How to prove it
I sometimes get frustrated at both sides of the debate over CAM* cancer cures. Yes, I even get frustrated at Orac! The reason is that the defenders of science tend to allow defenders of CAM get away with not proving their case.
The defenders of CAM like to whine that of course they can’t afford to do proper randomized placebo-controlled double-blinded trials, and the scientific community won’t fund those trials, and it’s just all very unfair because people are dying for lack of that proof.
This is bogus**, and defenders of science should hit them with that every time they whine about it.
CAM defenders claim to have cancer cure rates for all cancers upwards of 90%. If this is the case, they can prove it to the complete satisfaction of anybody (including the Nobel Prize committee) very easily and cheaply, just doing what they do now. All they need to do is pick one of the truly horrible cancers — pancreatic, for instance — and report two-year survival for a reasonably large group, say one hundred patients.
Two-year survival percentages for pancreatic cancer are in the single digits for all patients, regardless of age, medical condition, and treatment (though survival is lower without treatment, of course). Therefore it doesn’t matter whether people are randomly assigned to the CAM treatment or not, or whether there’s a control group, since there are no known variables that would make much difference in survival. The natural history of the disease serves as the control group.
Similarly, there’s no need for any patients to receive a placebo. Placebos may make a difference in how the patient feels about the situation, maybe in the pain or discomfort he suffers, but they make very little difference in survival. If placebos kept cancer patients alive, oncologists would have lots more surviving patients.
Likewise, there’s no need for blinding either side, if the measure is “Patient is alive” vs. “Patient is dead”. Barring fraud, there are very few doubtful cases where one might be inclined to score a patient as alive or dead based on one’s presuppositions about the effectiveness of treatment.
In short, to prove that a CAM cancer treatment works, the practitioner should proceed as follows:
- Line up a hundred or so pancreatic cancer patients.
- Make sure they are properly evaluated by a conventional oncologist, so there can be no question of their diagnosis.
- Document irrevocably their identifying information, so there can be no doubt as to who was in the original group***.
- Treat them according to the CAM protocol.
If, after two years, ninety of the group are still alive (as we would expect given the claims of CAM practitioners), then the treatment undeniably must have worked. There simply are no known confounding factors, biasing influences, or other treatments that could have produced that result.
And if CAM defenders don’t push for this simple demonstration … well, we know what that means.
————-
* I.e., pseudo-scientific.
** Simon Singh got in trouble for using that word …
*** I would put all the info in one big document then encrypt it using public/private keys, burn the encrypted document on CDs, and then ship it off to national health services, insurance companies, individual oncologists, mass media, bloggers, and anyone else I could think of. But then, I think in computer terms. The point of encrypting is that the patients’ private information isn’t revealed to the world. Part of setting up the test is that the patients agree that after two years their info can be revealed. Since they don’t expect to be around in two years, it seems to me that they’d be willing. After two years, you reveal the private key so anyone who wants to can verify that the surviving patients really do represent 90% of the original group.