An open letter to the medical community.

I’ve been following the DTC genomics and personalized medicine discussions for years now. I’ve learned that there are diverse and well-reasoned arguments by capable spokepeople proposing many possible futures for the doctor-patient relationship. Discussions during this weekend’s BIL:PIL conference and a recent exchange with Dr. Steven Murphy, with whom I’ve disagreed, occasionally vigorously, for years prompted me to put down my thoughts here. Actual meeting notes will follow in a separate post.

I believe, as does pretty much everyone, that medicine is on the cusp of great changes and that personalized medicine holds great promise. I believe that an informed patient is an empowered patient, and ultimately a healthier one. Every good doctor should want this, and every below-average doctor should pray this day never comes.

Email transcript follows:

Where Dr. Murphy and I differ is in the regulation of medical tests, diagnostic or otherwise. Why should diagnosis (the test + the interpretation thereof, sans recommendations) be considered to be treatment? Certainly insurance companies shouldn’t pay for quack treatments, but getting a test, even having the result of the test explained to you, isn’t a treatment for anything.

The laws were written back when people didn’t necessarily want to see the raw data because they wouldn’t have known what to do with it if they had it, and the tests were inaccurate enough that they didn’t mean much on their own anyways. That was a good reason to make doctors the gatekeepers, but that’s all changed now.

People neither want nor need a gatekeeper now. It’s not about self-treatment, it’s about empowerment. Patients should be empowered to understand what’s going on with their disease or condition or lack thereof, and they should be empowered to ask better questions. Every good doctor wants their patients to be empowered like this. The only thing total release of all information and interpretive notes, medical or otherwise, would do is reveal which doctors are the good ones who stay up-to-date and which are the slackers.

Patients might get together and discuss their tests and decide as a group that they’re going to undertake some quack treatment, but they might also get together and discuss their tests, bring their concerns to a good doctor, and he might realize that “Gee, maybe we should tell myeloma patients undergoing treatment with Zometa to avoid invasive procedures on their jaw.”

There’s upside and downside risks to these kind of patient rights, but Gilles Frydman at ACOR is an example of how to do this in a way to minimize the downsides while maximizing the upsides. They’ve published a study of how this works here. In short, an informed patient is a healthier patient when they’ve got a community like that. Take a look at what they’ve done at ACOR and you might realize the assumed or historical risks of self-treatment pale in comparison to the benefits that tens of thousands of patients are seeing today.

Things have changed and the medical establishment needs to get ahead of the curve on this one, or they risk losing the trust and respect they’ve been afforded for so long. In fact, it’s this loss of trust that has provided an opportunity for the dangerous quack treatments you see today among autistic parents, for example.

Please, in your advocacy of personalized medicine, abandon the paternalistic model. Seek to empower patients and regain the trust our profession has deserved for so long. Don’t let mistrust and arrogance leads patients away into the arms of the quacks.

Sincerely Yours,

William Gunn

Further reading:
In Iran and in the US Health Care System, Citizens’ Access to Computable Data Frees Everything!
It’s my genome: should researchers be obliged to return genetic data to research participants?
A Mom Brokers Treatment for Her Twins’ Fatal Illness
Participatory Medicine

17 thoughts on “An open letter to the medical community.

  1. Excellent post, thanks William.

    From a number of recent conversations with practicing physicians and geneticists I see this as one of the key issues for the medical profession to resolve in the coming year (and not much longer than that): whether an informed patient an empowered patient, or a dangerous one.

    There are many that share your view, but there also many that don’t, including those who believe that it’s too soon, today, for the sort of empowerment (or lack thereof) that genetic information, particularly of the DTC variety, provides.

    One question to consider is whether there are changes that can be made to the structure of the underlying healthcare system that would serve to shift the balance in one direction or the other. Reimbursement rates for genetic testing and genetic counseling come to mind. Higher reimbursement rates might (emphasis on the might) incentivize healthcare providers to incorporate genetics, even when probabilistic and/or not immediately medically actionable, into patient care and education.

    Whatever side of the debate you fall on, I think it’s probably appropriate to look at some of the larger structural changes as a means to address this issue, because I doubt it will be feasible to effect change one clinician at a time, at least in the time available.

