Terence Hines - Pseudoscience and the Paranormal

Spontaneous remission is another factor that can, incorrectly, lead to the belief that an ineffective therapy is effective. Obviously, not all diseases are fatal. Many diseases will go away even if no specific therapeutic intervention is undertaken. Our immune system and other physiological processes act to protect us against disease and fight off disease once we become ill. This is different from the placebo effect because these are real processes that act in the absence of therapy to restore health. I first learned about spontaneous remission from an episode of that well-known educational television program from the early 1960s, the Beverly Hillbillies . In the relevant episode, Jethro (I think) gets a bad cold. Granny sets out to cure him using one of her usual backwoods potions. She mentions this to their greedy neighbor, the banker Drysdale, who is always looking out to make money off the hillbillies. Granny tells him that her potion always cures colds. Wow! Drysdale gets very excited, convinced he’s stumbled across a wonder cure that will make him a fortune. He gets a business partner, who is president of a large pharmaceutical company, interested and together they plot to get the “secret” formula from Granny. It’s not really a secret—she ends up telling them what’s in it. Somewhat taken aback at the ingredients, they ask her how the potion is given and how she knows it works. Granny happily cackles in reply that you just give the tonic every day for two weeks and at the end of the two weeks, the cold will be gone! Obviously, when an illness goes away on its own, but the patient is engaging in some form of therapy, it is natural—but incorrect—for both the patient and the therapist to credit the therapy with the effect.

A somewhat less well known effect that can lead to unwarranted beliefs in therapeutic effectiveness is a statistical artifact called regression to the mean . The term “mean” means the average, as in adding up a group of numbers and dividing by the number of numbers to get the mean. Regression to the mean is simply the fact that an extreme score (either well above or well below the mean) will tend to be less extreme at some time in the future. Take a simple example of students taking the Scholastic Aptitude Test (SAT). Let us consider only the verbal test in which scores range from 200 to 800, with a mean score of 500. Two students take the test for the first time. Student 1 scores 225. Student 2 scores 775. Six months later both students take the test again. Assume that neither takes any kind of SAT preparation or tutoring courses during this six-month period. What would one expect about their scores the second time around? First, it would be pretty surprising if both students got exactly the same scores the second time, so we can be pretty sure that there will be some change. The question is, in what direction will the changes be? It would certainly be surprising if Student 1 did any worse (unless, I suppose, a safe had fallen on his head in the interim); there is really only room for him to improve. So his score will be, in all likelihood, greater than 225. Probably not much greater, but it will be greater and thus it will be closer to the mean. What about Student 2? He doesn’t have very much room to improve (and we can question his judgment about taking the test a second time), so it is most likely that, on the second testing, his score will go down—again, moving closer to the mean score of 500.

Regression to the mean is a nasty little statistical gremlin that can seriously cloud the interpretation of many scientific studies. But, in the present context, it can also lead to incorrect beliefs about whether or not a treatment works. A good example is the fad of using copper bracelets to treat arthritis pain. The pain in arthritis, like the pain in many other pain-causing disorders, waxes and wanes. Sometimes it’s worse; sometimes it’s better. When is it most likely that one will resort to some sort of treatment for arthritis pain? Obviously, when the pain is greater or above average. People are much more likely to use a copper bracelet when they are in greater pain than when they have little pain at all. Thus, over the next few days, by regression to the mean, the arthritis pain is very likely to reduce, for reasons having nothing to do with the copper bracelet. Of course, the placebo effect is also involved in situations like this, as there is a well-known placebo effect for pain (Watkins and Mayer 1986).

Readers should note that people will selectively take legitimate treatments for arthritis, such as aspirin, when the pain is greatest. Does this mean that aspirin effects are just due to regression to the mean? No. Certainly, part of the aspirin effect is due to regression, and part is due to placebo effects—but properly controlled studies have shown that aspirin does actually reduce pain (for example, Berole and Sethna 2002).

All the factors noted above combine to make it very difficult—I would say impossible—for even the brightest, best-intentioned therapist to determine without the aid of procedures like double-blind, placebo-controlled studies whether a therapy really works. There are other factors as well, and for a fine discussion of them, as well as a lengthier consideration of those mentioned here, Beyerstein (2000) is an excellent source. In any event, the complexity of the task of evaluating therapeutic effectiveness should put the lie to the “alternative” medicine claim that one can simply rely on subjective judgments to do the job. In fact, my own view is that it is highly unethical to rely on such judgments. In any area of therapy, one is dealing with very important—sometimes life-and-death-personal issues. To promote a therapy as being effective without being as certain as one can be that it is effective risks the well-being and maybe even the life of the patient. We will see examples of this in the sections below.

Another characteristic of “alternative” medicine is a heavy use of scientific-sounding terms and language, usually involving “energy fields,” “human auras,” and the like. However, the precise meanings of these terms are never specified. That is, the physical characteristics of the fields (i.e., frequency, amplitude, etc.) are not given. An excellent example is given by Raskin (2000) in his discussion of the nursing theory of Martha Rogers. In Rogers’s theory, “a unitary human being is an irreducible, indivisible energy field and a unitary one…. In fact human beings do not have energy fields; they are energy fields. They are open for exchange and extend to infinity” (p. 32, emphasis in original). Further, “[e]nergy fields are identifiable through dynamic-nonstatic wave patterns and organization that changes from ‘lower frequency, longer wave pattern to high frequency shorter wave pattern’ based on the principle of resonancy.” Sounds pretty scientific, right? From that description, a reader would naturally think that someone, perhaps Rogers herself as the main proponent of the theory, would have gone out and measured the frequencies of these wave patterns. To find out if this was the case, Raskin sent inquiries to several Rogerian nursing theorists and asked simply what the frequencies were and how the frequency measurements had been done. None of the theorists was able to say. In contrast, inquiries to physics teachers about the well-known Millikan experiment showing that charge is quantized were all answered by a description of the famous oil drop experiment. In a directly medicine-related area, it is well known that brain cells operate, in part, through electrical impulses. Even an introductory level text in neuroscience will contain a description of exactly how one measures the frequency (as well as other characteristics) of these impulses.

Andrew Weil, the “alternative” medicine guru mentioned earlier, is right in the mainstream of “alternative” medicine in his use of scientific-sounding jargon. He even invokes quantum physics in the support of his view that the mind can, by itself, cure diseases. Quantum physics (or quantum mechanics ) is a branch of physics dealing with interactions at subatomic levels. To be sure, there are a number of phenomena in quantum physics that are very strange and, to say the least, counterintuitive. This “weirdness” has allowed writers in a number of fringe areas to invoke quantum mechanics as a way of providing an explanation and seeming scientific support for the reality of phenomena they have been otherwise unable to demonstrate. The general argument seems to be that “well, there’s all this really weird stuff going on in quantum mechanics, so why can’t the really weird stuff that I’m pushing (ESP, mind cures for cancer, etc.) be real as well?” The fundamental difference is that the admittedly really weird stuff that goes on at the quantum level goes on in highly predictable and reproducible ways. These phenomena are produced regularly and predicted with high precision every day in physics laboratories around the world. Such is certainly not the case for psychic phenomena and mind cures of cancer! Thus, in the total absence of actual supporting evidence, the existence of weirdness at the quantum level does not, in any way, support the claims of alternative medicine.