kenny is saying (sorry to put words in his mouth!) is that PSA testing should not be considered an elective test. I don’t know what the right answer is there, but there is a reasonable case to be made that some men might want it, even after understanding as much as Kenny does about how lousy the test is. And it is a lousy test — as one researcher puts it, PSA is about as good at predicting your chances of having prostate cancer as the color of your eyes. And yes, the randomized controlled trials did not find an all-cause mortality benefit from PSA testing. But there’s evidence to suggest that it might possibly reduce your chances of dying from prostate cancer, though not by much. So for a man who would rather die of ANYTHING but prostate cancer, even the treatment for it, then maybe he wants the test. He values avoiding a prostate cancer death, and while early diagnosis is no guarantee that he will, he might.

      So while I think getting a PSA test is a really bad idea, and I don’t think I would get one myself if I were a man, and I’m glad my husband decided to forgo the test after viewing a patient decision aid, maybe it should still be considered an elective decision.

      What do the data say about shared decision making and its effect on patient decisions? A Cochrane Collaboration systematic review of more than 80 prospective RCTs comparing patients who had access to a patient decision aid (for lots of different elective decisions) and those who got usual care (the doctor tells them about their choices) found that patients who have access to an aid are better informed about the risks and benefits involved in their choices; are more realistic about those tradeoffs; and are on average 20% less likely to choose more invasive options. (That last finding is why policy makers get all excited about shared decision making.)

      When you look specifically at studies of shared decision making for PSA testing, it’s not so clear that men are less likely to choose the test. There are a couple of possible reasons. I can’t pretend to have looked at all of the studies, but from a few that I’ve seen, it’s not surprising that men opt for the test because the so-called patient decision aid that was used urged them to get tested. The other reason I think its hard to dissuade men from getting tested is we have been beating the screening drum in the US for a long time. It is an article of faith that catching cancer early is universally good. It is going to take a long time to unlearn that myth.

      Conclusion: discarding all of shared decision making on the basis of PSA testing is a bad idea.

      • Excellent comment. Before I respond/ask questions, let me point out that the main purpose of my post wasn’t necessarily to come to any conclusion on the PSA (and you’ll notice I didn’t, despite what I may think), but to suggest that there are certain treatments and tests that may be reasonably argued to properly belong outside the shared decision-making paradigm. This idea may not be controversial, but when we get down to specific treatments, it could very well be. Here, the PSA is a classic example, as evidenced by your response.

        My own views on this subject are in flux and at the moment are shaped by the notion of provider-induced demand. If it is real — and I think Wennberg’s work shows that it is — then it is natural to contemplate how to combat it. (I am presuming it is something we should want to combat. Wennberg’s work supports me there too.) Patient education may be effective, but there are likely limits to how widely they can be effectively applied. In the context of other financial incentives that “support” demand inducement, is patient education enough in all cases?

        What I’m getting at, is that a little pressure against some very low-value tests and procedures may be a good thing, not in an ideal world where everyone has perfect information, but in the real world where most people don’t, despite our efforts, and in a world where providers are subtly pushing their interests. In that world, I do think Kenny Lin’s view is reasonable. That’s not to say he’s right. There are other reasonable views too. But he’s not that far wrong! I think your own comments actually make that point. You just come out a hair’s breath on the other side of the line.

        Put it this way, in a world with no PSA test, would we want to introduce one knowing what we know now? It’s quite reasonable to “just say no.”

    • Great comment, and I’m glad you bring up the concept of supplier-induced demand. I too think it is a very real phenomenon in health care, and I agree a little pushback against low value treatments and tests is a very good idea. The question is, how do you define low value, and what form should that pushback take?

      To me, the answer to the second question depends upon what kind of care we’re talking about.

      Let’s start with things that SHOULD be done. There are certain treatments that all patients who have the condition should receive. You used the excellent example of hip fracture. If your hip is broken, it hurts like hell, you can’t walk, and it almost always ought to be fixed. Another example: a clot busting drug like streptokinase or aspirin or tPA when you are having a heart attack. I don’t want to have a long chit chat in the emergency room if I’m having a heart attack. If the diagnosis is correct, by all means, give me the drugs, stat.

      On the flip side are procedures that don’t provide much value if any and simply shouldn’t be done. Bloodletting, as you point out (unless you have porphyria . . . ) Avastin, the wildly expensive cancer drug that was recently shown to offer no benefit for metastatic breast cancer. Vertebroplasty (injecting superglue into the vertebrae) is another one, along with pulmonary artery catheters, cardiac stents for people with no symptoms (chest pain, shortness of breath) of heart disease, poudrage, high-dose chemo fro breast cancer . . . I’ll stop there.

      Despite the evidence out there saying these treatments are no good, some of them are still being done. Doctors are still prescribing Avastin for metastatic breast cancer, and doing vertebroplasty. Poudrage, happily, was abandoned. I sure hope high-dose chemo has finally been put to rest.

      Finally, there are treatments that are considered elective. There’s more than one way to treat the condition (often including no treatment at all), and each of those treatments involves tradeoffs. If I have stable angina (chest pain due to heart disease), I need to be treated with drugs, and I can choose to get a stent in addition. Stents add nothing in terms of mortality benefit, and for most patients don’t offer any added relief from symptoms. For such treatments, a conversation is in order, because it should be my choice and my values should come into play.

      We know from a variety of literature that my values are often left out of the decision, and it is the doctor’s preference that drives much of the geographic variation we see in rates of numerous elective procedures and tests, including the PSA test. That sure sounds like supplier-induced demand to me.

      So how you combat supplier-induced demand sort of depends upon which kind of care we’re talking about. When it comes to the low- or no-value stuff that doesn’t work and shouldn’t be performed, I think you stop paying for it. (For a list of some “don’ts” check out the American Board of Internal Medicine’s “Choosing Wisely” campaign.

      Shared decision making (and patient decision aids) comes into play for combating supplier induced demand for elective care (what Jack Wennberg calls “preference sensitive care in his book, Tracking Medicine .

      But there’s another category of care that requires an entirely different approach, and this comment has gone on way too long, so I’ll close.

      • We’re clear, and in agreement, that there are three types of care: stuff that should not be done to anyone, stuff that should be done to everyone (with the appropriate condition, that is), and stuff in the middle that isn’t so obviously of overwhelming benefit or isn’t so obviously overwhelmingly harmful, or for which there are multiple approaches that are reasonable to pursue.

        Now, which category is PSA screening in? I think it is reasonable to conclude that it is overwhelmingly harmful. There are those that think it is reasonable to conclude it is not, and that it should be offered to patients. Views will differ even among those with the same goal: to reduce the amount of harm done by routine PSA testing.

        The “it’s overwhelmingly harmful” camp says, “Stop doing it. Just stop.” The “not overwhelmingly harmful” camp says, “Educate patients and they will choose correctly.” It’s not clear which camp an “ideal” patient advocate should be in. It is clear that providers who benefit from PSA testing would choose the “patient education” route over the “stop doing it” approach. This causes me to be slightly distrustful that the “patient education” approach can work. On the other hand, there will be so much opposition to the “stop doing it” approach that that may never get off the ground either (case study: USPSTF changes in mammography recommendations).

        So, even among those that share the same goals, there can be conflict and distrust. This isn’t so helpful to the shared cause, but there it is.

    • I can tell you what I see in practice. I am seeing more patients forego care that we think they really should have, like not treating their diabetes or hypertension. I see patients making these decisions out of mistrust or misunderstanding, not based upon an understanding of the disease. I continue to see large numbers of borderline useful care, like back surgeries, and low yield surgeries like knee arthroscopies. Docs are still very capable of inducing demand (it is pretty easy for most of us if we want to do it). Patients (and doctors) need some way to determine which therapies/diagnostics are really appropriate for shared decision making.

      For those cases where a patient does not want what is viewed as evidence based, state of the art medicine (not provider induced), I think they still have very right to refuse that care. I think the rest of us need to decide if we want to pay for the ensuing complications.


    • I would also add the functionally health illiterate, and the valid tools required to ascertain treatment wishes. If folks cant read a pill bottle or calculate odds, eg, 33% chance you will no longer have erections, etc.,, conventional means (videos, etc) are unhelpful.

      This is a quarter to third of our country, and I doubt studies have captured the full grasp of the problem, ie, using these tools for the soon to be “health system” enfranchised. Not insurmountable, but troubling–and because SD making is still at the vanguard of care, we have not yet peeled back that layer of the onion,

      It may not be getting them to engage in SD making, its having these folks comprehend it. The MCO and Medicare world, with limited special needs and lower SES slices, our are only current barometers. I dont know how providers will engage with this population.

    • Shannon:
      Sorry to come to this conversation so late.
      Thanks for saying what I would have said if you hadn’t. Shared decision making is one model of communication, does not apply in all cases, but should be the default. It requires active listening, patient engagement, and collaboration. Even when I break my hip and am eligible for surgical repair, I want to be given all the information I need and I want to feel that we made a decision together.