    – Dan

  2. Thanks, Dan. The preponderence of evidence is building up. At this point we have a fear that an educated patient might be a dangerous one, and increasing evidence that educated patients are empowered and more healthy ones. Human nature might dictate that fear trumps evidence, but surely we can at least aspire to something greater?

    I know if I ran an insurance company, I’ve cover DTC genetic testing. I’d probably even pay my insureds to have it done, even if they didn’t share the results with me. Can you imagine that 10 years from now, which isn’t that long in terms of adverse health event intervals, we won’t be healthier because of that information, particularly if it’s available in the large numbers that insurance coverage would provide?

    In terms of the larger structural issues, I don’t mean to place this as another burden on doctors. They’re already overworked, as they never fail to mention. Higher reimbursement rates aren’t the issue if it’s patients that are shouldering the burden of keeping their own results and seeking competent analysis.

    This would all be so much easier under a single payer system, wouldn’t it, though?

  3. Great post!

    I have to agree with both of you as a supporter of the e-patient movement and also as a researcher in personalized medicine.

    Patients must have access to such data, they all should be empowered to be able to become equal partners in the treatment. But only at a time, when this information is totally reliable. These days DTC companies cannot provide patients with really medically relevant information just because this field is still not improved enough. I believe that is why Steven Murphy and many doctors are against DTC genetic testing. Now only geneticists and doctors together with the patients can analyze the results of genetic tests properly.

    In the future, I hope we will be able to let them access such information without any dangerous consequences. But this world is not here right now and won’t be as long as DTC companies are only driven by money and marketing instead of pure science.

  4. Dr. Murphy says he doesn’t think there’s a danger in people having their genome sequenced or SNPs read, provided people still see a genetic counselor for interpretation.

    We know that the best way to get a clearer picture of what a particular SNP or allele really means is to gather more and more data sets from as diverse a population as possible. That’s a strong reason to encourage this practice.

    What are some reasons to discourage the practice? It’s possible that DTC genetic testing might feed some people’s hypochondria, and a misestimate of risk might inappropriately alarm some people. What’s the worst that could happen here? Opponents of DTC genetic testing are acting like people are going to go and jump off a bridge because they have a certain SNP, but in the absence of evidence this sounds more like fearmongering than true concern. Many people feel that the
    “we don’t really know and it’s risky giving people incomplete information” argument is just a way for doctors and geneticists to rationalize legislation mandating they get a slice of the action.

    In the final analysis, it’s up to the individual to decide their own level of risk tolerance, not for the AMA or anyone to decide that what level of detail someone can handle. That’s what people mean when they talk about the paternalism of the process. If you tell me I have a SNP that gives me an elevated risk for brain cancer, I’m not going to go jump off a bridge. I’m going to continue living as I have because I understand what risk and probability means. I’m not the only one who does, either, and if doctors and medical professionals don’t get their act together you’re going to lose us and lose any chance in being able to participate in the development of the science because researchers and patients are going to get together and figure out what this stuff means for ourselves. This is already happening, and the community wants to include doctors in this, as our equals.

    As Dan alludes to above, the medical community has a choice it must make quite soon: swallow your pride and join with patients and scientists and let’s figure out what this means together, or try to maintain your gatekeeper status through fear and legislation. Before making your decision, however, consider that what’s going on here, the underlying social dynamic, is disintermediation. Disintermediation is what disrupted the music industry by taking out record companies as the gatekeepers. Disintermediation disrupted newspapers by taking out professional journalists as the gatekeepers. It’s currently shaking the foundations of academic publishing by targeting academic journals as gatekeepers of scientific knowledge. Do you really think you’ve got a chance in hell of remaining as the gatekeepers of medical knowledge?

    The real danger and the urgency, why I’m writing this now, is that if you keep trying to tell people they can’t handle the information, it’s just going to make them trust you less and give an opening for quacks and fraudsters to take advantage of people. Please, don’t lose sight of the goal we share – to improve people’s lives.

  5. William –

    Your final point is an excellent one that illustrates very concisely one of the reasons DTC genetic testing has become such a flashpoint. In a very real way the hue and cry over DTC genetics has almost nothing to do with what those companies are adding to the traditional healthcare framework (i.e., personalized genetic data of debatable personal or clinical utility) and everything to do with what they stripping away from that framework (i.e., disintermediation).

    – Dan

  6. There is certainly widespread resistance in the clinical genetics community to removing the health care professional from an intermediary role. Attributing this resistance to fossilized paternalism and protection of privilege is an inspiring meme in this age of self-empowerment through free access to information. But there is no evidence base to support this assertion.

    Look at it objectively. No-one makes money interpreting genetic test results or doing genetic counseling (all due respect to Dr Murphy). So there is no financial incentive to holding back on DTC genomics. What providers are trying to do here is follow the classic injunction to first avoid harm, both individual and societal. And, yes, this is a situation when the potential harm is more evident to professionals. Would you have expected the average investor to understand the risks inherent in CDOs before last year? How many people’s 401(k)’s went underwater in the era of self-empowered investing?

    Even those in favor of DTC testing should admit that the downstream effects (good and bad) are unknown. It remains to be seen whether or not people who receive genomic risk assessment without appropriate education will “raid the medical commons.” Early studies suggest that individuals with genomic risk information anticipate seeking guidance (and presumably care) from the medical community in response to that information. And the question of adverse outcomes remains. Remember, people in REVEAL received counseling. Even if there is no harm on a group basis, individuals may suffer unanticipated adverse outcomes. You may not throw yourself out the window learning that you are at increased risk for brain cancer, but you might be more distressed if your father and brother died from that awful disease and this suddenly popped up on your profile.

    All that said, I think the focus of this debate is in the wrong place. In my opinion, too much of the discussion surrounding this issue has centered on the question of the utility of the information, be it clinical utility or personal utility. The more relevant question is the clinical VALIDITY of the information. I doubt that many providers would care much if everyone carried their sequence around on their iPod. The issue today, which will hopefully become less of an issue in the future, is that the meaning of that sequence in terms of actual risks remains unclear. There are different algorithms for translating sequence data into risk estimates, all of which are founded upon reasonable but unproven assumptions and none of which are validated in any scientifically acceptable way. Studies of self-selected volunteers who may or may not choose to share relevant outcome data are not scientifically rigorous enough to define those risks robustly. Until that research is done, the beliefs that people have regarding the meaning of their genetic sequence data are completely magical. Once the day comes that the actual health implications of sequence information are established more clearly and robustly, I think the perceived resistance on the part of the health community will fade. Noone objects when people measure their own blood pressures or get their cholesterol checked at a health fair.

  7. Thanks for your comments, Mark. The history of the effects of disintermediation brought about by the internet and those who operate within it is pretty clear and supports my thesis that the gatekeepers will become marginalized. It doesn’t matter what their motivation is, it only matters that they’re in the way.

    I agree that we don’t fully understand the risks either upside or downside. If that’s the case, though, given that the default action would be to let the businesses operate without interference, where is the motivation for interfering coming from? You mention that professionals have a greater understanding of the risks. What risks does your allegedly greater understanding bring to mind? Anything specific? I can point to many examples where distrust of doctors has caused harm (vaccinations, for example), so if you’ve got real evidence that DTC genomics is or would hurt people, why aren’t you talking about that instead of this vague fearmongering?

    The issue is simple: on one side we have consumers getting all the data and analysis anyone can provide, both from DTC sources and via genetic counseling and on the other side we have consumers only getting information from doctors because “it might cause distress”.

    My argument is simple: The risk of perhaps causing distress in someone who we’re assuming is really distressed but for some reason doesn’t seek genetic counseling is a far smaller risk than
    the risk that public mistrust of doctors will grow and quacks will prey upon the vulnerable.

    If you’re serious about “doing no harm” you should really be on the other side of the issue.

