Evaluating complementary and alternative medicine: The limits of science and of scientists
Science provides the most important set of tools for the evaluation of complementary and alternative medicine (CAM). Nonetheless, there are important limits in science that constrain its ability to evaluate CAM effectively. Some are the limits encountered by science in conventional medical research.
My assignment for the Ninth Annual Thomas A. Pitts Memorial Lectureship1 was to discuss the limits of science in medicine. Certainly there are limits of science in general, such as the classic difficulty of using science to confidently derive value statements in such domains as ethics and esthetics. But taking a simple dictionary definition of science, such as "the observation, identification, description, experimental investigation, and theoretical explanation of phenomena,"2 it seems unlikely that we can find a priori limits on its usefulness for studying health practices. If we grant that experiment is not always possible at every stage of activities properly called "science," as in some aspects of astronomy, but still require the centrality of prediction as a test, science seems quite well-suited to any empirical question. This is not a call for a special exemption for CAM studies: nonexperimental research that nonetheless utilizes prediction as a test for hypotheses is common in science, e.g., in epidemiology. Whether science can provide complete and satisfactory answers to all such questions is an empirical issue that must await the future; that is, there may well be ultimate limits to the understanding of the universe that science can eventually provide. But that limit is of no particular use in adjudicating arguably scientific issues at present. Nor are most of the limits internal to science, such as Heisenberg's uncertainty principle. Granted that many advocates of CAM like to bring quantum mechanisms into their speculations about some aspects of CAM practice. I prefer to stay with the more basic aspects of the sciences that are obviously relevant to health practices.
THESIS
My presentation was set as a counterpoint to the presentation by Lawrence Schneiderman, M.D., "Alternative Medicine or Alternatives to Medicine."3 In this talk, Dr. Schneiderman vigorously critiqued CAM on the basis of evidence-based science as opposed to what he called "the collective romantic fantasy" of CAM. I will challenge this science-versus-CAM view on the basis of several limits to science. My thesis here is: (1) the basic methods of science are as appropriate to the study of CAM as they are to the study of conventional medicine (that is, there are some limits in both cases, but they are not special exemptions for CAM); however, (2) the human actors who practice science are cultural and emotional beings and the enterprise of science is therefore limited by human capacities. This is not a controversial claim, because human limitations in science are well-recognized. For example, human eyesight and hearing, prior to the development of optical and acoustical instruments, posed important limits on science. Furthermore, modern ethics places a variety of limits on science, such as those imposed by the informed consent process. Consequently, there are many technically feasible studies that simply can-not be performed on humans by civilized scientists. The human limits that will concern me most here are those imposed by the personal and social commitments of scientists and those who depend on science as an enterprise in modern society, and the relationship of those commitments to the rest of society. Thus, this includes both limits internal to science, and others that are external to science. I will sketch these human limits, and then I will return to them with CAM examples.
"Objectivity" or Fairness Limit. "Facts" cannot speak for themselves; all observations require interpretation and more than one interpretation is generally possible. The ideological commitments of scientists introduce bias into interpretation. These tacit biases interact with such explicit biases of interpretation as preferences for quantifiable measures, parsimonious and elegant theories, and so forth. When the tacit commitments are strong, the resulting bias can be overwhelming despite the best intentions of the interpreter.
Resource Limit. All observations have costs and resources are finite.
Autonomy Limits: internal and external. Limits on the autonomy of science constrain the ability of scientists to freely explore phenomena. Commitments to existing theory, usually helpful, serve as an internal limit on the ability of scientists to respond appropriately to novel information. Reliance on outside funding is an external limit, restricting autonomous agenda-setting by scientists. Because the public does not share all the commitments of scientists, the preferences of taxpayers are in tension with the preferences of scientists. Private funding, as by pharmaceutical companies, brings a similar limit. Commercial interests and public interests are often at odds with one another and neither directly represents the interests of scientists and their institutions.
Subjectivity Limit. A preference for observations that can be registered by machines has been productive for science historically, but has limited its investigation of important human issues in medicine, such as pain.
Religion Limit. Much, but not all, of CAM includes an implicit spiritual framework. Some critics have suggested that this is an important source of scientific resistance to CAM. Dr. Arnold Relman, well-known critic of complementary and alternative medicine, made this connection at a conference on CAM held in Philadelphia in November 1999, asserting that "science denies religion and that is what distresses advocates of CAM because CAM," in Dr. Relman's view, "has a spiritual foundation."4 be that as it may, many aspects of CAM are explicitly religious or spiritual in nature, and the reluctance of most scientific researchers to cross that boundary is well-recognized.
These limits suggest the inevitable competition between ideology and the quest for "facts." But my argument is not centered on a postmodern skepticism nor is it a relativistic account of science and rationality. Rather, it rests squarely within the conventional views of science. Until we have fully understood the capacities and limits of conventional modern science, a wonderfully productive enterprise, there is little point in going "beyond science."
APPROACH
To demonstrate my proposition about the human limits of science, I will respond to specific claims and arguments by those who use science to challenge CAM. Foremost, I use as the source for my response an article by Dr. Schneiderman, "Alternative Medicine or Alternatives to Medicine? A Physician's Perspective," which he published in the Cambridge Quarterly of Healthcare Ethics,5 and which he contributed to the Pitts Memorial Lectureship bulkpack. This article presents a vigorous critique of CAM. Dr. Schneiderman published a similar article in the Scientific Review of Alternative Medicine,6 and these articles include the criticisms of the field commonly found in the medical literature and discussed by Dr. Schneiderman in his talk.
At the conference, Dr. Schneiderman and I had a chance to discuss and respond to each other's presentations, and I also had an opportunity to hear and respond to additional challenges to CAM from other participants and audience members. I will draw on these arguments and my responses to them in my discussion here. Doing so will enable me to move from a general discussion of the issues abstracted from the literature and from conversations in medical circles the most common format for discussions for and against CAM - to specific challenges by knowledgeable commentators. Dr. Schneiderman is a nationally known critic of CAM and the Pitts Memorial Lectureship brought together a knowledgeable faculty and audience. The materials and events of the conference are, therefore, appropriate data for an investigation of the human limits of science with regard to CAM.
DEFINITION
First let us be clear about what we are talking. The title of Dr. Schneiderman's talk and bulkpack article suggests a criticism that he and others have repeatedly made about CAM: that it is not medicine at all and therefore it is wrong to "seek medicine's cover" by calling it "complementary and alternative medicine."7 Defining medicine as efforts at "restoring health and alleviating suffering," he says that most activities of CAM lie "outside the range of medicine.... They are alternatives to medicine."8 This is related to the common assertion that "[t]here is no alternative medicine. There is only scientifically proven ... or unproven medicine."9
"Complementary and alternative medicine" is a hybrid term arising in public discourse and combining natural language processes with the efforts of journalists, researchers, and policymakers to craft a semitechnical expression to describe a set of social phenomena. The people and ideas to whom it is applied neither created it nor applied the label to themselves until the usage had become widespread. Many still do not use it, and many actually reject it as a description of their beliefs and practices. While they believe their practices to be relevant somehow to health, that does not necessarily mean that they want these practices to be considered medicine. The reason that "medicine" has consistently recurred in connection with these practices is to be found in the way that the medical profession has succeeded in medicalizing health during the past century. Dr. Schneiderman's assertion that only medicine restores health and alleviates suffering illustrates how radically the territory of medicine has been extended. Fresh air, rest, chicken soup, the comfort of concerned friends, thoughts of Heaven - the list of things that can either restore health or alleviate suffering without a physician's assistance is practically endless. The exaggerated sense of the boundaries of medicine is responsible for the tendency of our language to attach the word "medicine" to all sorts of health-related efforts. If medicine is everything healthful and comforting, then the profession of medicine cannot rightfully claim to own it.
The common definition of CAM, used in David Eisenberg's landmark surveys10 and further articulated by a National Institutes of Health (NIH) panel in 1995,11 refers most basically to health ideas and practices not taught in most medical schools. That is a lexical definition that captures the customary usage of the term. It specifies the only characteristic shared by all CAM practitioners that most people refer to when using the term "complementary and alternative medicine," or its constituents, "alternative medicine" or "complementary medicine." This includes everything from chiropractic to Reiki, traditional Chinese medicine to Indian Ayurveda, "feed a cold and starve a fever" to the use of herbs and foods for health purposes, prayer for healing and visits to shrines, and so on. Descriptions of CAM can show some themes common to many of the ideas and practices covered by this definition, such as a tendency to refer to "energies" not currently known to science, and various forms of vitalism. Such description and analysis are important aspects of the study of CAM and its popularity, but none of these themes is present in all the ideas and practices covered under the definition. CAM is a term of art and is not employed or even known to all whose practices it fits. Therefore, it is not accurate or fair for critics to blame advocates for the various practices that are referred to as CAM for the use of the term. And even for those who do use the term for rhetorical advantage, any claim of cynical manipulation must first address the ambiguity of the term "medicine" noted above.
Given that CAM is not one thing, but many, and that CAM advocates are all sorts of people with diverse cultural backgrounds and education, it is nonsensical to attempt unitary criticisms of CAM. The issues involved in chiropractic treatment of back pain are not the issues involved in my mother's practice of giving me ginger ale for an upset stomach when I was a child. Both practices fit the CAM definition, but one is the practice of a licensed professional and the other is folk medicine. Furthermore, both have been evaluated scientifically and found effective - although my mother never knew that it was the ginger that helped, nor that ginger has outperformed both dimenhydrinate and placebo in the prevention of nausea in double-blind, randomized control trials.12 Neither chiropractors generally nor my mother in particular has opposed scientific medical research on their practices. Some religious people, including medical researchers such as Chibnall and colleagues,13 have opposed controlled scientific studies of prayer - a clear fit with the CAM definition - on religious grounds, but others have carried them out. Opposition to scientific research on CAM is found among both CAM advocates and CAM critics, as is the call for more and better research.
AGREEMENT AND CRITIQUE
I agree with several of Dr. Schneiderman's general points. I consider fraud to be bad, I strongly prefer good science to bad science, and I think that bias distorts research, both in the collection of data and in the interpretation of results. I believe that patients should be well-informed and should not be harmed or exploited. It was on the basis of these principles that in the mid-1960s I began my research on what is now called CAM, and these principles have been the foundation for my teaching about CAM to medical students at Penn State since 1974. So, clearly, Dr. Schneiderman and I have much common ground. Also, in my experience, most advocates for CAM research and practice - whether medical doctors, alternative healers, or patients - share these general principles. But I disagree with Dr. Schneiderman on several crucial points, both implicit and explicit in his remarks at the Pitts Memorial Lectureship and in his publications noted above. The disagreement, often extreme, is about how to draw proper applications of the principles on which we agree.
THE "OBJECTIVITY" OR FAIRNESS LIMIT
"Facts" cannot speak for themselves; all observations require some interpretation and more than one interpretation is usually possible. This is generally recognized within science and there are several explicit biases that valorize certain interpretations, such as the preference for elegant and parsimonious theories founded on quantitative data. The preference for interpretations that are consistent with previously accepted interpretations, "the consistency condition," is another sensible scientific bias in interpretation. These explicit biases have shown themselves generally useful as science has developed, so they have become standard values of scientists. But despite their general usefulness, these values can become limits on scientific autonomy, and will be discussed under that heading.
There are also tacit biases implicit in the culture of science that influence interpretation. These tacit biases, following from the ideological commitments of scientists, constitute an influence strongly conservative of the status quo. This increases the stability of science and retards the occurrence of what Thomas Kuhn calls "scientific revolutions."14 As Kuhn shows, this is ordinarily a good thing. It is not a good thing, however, when science must address observations and claims that are truly novel from the perspective of current scientific theory. There is no need to assert that CAM justifies a scientific revolution to see that some CAM claims are truly novel to contemporary science; that is, they appear to be outside the contemporary scientific paradigm. Since many CAM practices have developed in other cultural settings, this should not be surprising.
Kuhn's description of the sources of resistance to anomalies is a useful framework for considering the tacit commitments of scientists.15 His analysis also brings into focus the observation that much of CAM is not outside the current scientific paradigm. For example, the idea that plants may be useful medicines is not strange in the context of pharmacology. Even the naturopathic idea that whole plants may have benefits that synthetic active principles do not requires no concepts or methods alien to science. Here we encounter a different set of commitments from those Kuhn discusses; for example, commitments to a business model in which science and the patent process are entangled. There is no need to take a relativist view that debunks scientific knowledge to recognize that such commitments influence all kinds of social activity, including science. There is extensive current discussion of such tacit biases in medical research and ways to counter them; for example, the requirement that researchers disclose financial connections relevant to their research projects. The commitments most relevant to influences regarding CAM are more complex and the vested interests less obvious than simple personal financial advantage.
When the tacit commitments are strong, the resulting bias can be overwhelming despite the best intentions of the interpreter. The possible sources of this kind of bias have been extensively discussed in a variety of fora. Critics of CAM frequently offer various kinds of bias, especially financial interest, to account for "unsubstantiated claims" being made with vigor, and CAM supporters have offered plausible theories about the sources of scientific bias against CAM. But more useful than speculating about reasons for the bias is the harder work of showing that the bias clearly exists and exhibits general patterns; that is, that biased interpretations are not simply individual errors, or personal aberrations. That is why clearly instantiated descriptions of scientific bias are centrally important in assessing arguments over what "the evidence" does and does not show.
In offering examples of biased interpretation below, I do not suggest that Dr. Schneiderman intends to mislead the reader. As is usually the case, I believe that the bias revealed is unexamined and unrecognized, and that the resulting interpretations are errors rather than deceptions. That is what makes this issue so important. This problem occurs not despite good intentions, but because of good intentions influenced by unrecognized bias.
Example 1: Advocates of CAM are not influenced by scientific evidence
In his Cambridge Quarterly of Healthcare Ethics article, Dr. Schneiderman states that the empirical evidence for CAM efficacy is not sufficient to explain the enthusiastic support it receives.16 he goes on to question "whether those who are emotionally and financially invested in unconventional treatment will allow their advocacy to be undermined by data."17 This is a frequent charge against CAM advocates and it suggests a kind of external social limit to science: a limited capacity to influence the behavior of the public. This position is historically part of the basis for expert paternalism,18 and it is a serious allegation.
In contrast, I argue that good science does influence CAM utilization, and that for this reason widely used practices require careful study. This argument provides some common ground for those who disagree about CAM to find a research agenda that addresses both public demand and scientific interests. This is important because it provides a consensual rationale for studies that may also yield startling positive results - startling because they were so unexpected on the basis of conventional theories. In essence, this is an effort to expand the limits of scientific practice, and it hinges on the willingness of CAM proponents to be influenced by high quality scientific results and the willingness of conventional scientists to do really excellent studies, where excellence includes accurate representation of the practice being studied. If it were true that CAM providers and users are uninterested in high quality scientific data that do not support their views, my argument would fail.
Dr. Schneiderman supports his skepticism about the openness of CAM proponents to scientific evidence by footnotes referring to Berthold and colleagues' 1998 study in JAMA19 that reports negative results concerning garlic's lipid-lowering properties, and a Los Angeles Times story reporting that a garlic proponent's faith in the plant was not shaken by the JAMA study.20 Does this illustrate the close-mindedness of CAM proponents? The study in question is entitled "Effect of Garlic on Serum Lipid." The study was double-blind, randomized, and placebo-controlled. No fault has been found with the investigators' methodology, and I agree that they showed that the commercial garlic preparation studied was ineffective in lowering serum lipids. But despite the study's rigorous design, numerous researchers, especially in pharmacognosy, criticized the study as fatally flawed. The problem is that the investigators used a steam-distilled garlic oil product, yet generalized their results to all garlic therapy.21
This distillation process yields an extract devoid of aliin, allicin, or diallyl ajoene, the constituents of fresh garlic thought to have a variety of pharmacological actions including lipid reduction. Allicin is a water-soluble thiosulfinate that is converted to oil soluble allyl sulfides when exposed to steam. Further, it is common knowledge among natural food proponents that heating garlic reduces its health-promoting effects and, for the same reason, reduces the plant's characteristic odor. The JAMA study is useful in casting further doubt on the effectiveness of commercial garlic preparations, although not all are steam-distilled. But the study does not support the general conclusion stated by the study's authors.22 Anyone knowledgeable about garlic is justified in considering the study's global dismissal of garlic as a treatment for hyperlipidemia to be unjustified. What would we make of a study that used a conventional drug known to be unstable at room temperature but did , not refrigerate the drug? The broad, unqualified generalization in this study no more justifies the claim that CAM advocates will not "allow their advocacy to be undermined by data," than the study itself justifies its concluding statement about garlic.
The problem in the Berthold and colleagues study is a common kind of error, in part because CAM remedies are often complex and exist in a wide variety of forms. It results not only in wasted effort and resources, but also in growing distrust by the public and CAM practitioners of conventional findings critical of CAM practices. One way of overcoming this problem is through the collaboration of CAM practitioners and medical research scientists. Issues such as the selection of subjects, the precise details of the intervention, and selecting appropriate placebos would all benefit from the involvement of knowledgeable CAM practitioners. Critical reading of proposed designs and other consultative functions, carried out by expert CAM advocates, would provide a strong protection against unintended bias as well as the effects of ignorance, and the result would be greater public confidence in research results.
Example 2: The evidence does not account for CAM advocacy
Dr. Schneiderman tells us that "[o]ne must look for other reasons" than the evidence to support CAM.23 He goes on to state:
carefully conducted reviews of empirical studies reveal no definitive evidence for efficacy in homeopathy, naturopathy, and acupuncture. In addition, after completing a study of therapeutic touch (TT), an intervention practiced by thousands of healthcare professionals, the researchers concluded that "the claims of TT are groundless and that further professional use is unjustified."24
I will now briefly comment on the evidence cited by Dr. Schneiderman for each point.
Homeopathy
The study Dr. Schneiderman cited for homeopathy is Linde and colleagues' meta-analysis of homeopathy trials published in Lancet in 1997.25 This was a well-designed analysis using state-of-the-art techniques to control for study quality and publication bias. The results of this study were consistent with a previous review published in 1991 by Kleijnen and colleagues in the British Medical Journal.26 The Lancet study found 189 randomized and/or double-blind, placebo-controlled trials of homeopathy, 89 of which met strict inclusion criteria. Using the computational method in which an odds ratio larger than 1.00 indicates homeopathic efficacy greater than that of placebo, the combined odds ratio of the 89 studies was 2.45 with a 95 percent confidence interval. Although the reviewers found a lack of evidence supporting clear effectiveness within any single condition (not surprising given the differences between homeopathy and regular medical nosology), their results strongly suggest that homeopathy is not merely the manipulation of placebo effect. In other words, this is a study that supports homeopathy, published in a prestigious, peer reviewed journal.
Naturopathy
The paper cited by Dr. Schneiderman27 is a review of the field of naturopathy, not a "carefully conducted review of empirical studies" of naturopathic practices. It was published in the Scientific Review of Alternative Medicine in 1998. That review focuses primarily on the theoretical dissonance between the naturopathic idea of vis medicatrix naturae, which it characterizes as "magical thinking," and conventional medical views. It grants that in accredited naturopathic schools "[t]he basic-science portion of the required curricula appear acceptable," but goes on to say that the reviewers' "investigations incline us to believe that actual delivery has improved little since the ... Australian Committee of Inquiry issued its findings in 1977."28 The relevance of the Australian findings of two decades ago is not pursued, but the reviewers' "investigations" turn out to be "discussions with graduates of the naturopathic schools."29 No sample size or selection criteria are offered for this sample, nor do the authors give the questions used to rate the knowledge of the graduates. But the authors do go on to say that obviously the basic-science training was poor or else the graduates could not continue to believe in the various naturopathic practices and principles - the "magical thinking" - of the profession. Some of the most damning statements in the review, such as that many naturopathic colleges are "little more than for-profit 'diploma mills,'"30 are supported by reference to a publication that is cited three times,31 but each time with an inadequate footnote that omits source, date, and page information. As a result, I have not been able to track it down.
This essay does not constitute any kind of empirical investigation of naturopathic practice, which includes a very large array of treatments, many widely accepted in conventional medicine. Clearly, this study cannot be used to show the absence of evidence for naturopathy in general.
Acupuncture
Dr. Schneiderman cites two studies to support his assertion that the literature contains no definitive evidence of efficacy for acupuncture. The first study cited, by Vickers and colleagues, was published in 1996.32 The study reviewed all thirty-three controlled trials found in a worldwide search that evaluated the antiemetic effect of stimulation of the P6 acupuncture point. In four trials, the stimulation was performed under anesthesia and did not show an effect. In "27 of the remaining 29 trials acupuncture was statistically superior."33 A second analysis considered only the twelve best quality randomized, placebo-controlled trials. "Eleven of these trials, involving nearly 2000 patients, showed an effect of P6. The reviewed papers showed consistent results across different investigators, different groups of patients, and different forms of acupuncture point stimulation."34 The investigators state that "except when administered under anesthesia, P6 acupuncture point stimulation seems to be an effective antiemetic technique."35 They conclude with an exasperated question for those who will not accept the evidence of such reviews as showing that acupuncture has "specific effects on health":
In short, how could we answer "Can acupuncture have specific effects on health?" other than by a conventional systematic review of what is, after all, conventional research? How much evidence would constitute "enough" evidence?36
This seems an odd citation to support the claim that "carefully conducted reviews of empirical studies reveal no definitive evidence for efficacy" in acupuncture! Dr. Schneiderman's other citation is a 1990 study, which simply found that, at that time, there were no high quality clinical studies of acupuncture for chronic pain.37
The enormous volume of acupuncture research is such that there is little point in pursuing further the question of whether there is adequate evidence that it "does something" beyond the placebo effect to which Dr. Schneiderman attributes its activity. The NIH consensus conference in 1997 is cited by advocates as evidence for efficacy and by opponents as evidence of NIH complicity in CAM fraud.38 Acupuncture is a fine example of the frustrations involved in attempting to find "just the facts" using research methods and concepts from one culture to study the practices of a very different culture.
Therapeutic touch
The quotation above by Dr. Schneiderman dismissing therapeutic touch as groundless and unjustifiable in professional practice appeared in the 1998 JAMA paper written by Emily Rosa, a fourth grader who conducted the study as a science fair project, and her mother and stepfather, both members of a debunking organization (Questionable Nurse Practices Task Force, National Council Against Health Fraud), and Stephen Barrett, the publisher of the website Quackwatch.39 This paper is a remarkable event in the history of JAMA. I will offer just a few comments.
The design used by young Emily tested the ability of self-identified therapeutic touch practitioners to detect the presence of Emily's hand above one of their own hands, with the therapeutic touch participant's view of the hands blocked. The participants apparently thought they could do this, although the ability does not properly represent any specific facet of actual therapeutic touch practice. The study did not address in any way the ability of the practitioners to reduce pain or anxiety, the main claims of therapeutic touch.
The extensive literature review in the paper includes studies with interesting positive results according to the reviewers themselves. For example, the University of Alabama study of therapeutic touch for burn patients, controlled with the use of sham therapeutic touch, yielded "statistically significant results" in patient reports although it did not result in the use of less pain medication.40 The reviewers treat this as a negative finding but such a view seems unjustified given the severity of burn pain and the typical use of therapeutic touch as an adjunct to conventional treatment.41
The study by Emily and colleagues also presents serious ethical problems in that the authors state that their success in getting the cooperation of so many therapeutic touch proponents probably hinged on the presentation of the study as merely a fourth grade science fair project. This is not just failure of full disclosure; it is actual deceit. Finally, one cannot help but note the strong negative bias of the investigators conducting research in an unrefereed setting. Research by CAM proponents is often treated with skepticism because of the investigators' commitments. How would we react to a positive study of therapeutic touch carried out as a science fair project by the granddaughter of Delores Krieger, the creator of the practice? Would JAMA publish it?
Lorenzo's Oil
In his Cambridge Quarterly of Healthcare Ethics article, Dr. Schneiderman discusses the movie Lorenzo's Oil,42 the film in which the parents of a boy named Lorenzo, who has adrenoleukodystrophy (ALD), are involved in the development of a novel treatment with a special oil containing the monounsaturated fatty acids oleic and erucic acid. In the film, the parents are portrayed as pushing the medical estab-lishment and creating a new treatment for this devastating disease. Dr. Schneiderman says:
Unfortunately, the movie was a fraud. Even by the time the movie had been made, it was clear that the oil produced no such miracle. Worse than being merely useless, it was toxic as well.43
This assertion was supported with two footnotes, and I initially gave this point to Dr. Schneiderman-without looking at the sources cited. But in October of last year, a story appeared on the front page of USA Today, entitled "Lorenzo's Oil Brings Hope for the Afflicted: New Findings Support Unorthodox Treatment of Rare Brain Disorder."44 In the story, Hugo W. Moser, M.D., the physician who cared for Lorenzo during the time that the oil was developed, was interviewed on the occasion of his presenting the findings of a study showing that Lorenzo's Oil delays the onset of symptoms in boys with ALD, if it is started while the patients are asymptomatic. Dr. Moser was quoted as saying:
I think if I had a son (whose ALD symptoms had not yet appeared), I would put him on it [Lorenzo's Oil]. Things have been publicized as treatments with much less evidence.45
This, naturally, led me back to Dr. Schneiderman's footnotes. The publication intended to support the statement that the movie was "fraudulent" was entitled "Suspended Judgment: Reactions to the Motion Picture 'Lorenzo's Oil,'" published in 1994 in Controlled Clinical Trials, by Dr. Moser.46 In the article, Dr. Moser criticizes some inaccuracies in the film's account of events and says that the oil, on the basis of initial clinical trials, "is not nearly as effective as portrayed in the movie.... results in patients who already have symptoms have been disappointing."47 Dr. Moser expresses doubt that Lorenzo's "long survival" is attributable to the oil. (Eight years later, as the newspaper article documents, Lorenzo is still alive although severely incapacitated. His father attributes his survival and ability to communicate minimally to the use of the oil named after him.) Dr. Moser closed this article stating that he expects that with time the public and the medical profession "will achieve an appropriately balanced view of the therapy of ALD and the role of Lorenzo's oil."48 He called for caution and "suspended judgment"49 of a treatment that eight years later, based on his own clinical trials, he now says he would use for his own family. This is not a statement that Lorenzo's oil was "useless," nor an assertion of fraud.
The other reference, apparently to support that the oil is "toxic," is to a letter to the editor of the New England Journal of Medicine by Dr. Moser.50 In the letter, he reports that some patients receiving Lorenzo's Oil in a clinical trial developed a reduction in their platelet count. He says that "there was no clinically important bleeding" secondary to this thrombocytopenia, and platelet counts returned to normal when erucic acid was removed from the diet.51
Had the newspaper story not caused me to go back to these citations with a critical eye, I would have placed Lorenzo's Oil in the fraudulent-and-dangerous column. Now I have it in the wonderful-contribution-to-the-treatment-of-a-devastating-disease column.
I do not argue that CAM in general is supported by an abundance of high quality scientific evidence. But it is not only evidence that is lacking, it is also studies. Given the history of medical research funding and the recency of new funding sources for CAM, this is hardly surprising. Lack of evidence for effectiveness, in the absence of substantial, well-supported research, is not evidence of lack of effectiveness. And when critics of CAM state that the evidence runs against CAM, and then support it with references to weak studies with negative findings, to reviews showing a lack of studies, and to systematic reviews that actually support CAM efficacy, it is hard not to see that as prejudice against CAM. Strong negative prejudice can easily lead the scholar to feel disinclined to do thorough and time-consuming research. If you are certain that the answer is negative before you start, there is little motivation to be scrupulous. It seems clear that conventional scientific bias is as harmful to serious criticism as it is to support.
Example 3: The anti-CAM literature
I have said that Dr. Schneiderman is not alone in the approach I have criticized, and that these issues represent a systematic bias among the strongest critics of CAM. Dr. Schneiderman's recommendations of reliable information sources on CAM are useful in supporting my contention. Those that he says are "the best currently available sources for gaining accurate information about alternative medicine" are the website www.quackwatch.com and "publications like the Scientific Review of Alternative Medicine and Alternative Medicine Alert."52 I have often found Alternative Medicine Alert useful. The other two sources I primarily turn to in order to find further examples of systematic bias.
For example, in 2002, Joel M. Kauffman, of the Department of Chemistry and Biochemistry at the University of the Sciences in Philadelphia, published a website review of Quackwatch.53 For the review, Kauffman, a self-proclaimed member of a "local skeptics group," used eight webpages on topics with which he was familiar, and these he "examined minutely.... to make generalizations about the website."54 The topics were "Tips for Lowering Your Dietary Fat Content," "Low Carbohydrate Diets," "Chelation Therapy," "Glucosamine for Arthritis," "Magnet Therapy," "Homeopathy," "Dietary Supplements," and "Stanislaw Burzynski and 'Antineoplastons.'" Kauffman says that "[a]ll eight pages from www.quackwatch.com ... were found to be contaminated with incomplete data, obsolete data, technical errors, unsupported opinions, and/or innuendo.... Hostility to all alternatives was expected and observed from the website, but not repetition of groundless slogans from mainstream medicine...."55 A good example was the website's use of a Danish study of chelation in which the investigators used a solution different from that used in chelation therapy - a dietary supplement including iron, the chelating properties of which "guaranteed a lesser effect" - and used a sample including 70 percent smokers "despite the fact that it has been shown that smoking will neutralize the effect of chelation."56 When the study was investigated by the Danish Medical Society's Committee on Investigation into Scientific Dishonesty, it was found that "the double-blinding was broken," and that the investigators falsely claimed to be using the correct solution.57 Additional flaws were cited from a variety of peer reviewed publications.58
The Scientific Review of Alternative Medicine provides a similar opportunity to see the result of "hostility to all alternatives." It was this journal that Dr. Schneiderman cited concerning the waste of time involved in studying acupuncture. The article was a repudiation of the 1997 NIH consensus panel on acupuncture.59 The Scientific Review is published by Prometheus Books, another good source of such examples. The problem is not limited to a few critics or sources. The problems that one finds in the most adamant critics are simply more obvious. Similar bias is frequently visible in most mainstream publications.
It is not surprising that poor arguments do not influence CAM practice. Worse, consistent and predictable bad arguments reduce the credibility of medical research in the eyes of CAM proponents. Yet, when rigorously designed studies yield negative results, they have led to changes in practice. For example, Moertel and colleagues' excellent study of laetrile, published in JAMA in 1981,60 marked the end of the widespread use of laetrile as a stand-alone cancer treatment by most CAM proponents.
THE RESOURCE LIMIT
Despite modern society's vast investment in scientific research, resources are finite. Because resources are limited, so are relevant scientific observations. Here we encounter the problem of induction; that is, the need to justify general statements when all pertinent observations have not been or cannot be - made. For instance, the justification of sampling is a scientific practice required by the problem of induction. Treatments work better for some patients than others. Side-effects and other risks are encountered by some, but not all patients. When a treatment is pronounced effective or safe for a class of patients - people with diabetes, for example - the treatment has not been and could not be observed in all people with diabetes. For a study to validly state that a treatment is safe and effective requires the use of statistics, which permits an assessment of the probability that the observations are generalizable with confidence. In biomedical research, even the relatively "simple" evaluations required for a new drug require large numbers of subjects to achieve an acceptable level of probability and confidence, making such studies very expensive. Therefore, problems and observations must be prioritized, and deciding which questions to ask and which possible answers to entertain as hypotheses crucially determines what can and will be learned. CAM suffers a variety of disadvantages in this prioritizing process that limit the ability of science to produce confident assessments of its efficacy and safety.
The selection of questions carried out within conventional science and its institutions follows from current theory, which itself is somewhat limited by the personal investments of scientists. It is not just resources in terms of tax dollars that are at stake. Rather, it is the resources of particular schools, laboratories, investigators, and businesses, and this produces strong motivations to protect those particular resources. It is only natural that scientists proceed in a logical fashion consistent with their past findings. "Breakthroughs" can make careers, but this is usually true only when they remain within the bounds of the contemporary paradigm. But when a well-established scientific paradigm encounters radically different claims, the conservative trend reduces flexibility. All the rules of academic medicine, from peer review through tenure and promotion criteria, and the operating procedures of NIH, generally guarantee that radically unconventional topics receive little attention and that when studies are done they are rarely published. Unconventional interests - let alone actual research - have such a strong antitenure effect that most of us know at least several research scientists who have strong interests in CAM but never mention those interests within their medical school. Access to resources governs careers, and the resource limit is among the most important factors that produce a strong political/economic incentive against embracing exotic research topics.
In his Cambridge Quarterly of Healthcare Ethics article, Dr. Schneiderman invokes the resource limit under the heading of "Justice."61 Noting that "our country is concerned about the high cost of medical care, the limits of medical resources, and access to these resources....[, p]hysicians must act in the role of stewards in order to maximize benefits to patients and protect the limited resources of society."62 Suggesting that if, as he believes, most complementary and alternative medicine achieves any apparent effects through a long list of
"well-known psychological mechanisms," then we have wasted a lot of time and effort.... The time has been wasted on all the people who have spent years learning falsehoods about acupuncture points and principles of homeopathy.63
Dr. