The key difference between alternative medicine and evidence-based medicine

Recently, I got an e-mail from someone who had just discovered my blog that made me think a bit, which is usually a good thing. At least, in this case it was. Basically, this reader asked me a question I hadn’t been asked in a very long time and hadn’t thought about in a very long time, specifically: If I had to pick just one and only one, what is the single most characteristic difference between alternative medicine quackery and science-based medicine? True, there are several key differences, and I’m sure many of you could tick off a list of five to ten characteristic differences without even thinking about it. So could I. However, if I had to pick the single biggest characteristic that distinguishes alternative medicine quackery from SBM, it lies in the approach to clinical evidence. In brief, in SBM anecdotes are considered the lowest, least reliable form of clinical evidence. In alternative medicine, anecdotes are the highest form of evidence, and you must, above all, believe them. Failing to do so is to attack the patient. Anecdotes must never be questioned, no matter how respectfully it is done. The patients’ interpretation of the medical implications of his experience must never be questioned.

Consider this. In the evidence-based medicine (EBM) paradigm, observations begin with clinical observation. A careful clinical observation in which the patient’s history, diagnostic tests, treatments, and clinical course are carefully documented is basically an anecdote. (I realize I’ve been critical of EBM for ignoring prior plausibility and fetishizing the randomized clinical trial further above all other forms of clinical evidence than is deserved, but that doesn’t mean the general paradigm isn’t generally correct when prior plausibility is taken into account.) In EBM/SBM, anecdotes are not, straw men by quacks to the contrary, dismissed out of hand. Rather, they are useful as hypothesis generating observations, but they are not sufficient. The next step must be to test the hypotheses generated in more patients. This sort of observation can range from a case series (several anecdotes examined together for commonalities and differences) to retrospective observational studies, to a pilot study, which is usually a small study without a control group to test safety and look for indications of efficacy. Trials of new drugs, for instance, are divided into different phases. The purpose of phase I trials being to test the safety and maximal tolerated dose of a new drug and phase II trials being small pilot trials that might or might not have a control group designed to look for indications of efficacy. Finally, when there are sufficient data from more preliminary trials, large randomized double-blind phase III clinical trials are performed to test the new treatment either against a placebo control or against the existing standard of care. In the case of cancer trials, for example, these are often trials of a new drug plus standard of care versus standard of care. Finally, once a drug is approved by the FDA, phase IV studies are performed to look for uncommon complications that might have been missed in the original trials used to approve the drug.

There’s no need to go more into the details of clinical trials here other than what I’d discussed above. The point here is that in EBM/SBM, anecdotes are considered the lowest form of clinical evidence, with case series only slightly “better” than single anecdotes. As I like to say, the plural of “anecdote” is not necessarily “data.” Rarely are treatments adopted based on only case series, because case series can easily mislead. The only time treatments are adopted on the basis of such unreliable evidence is when the disease or condition being studied is so rare that randomized clincial trials are not feasible. Other questions can’t be subjected to randomized clinical trials because it would be unethical to do so (the usual example is a randomized, placebo-controlled, double blind clinical trial of “vaxed versus unvaxed,” which would leave one group unprotected against vaccine-preventable diseases). In these cases, large epidemiological studies must suffice to demonstrate that, with only a very small margin of error, there is no detectable association between vaccination and autism or between the thimerosal preservative that used to be in vaccines and autism, and they do. There are multiple such large epidemiological studies that do just that, to the point that, barring new evidence as compelling as the evidence that exists, scientists have concluded that vaccines do not cause autism.

The contrast from quackery couldn’t be any starker. I’ll use cancer quackery as an example. In cancer quackery, you’ll virtually always see glowing testimonials aplenty in which patients claim to have been cured or made substantially better by the treatment in question. Never mind how alternative cancer cure testimonials often result from a misunderstanding of cancer prognosis and in particular the difference between primary curative therapy and adjuvant therapy. Strike that. It’s not just cancer quackery. For any quackery, you will find that the key evidence presented by practitioners of that quackery for its efficacy will usually not be randomized clinical trials. For some forms of quackery, bias-prone clinical trials, either poorly designed and performed or actually well designed but misinterpreted as positive, either willfully or inadvertently. (Acupuncture studies are notorious for the latter.)