  8. Lots to unpack there. First, the risks that professionals are more inclined to take seriously are those that we have seen in our own practices with respect to testing for more highly penetrant conditions, where there certainly are individuals who experience significant distress even if the “average” person does not. Will that happen in the genomic testing setting- maybe, maybe not. But it is not “vague fear-mongering” to suggest the possibility. Even James Watson decided he did not want to learn his ApoE status. This is particularly a concern in light of my contention that you did not address, which is that the risk estimates provided by the companies resulting these tests are not validated, and may or may not bear any relation to reality. So, we are potentially exposing people to incorrect information that MAY cause distress. And the recipient (the DTC consumer) does not usually have the expertise to determine whether or not what he/she is being told about his/her risk is likely to be accurate. “Trust me” says the company, while reserving the right to “adjust” the risk estimate in light of future data- not peer-reviewed, no transparency- and promoting it as a seling point. You have to stay in touch with them, paying the subscription fee, because, well, what you have been told today may be completely changed tomorrow.

    The societal risks of raiding the commons are also very much a concern. If individuals are going to seek medical care consequent to receiving results of these tests, then they are going to have to get that care from the medical establishment. Take Jeff Gulcher’s case- it worked out well for him. But are we (as a society) prepared to respond with prostate biopsies in a large group of men who happen to have risk alleles? And what will the societal costs of that be? And personal costs- the idea of having a prostate biopsy makes me cringe. I contend that professionals are more aware of the nature of those risks because they are more aware of the clinical pathways that follow on from a given presentation (in this case a test that suggests increased risk based upon incompletely validated data).

    But forget about the risk and protection from harm argument for now. Forget about the issue of raiding the commons. Even forget about the issue of social justice (where the wealthy, or relatively so, can get their genomic information while the less well-off cannot). My fundamental problem with DTC genomics is what I described before and above reagrding clinical validity. The sequence data is fine, but the interpretation of that sequence in terms of what it “means” is not robust. It is, at this point in time, basically high-tech fortune telling with a veneer of legitimacy that comes from it being “scientific” and “genetic.” People can have their sequence data- in fact, I think everyone will one day have their sequence data. Whatever. The reasoning behind “interfering” now is that there is insufficent evidence at present to support the accuracy of what companies tell consumers their sequence means. And that is a “truth in advertising” problem, if you will. That may change, but not without significant investment from the research community in validating risk estimates. It will not come from crowd-sourcing, no matter what some concerns would have one believe.

    Lastly, with regard to disintermediation (quite a word, there)…I think you are correct and that the well-off will be able to get sequence data without going through the medical system (see above with regard to social justice). That cat is likely to be out of the bag unless some regulatory agency decides to intervene in a way I think unlikely. But the Internet will not “disintermediate” medical care. To get a prostate biopsy one is going to have to go to a urologist. And that urologist is going to want to be very comfortable with the data that generated the risk estimate that is leading him/her to stick a needle through your rectum. Unless the test results are validated through a traditional (sorry about the word, but there it is) mechanism of peer-review and so on, it probably isn’t going to happen. So, for genomic risk assessment (DTC or traditionally-mediated) to reach its potential for improving health, we’re all going to have to work together to define the accuracy of risk estimates flowing from that assessment. It is my opinion, fairly strongly held, that premature commercilization does not facilitate that end because it introduces an inherent conflict of interest.

  9. So you’re saying that in your professional experience, there are probably dangerous risks of emotional distress that may or may not occur that the average layman wouldn’t necessarily know to watch out for? Sorry, but that’s EXACTLY what I meant by vague fearmongering. It’s a textbook example.

    We’re in agreement that it’s quite early and most risk estimates are subject to change. So what? Isn’t it up to the person to decide whether or not they’re comfortable with a risk estimate that might change? Don’t presume to make the decision for them. Perhaps where we differ on this point is that I believe people, especially the self-selected group paying hundreds of dollars for the service, are smart enough to handle it. I’ve still seen no evidence that any harm would come from someone being unduly alarmed, certainly nothing serious enough that the merest possibility of which is enough to shut down a whole industry. Surely whatever harm did arise would pale in comparison to the harm that frauds and quacks are doing this very moment to patients turned off by this medical paternalism?

