Anoop Kumar, MD

Yes, another informal logical fallacy

the idea of “traditional treatments that… have been used for centuries.”
__

a.k.a “Argument from Tradition”

See my Dec 4th REC Blog post on Mario Bunge.

Dr. Katz, I once again genuinely thank you for your continued discussion in the matter. I can absolutely assure you that I care not from where or whom a therapy comes – so long as I can be convinced it can help my patient I will use it.

I find it interesting that you disagree on the topic of medical students learning the hierarchy of evidence. In my first year of medical school we had an entire year-long course on it, with exercises, projects, and questions on our exams pertaining to it. Similar for a number of friends and classmates at other medical schools. Perhaps things have changed since you graduated? Or perhaps my small sample size is biased. In any event, I wholeheartedly agree that it should be taught to all medical students in a rigorous manner and with an added emphasis on Bayesian frameworks which incorporate all relevant evidence, including bench sciences, to inform decision making. The idea that only RCTs and meta-analyses can inform decisions seems perverse to me (and indeed has been dubbed “methodolatry” by some) and is not at all what any of my colleagues and classmates think is appropriate.

But that hierarchy of evidence cuts both ways. For example, when there are equivocal clinical studies in homeopathy, we can look at the bench sciences to inform us that the prior plausibility is so low that equivocal studies can be rightfully regarded as resoundingly negative. A point on which you have failed to expound – your own admitted use of homeopathy with patients.

You certainly make a valid point about perverse incentives (as I like to call them) nudging us against non-proprietary, low profit modalities. There is a very good reason why my institution has banned drug company lunches for many years and why I support such initiatives. But that still does not justify the use of unproven or disproven modalities in patient care. We can – and should – lament when we are at a lack of tools and knowledge to help our patients, but I fail to see how picking something at random is any better than admitting the limitations of our current abilities. It seems to me much more productive and better for our patients to instead focus our efforts on improving the process.

Which brings me to the idea of “traditional treatments that… have been used for centuries.” I wholeheartedly disagree that this is a valid surrogate to use for establishing the probability of safety. To use a cliche, bloodletting was used for many thousands of years and it is clearly not safe nor therapeutic. But in a more serious example (of which there are myriad) I would like to draw your attention to aristolochia. That is an herb that was used for at least many hundreds of years, if not thousands, and is in the Bencao Gangmu, now known as the Compendium of Materia Medica of Traditional Chinese Medicine (TCM). The Bencao Gangmu was compiled in the 16th century and there is evidence of the use of aristolochia dating back to the 1st century CE and even ancient Egypt. By your argument one would expect that to be an excellent establishment of a high probability of safety. Yet, in 1993 it began to be noted that people who ingested this herb had unusually high levels of urothelial cancers. A decade later it was noted that so-called “Chinese Herbs Nephropathy” and “Balkan Endemic Nephropathy” are likely the same thing and finally in 2013 it was demonstrated that aristolochia is a potent carcinogen, with it’s effects not limited to patients ingesting it, but also the herbalists producing it. As is noted in this study a full one-third of the population of Taiwan was prescribed something with aristolochia in it. In the words of the authors:

…it is likely that upper urinary tract carcinomas and their attendant aristolochic acid nephropathy are prevalent in China and other Asian countries where Aristolochia herbs have been used for centuries in the treatment and prevention of disease, creating a potential public health problem of considerable magnitude.

By your proposed approach to judging the potential safety of therapeutics, aristolochia would be a reasonable therapeutic to try (prior to 1993 and arguably up until the early 2000’s) as it meets all of your requirements as outlined in your CURE paradigm: it has been used by many cultures for thousands of years (probable safety), it would come on the recommendation of a person trained in an accredited program, with national certification, standardized education, and government regulation and licensure (an acupuncturist/oriental medicine specialist; possible efficacy), and unclear scientific evidence of efficacy (ambivalent evidence of efficacy prior to 1993). In our hypothetical patient for whom we are low on options, it could not have been argued by you that trying aristolochia would be anything but acceptable.

And yet, we would have been exposing our patients to a greatly increased risk of cancer with no evidence of benefit and no rational reason to assume there would be, beyond TCM use based on pre-scientific approaches to pathophysiology.

My own background includes an undergraduate degree in medical anthropology in which I studied many different systems of medicine, as well as advanced coursework in molecular pharmacology, and post-graduate research in pharmacognosy (specifically investigating the anti-senescent properties of botanical extracts). This is but one of many examples of the pitfalls of assuming safety based on proxies such as traditional use and an illustration that relying on traditional use as indications for therapy is almost the same as walking into a field and picking a random plant to give your patient. Even thousands of years of use is worthless if the results aren’t carefully recorded. That’s why we invented science in the first place.

The key to your argument is “when risk is nil” – but that is not something I believe can be established based on the methods you are proposing. And it is well demonstrated that something is more likely to harm or do nothing than be of any benefit. Else we would have our patients chewing on yew trees and periwinkle flowers and the field of pharmacognosy would be of very low value.