The point is that in EBM/SBM it is fully understood that individual anecdotes can be profoundly misleading regarding the efficacy and safety of a therapy. Placebo effects, confirmation bias, regression to the mean, and any number of other confounders can provide the illusion of efficacy when there is none. Individual human observations and human memory are extremely fallible. You only have to consider the seemingly never ending testimonials for homeopathy. If there is any form of quackery that is, quite literally, nothing (well, nothing more than water or water embedded in sugar pills), it’s homeopathy. Similarly, if there is any treatment that is even more clearly nothing than homeopathy, it’s reiki. Reiki, as you might recall, claims that healers known as reiki masters can channel the “universal energy” into a patient for healing effect. It’s nothing more than faith healing that substitutes Eastern mysticism for Christian beliefs. It’s magic.

In contrast, in quackery individual anecdotes, which I like to refer to as testimonials, are all. They are not to be questioned. If you do analyze or question them, no matter how reasonably and even if you are not doubting what happened but rather questioning the patients’ interpretation of what happened, you can count on being accused of attacking the patient or of accusing the patient of lying. If you point out that an anecdote does not necessarily show what it is being claimed to show. I’ve done this several times for Stanislaw Burzynski patients, and the result is that I’ve been accused time and time again of “attacking patients.” One patient of Burzynski’s even complained to my state medical board because I analyzed her testimonial and found it not to be evidence that Burzynski’s “personalized gene-targeted cancer therapy” cured her cancer. When it comes to Burzynski and much of the rest of alternative medicine, the cult of the anecdote rules. The sole exception that I can think of is acupuncture, but it’s not really much of exception at all. There are lots of clinical studies of acupuncture that, when taken in their totality are negative, but there are enough positive studies that can be cherry picked.

If you really want to see where the cult of the anecdote rules above all else, however, the antivaccine movement is the prototypical example. Antivaccinationists fervently believe that vaccines cause autism primarily because they have confused correlation with causation. Because children receive several vaccines during the age when autism and autism spectrum disorders are most commonly diagnosed, by random chance alone there are many whose children either regress or whose children’s autistic symptoms become apparent within a short period after vaccination. Add a dash of confirmation bias, and it’s easy to understand why parents so easily confuse correlation with causation, as we humans are pattern-seeking machines and our psyches seem to require identifying a cause when something bad happens. These anecdotes are Gospel, and the “mommy instinct” that leads to the linking of vaccines with autism in these parents’ minds is not to be questioned.

The overall problem is that our cognitive biases are such that we are easily misled. That we are so easily misled is not because humans are stupid or dense (although surely a few of us are). Rather, it’s because the way our brains are wired is such that we are too quick to ascribe causation to correlation and cannot recognize just how fallible we are. It’s also not because we are inherently deceitful (although surely a few of us are). The people repeating testimonials of cancer cures due to alternative medicine or that vaccines made their children autistic fervently believe that the alternative medicine or that vaccines caused their children’s autism. That’s why they are so fervent, so absolutely certain in their beliefs. They think they saw it with their own eyes, but what their own eyes told them was wrong.

In medicine, anecdotal evidence is evidence, but it is only a starting point. SBM/EBM takes that into account and relegates anecdotes to the lowest rung of clinical evidence, to be considered carefully and used for hypothesis generation, but they are not particularly useful for making generalizable conclusions. It took medicine a long time to realize this. Indeed, compared to the 3,000 year history of medicine, the randomized controlled clinical trial as we know it today for testing drugs and devices is very a recent development. It is less than 70 years old. Thinking scientifically in this way and not trusting our own experiences and memories are not easy; they do not come naturally. We had to understand that before medicine could truly advance beyond bloodletting and itinerant healers. Antivaccinationists and alternative medicine apologists would lead us back to those bad old days simply because very human cognitive biases have led them to believe what they want to believe, and, like most humans, they believe their senses more than they believe science. Science is, after all, a very unnatural way to think, particularly when it comes to our health. It takes effort, but it’s worth it because it is how SBM changes and improves based on new medicine.