    I also find it a little hard to believe that real harm would come from bunches of people suddenly deciding to seek preventative medical care. DTC genomic customers are coming in a trickle right now. There’s plenty of time to see a problem before it occurs here and considering the millions of unnecessary medical procedures that are performed every day, not to mention the more serious incidents of blatant up-coding, it’s actually kind of galling to hear this even suggested as an issue. Isn’t there something in the Bible about “the beam in thine own eye”?

    Your main point was about the validity of the tests. Confidence intervals narrow as you add samples. Surely if the validity was a concern, we’d want more data, not less? I understand if you’d rather wait and let research take its course. That’s a prudent decision on your part, but why take away the right of anyone else to make their own decision about how comfortable they are with the data as it is today? If a DTC genetics company were to claim that they had all the answers to life, love, and money based on a SNP scan, they should certainly be shut down, but that’s not what 23andme is doing. They go out of their way to try to make sure people know what they’re getting and their clients are predominantly educated people who can understand what they’re getting. There’s no reason at all they should be denied this. Truth in advertising claims should be decided on a case-by-case basis, not used as an excuse to shut down a whole industry where many of the players are in fact operating responsibly.

    The effect of disintermediation will happen first on the interpretation side of things, not on the physically getting a biopsy side of things, but who knows what the future holds. Either vague unspecified risks or great promise of better days, right?

  10. I guess I must not be making myself clear. I am not promoting “shutting down” the DTC industry. From my perspective, it is not desirable to frame the discussion that way, because it is a distraction from a consideration of the real issues with genomic risk assessment. As I said in my first comment, the meme that DTC companies are some sort of David standing up to the corporate medical-industrial complex is a romantic fallacy that the companies promote to appeal to their constituency and avoid a deeper discussion of the flaws in what they are purporting to do. If people want to spend their money to play in the sandbox, by all means they should be allowed to do so, within the usual constraints of truth in advertising. But they should be very clear about what they are getting. And, to address another straw man that you’re setting up- that this is all about paternalism and people should be free to choose- I am not saying that the risks I’m concerned about are not foreseeable by non-professionals. However, I do believe that the risks are under-appreciated by non-professionals, and even sometimes by people who choose to undergo genetic testing after counseling. People are sometimes surprised by strength of their own negative reactions- personal experience, not fear-mongering. As long as DTC consumers are such a select group, I agree it’s less likely to be an issue. The concern is that the DTC companies know that the current model is not sustainable, so they want to push their product out into broader settings- “medicalize” it, if you will. The problem is not so much how people get the information- it is just as useless (and yes, potentially harmful) whether obtained through the Internet or by order from one’s family doctor who is presenting themselves as a “personalized preventive medicine” specialist. The information itself is faulty- not the sequence, but the interpretation.

    So, yes, we need more data(research) before employing genomic counseling in medical care. But that data is not going to come from the kinds of convenience samples being gathered by the DTC companies. It will come from things like the Wellcome Trust study and the nascent US study. Unfortunately, those data will be a long time coming. And, until those data arrive, the use of genomic information for risk assessment is premature.

    Your overall tone seems to be is one of anger and disappointment with the medical profession. The problems with the current billing system, for example, have nothing to do with the question of whether or not DTC consumers will attempt to access care on the basis of their genomic risk predictions. It’s unfortunate that you have had experiences that have brought you to that place. I don’t see “disintermediation” as an unalloyed good. What would happen if we “disintermediated” access to medications? The Internet is a wild and woolly place when it comes to health information, and most people have not been trained in a way that helps them differentiate between good information and bad. Furthermore, the Internet is an unregulated information marketplace in which success is basically all about who has the best marketing, not who has the best information. The idea that people who have no training in interpreting genetic epidemiology studies will be able to consistently make informed judgments about the accuracy of risk estimates strikes me as naive, just as it is unreasonable to suggest that a non-specialist could accurately assess technical arguments provided on the Internet about, say, cold fusion. Most people still need trusted networks to help them determine what is more or less worthy of belief. As I suggested earlier, DTC companies have an inherent conflict of interest that should lead one to at least question their good faith in interpreting your sequence data for you. Despite all the changes in medicine, and all the problems with our system, your doctor still has an ethical and legal duty to act in your best interest. And he/she doesn’t yet make you sign a disclamatory EULA before giving you advice. And you can sue them if the advice is bad- not so with the DTC companies.

    Well, it’s your blog and this is a bit of a two person conversation, so I’ll bow out now and let you wrap it up. Thanks for the chance to articulate my thoughts.

  11. Not at all, Mark! You’re making yourself quite clear and I’ve enjoyed discussing this with you.

    Tell me if I have anything wrong here:

    You feel that there are risks, underappreciated by non-professionals, that someone might experience undue emotional stress upon finding they have a certain SNP or mutation, and all their worry might be for nothing because there’s so little we currently understand about what DTC genetic testing can really tell us.

    You think more research can be done, and consumer genetic testing has no role to play in this – all the valid research will come from future research studies conducted by traditional methods of study recruitment.

    DTC genetics companies are selling consumers a bill of goods that aren’t quite what they say they are.

    Is that a more or less accurate summary?

    I agree there are risks to letting people have access to this information raw, without the mandatory guidance of a health professional. Imagine if someone were to say, “I don’t think people should be prevented from traveling around in automobiles, I just think it’s too dangerous to let them do it on their own. Professional drivers undergo a lot of training and if the public were to all drive around themselves, we’d see a large amount of injuries and deaths from wrecks caused by unskilled drivers.” In fact, automobile accidents kill tens of thousands of people ever year, but it remains a freedom accorded to citizens, provided they meet certain criteria. Doesn’t it make far more sense to talk about what sensible regulations for the industry should be than to simply state that it’s risky and potentially useless? There’s a wide gulf between what a company like 23andme is doing and some of the others, so it’s not useful to consider them all under one label and your comment about truth in advertising and conflict of interest is very much valid in some cases, less applicable to others.

    While I agree that more research needs to be done, I think you’re perhaps not aware of the things going on with communities such as ACOR. They’re doing research and publishing it, and the data comes directly from the community. Crowdsourcing DOES work for research, if it’s done right. Isn’t it at least worth a shot, especially since there’s a company like 23andme who’s willing to throw large amounts of resources at the problem? Research is being done, people’s lives are being improved, and no one is being hurt. Why discount this as a possibility?

    I mention disintermediation as a way to illustrate what the past has to say about the future here. Disintermediation doesn’t need a spokeman. Good or bad, it’s an inevitable a social force as gravity is a physical one. What the past says will work is to join with the trusted communities that users of social media are developing among themselves and be an educated voice of reason within the community. Don’t expect your real-world authority to give you any elevated status immediately. Treat everyone as if they were your equal. Build anew in each community the bonds of trust. Yes, this takes lots of time, patience, and swallowing of pride. I wouldn’t blame anyone for thinking “How do I have time to do this and see my normal rounds of patients, too?” You don’t have time. Others do have time and new roles are being created. The institutions of the past are failing large groups of people so they’re creating their own solutions. Be there to help them if you can, or support others in that role. What won’t work is to try to remain the gatekeeper of interpretation and knowledge. An individual in one of these groups may not have your education and clinical experience, in fact few do, but as a group they’re still capable of making decisions, and that’s what these social networks are doing. Sometimes, as with ACOR, they make good decisions guided by wise voices that have built trust equity within the group, and sometimes they make bad decisions, as with some of the autism groups, but they’re doing it and they will continue to do it. You can help only by being part of it or supporting those who are.

  12. Pingback: Tweetlinks, 11-06-09 [A Blog Around The Clock] « Technology Blogs

  13. As a 30-year veteran of the health care business, it is my firm belief that too much health information is not a good idea…mainly because it is never interpreted in the right manner.

    Human nature? The evil side of us? Maybe not. But sometimes too much of a good thing is not so good!

  14. Mr. Gunn, You surely have too much time on your hands! How doing some work and paying some taxes!

  15. John Doe –

    As a scientific consultant, I get quite a bit of work coming in via my blog and social networking profiles. In fact, I’d be willing to bet that I pay more taxes than you do, especially since independent contractors pay both their share and the share otherwise paid by the employer.

    Don’t hate me because I get paid to do this 😉

Leave a Reply

Your email address will not be published. Required fields are marked *