I would posit that in cases of few options the best thing we can do to genuinely help our patients is to support and empower them, not give them placebos or experiment on them. If I had tried aristolochia and later learned that I was likely the cause of my patient’s urothelial cancer and nephropathy that would seem to outweigh any slim chance of benefit for both of us.

In conclusion, I do fully agree with your conclusion in principle. But I disagree that a “more fluid form of evidence” as you describe is the way to achieve those laudable goals. There is no such thing as a free lunch and with any intervention that could possibly have benefit there is also necessarily the possibility of harm. Without due diligence beforehand, we have absolutely no idea where that balance may lie and are betting on the benevolence of Mother Nature and the wisdom of pre-scientific thinkers. Those are not bets I would take, if I were on the other side of the exam table.

I would also note that forces in our culture conspire against the utilization of non-proprietary, low-profit modalities. It’s just naive not to notice this, or think otherwise.

Finally, there are ‘traditional’ treatments that in many cases have been used for centuries. While the accumulation of such experience is no substitute for randomized trials in proving efficacy, it can serve quite well to establish the high probability of safety. That, of course, is an important consideration- since what matters ultimately is the risk/benefit ratio. When risk is nil, and options are few- even a slim chance of therapeutic benefit may be worth taking.

What I am pleased to see is that we are not really debating evidence, or nomenclature- we are debating methods of practice, HOW to apply evidence to greatest effect. That, I believe, is the correct focus. The notion of ‘this or that,’ ‘conventional or alternative’ is a distraction. What we should care about is: what is the best thing I can offer any given patient in need, and how do I get there from here? Were we all to devote our efforts to that challenge, I think we- and more importantly, our patients- would benefit considerably.

Once again, thank you for your time Dr. Katz.

Superficially we certainly agree. I wholeheartedly agree that ultimately getting the job done is what matters and there are, as you know, many barriers to achieving that in practice beyond the limitations of scientific knowledge.

Your CURE construct is also superficially agreeable. But it also strikes me as excessively vague. It also, once again, does not jibe with your use of homeopathy. All of your categories include a patient preference for something that works. Homeopathy cannot work.

It seems that here we disagree on the what constitutes that scale of evidence. Homeopathy is a good litmus test because it cannot possibly meet any definition by your own rubric that exceeds a “never” in terms of utilization frequency. Yet you seem to have used homeopathy at the suggestion of a naturopath, which goes against your own (albeit vague) recommendations.

Which raises another point from the paper you linked – your standard is:

As the term CAM refers to any therapy or provider outside the mainstream, any modicum of efficacy or legitimacy satisfies this label. The CAM disciplines with the most promise of successful integration into conventional medicine are those with tangible standards. Specifically, those that have accredited training programs, national certification, standardized education, and government regulation and licensure.

…essentially nonexistent. And certainly not at all based on actual scientific legitimacy. By this standard if a sect of voodoo priests had enough lobby power in Congress they would be deemed worthy to collaborate in patient care and offer suggestions for treatments. And scientific legitimacy is demonstrably not a requirement for accreditation, standardized education, or government regulation. Bob Jones University is a prime example of this, as is Bastyr and the entire field of naturopathy, which you none the less seem to embrace. You even say:

The ultimate goal of integrative medicine should be to make the widest array of appropriate options available to patients. Appropriateness should be predicated on fundamental considerations that pertain equally to conventional and CAM practice: treatment safety and treatment effectiveness. Treatment safety and treatment effectiveness must, in turn, be interpreted in light of the available evidence

Once again we agree in principle, yet in execution something is profoundly lost. The entirety of your argument – and this particular paper – boils down to nothing more compelling or novel than saying we should follow evidence, starting at the top and working our way down as needed. This is something I knew before starting medical school but every first year student learns (or should at least). The only difference is that you seem to have no limit as to how low down the evidence totem pole you are willing to go. If a naturopath recommending homeopathy seems like a reasonable option, why not every old wive’s tale that any random person can conjure up?Do we not owe it to our patients to do more than just flail at them with every nostrum we can throw at them, no matter how ridiculous?

There are limits to medical and scientific knowledge. Limits that are continually pushed farther and farther out. Not fast enough, but we can (and should) always feel that way. When we’ve reached those limits, abdicating our responsibility to our patients to understand, to empower them to cope and manage their condition in favor of what are nothing more than wild guesses at best is no better than the heavy paternalism of our forebears.

If you genuinely do wish to “get the job done” (and I truly believe you do, as do I and all of our colleagues) abandoning the principles of rigorous scientific inquiry to the ilk of naturopaths, acupuncturists, TCM practitioners, and chiropractors is not the way to do it. One does not move forward by traveling back in time to pre-scientific nostrums, nor do we improve patient care by abandoning standards.

I still genuinely fail to understand what value your proposed integrative medicine adds to patient care beyond treating ourselves and the patient to placebo medicine.

Actually, I prefer no separate category. What matters is how we proceed when we have exhausted all of the ‘high quality’ evidence, but our patient’s need remains inadequately addressed. Call that CAM, or ‘the next best thing,’ or the ‘art’ of medicine. The rubric matters little; getting the job done matters a lot. The method we use in my clinic is the CURE construct on p. 27 of this paper-

http://www.iom.edu/~/media/Files/Activity%20Files/Quality/IntegrativeMed/Preventive%20Medicine%20Integrative%20Medicine%20and%20the%20Health%20of%20the%20Public.pdf

Dr. Katz, thank you for taking the time to respond.

I am generally overly wordy so I’ll try and be concise as best I can.

In essence, I am reading your argument as saying that in the face of immediate patient need we should lower the standards of evidence to allow for less evidence based therapies to be administered to our patients and that further this is what “CAM” allows us to do. In other words, “CAM” is nothing more than a term for creating a double standard in which therapies with inadequate evidence to support them can be ethically administered to patients outside of a clinical trial.

I take multiple issues with this approach and this argument.

First and foremost it does not at all explain how you can possibly find homeopathic treatments to fit into this proposed paradigm. By what possible rationale can you assert that homeopathy is a “promising, but not truly ripe” modality? If there is one “alternative medicine” that can be less promising I cannot think of one. This can be repeated for nearly all things labeled “CAM”, including many of which you are on public record as advocating (or at least having administered to patients).

But to address the actual argument you are offering here, it still fails to provide any reasonable rationale for the notion of calling these lower-evidenced therapies “CAM.” Why create a separate category? If the premise is simply that the (admittedly flawed) machinations of medical research are too slow in certain regards, then it makes no difference whether something is “CAM” or not. There is no special validity that somehow grants a higher prior plausibility to something labeled “CAM” than any other random thing you may wish to experiment on patients with (and make no mistake, providing an a treatment in the absence of evidence is experimentation). In fact, you explicitly make that argument but to a degree I am forced to disagree with. Was it not you who said:

I would not deny the implausibility of the therapeutic influence from not touching, nor contest the improbability of healing messages left behind by molecules diluted out of solution. But once we acknowledge that little could be more unlikely or wondrously implausible than a handshake, the topic of plausibility must be broached with greater purpose. I don’t understand homeopathy, or believe in it per se. I find it inexplicable and farfetched, but perhaps slightly less so than the experience of a caress, and incalculably less so than our existence.

Neglecting the fact that you are attempting to say that everything is so equally implausible as to be equally plausible, does this not directly contradict the idea of giving an special quarter to something labeled “CAM” as being somehow more likely to be useful in the absence of evidence?

So how on earth does one pick from the literally limitless possible therapies without evidence to experiment on our patients with? And how would you know if it worked? You say:

And the anecdotal evidence that homeopathy might be effective was brought up by the naturopaths. We tried it. It worked.

Besides the fact that there are anecdotes for everything from alien abductions to mind control, you simply cannot say “it worked” because such interactions can, by definition, only demonstrate correlation, not causality. To say otherwise is to abandon the science you claim you believe in.

One place we do agree is when you say, “patient need is immediate, and data generation is slow, there is a need to make the best use today of the data we have today.” But if you are resorting to using homeopathy and listening to naturopaths then you are very clearly not making the best use of the data we have today. At that point one may as well be trying to use unicorn tears to address immediate patient needs.

The failures and flaws of the current research paradigm – the one that declared CoQ10 dead – do not magically validate incredibly implausible and unproven claims. As Ben Goldacre said, flaws in aircraft design does not mean that magic carpets can fly. I would argue that using CAM as a crutch to defend what would otherwise be unethical patient care – either providing placebo medicine or doing experimentation on humans without IRB approval – is contra to your stated claims of giving patients needed treatments in an accelerated timeframe.

We disagree on a number of points, clearly- but I will only address one. The difference between, for instance, co Q10 as “alternative” medicine and co Q10 as “conventional” medicine is measured in years, if not decades. Yes, the research in this case did eventually get done- but it was ten years after under-powered studies were used to declare the ‘co Q10 for heart failure hypothesis’ dead. Does that happen with a conventional drug: it is studied in too small a sample, little is seen, and the ‘hypothesis’ that the drug could work is declared ‘dead’? It does not.

You seem to be dismissing the significance of the timeline. For any given patient in need of an option today, data that will ripen ten years from now are of no real use. What “alternative” medicine does is apply promising, but not truly ripe, data so that today’s problem can be addressed today. The down side to this, of course, is that the agent in question may be ineffective, and potentially dangerous. The up side is that when they are right, they are right ten years or more before we are.

The less profitable the product, the longer the timeline to get a critical mass of research done.

Because patient need is immediate, and data generation is slow, there is a need to make the best use today of the data we have today. That simple imperative leads to therapies for which the evidence is suggestive, but not truly ready for prime time- and thus, integrative medicine. The need to work diligently to differentiate baby from bathwater is, I think, self evident. Best- DK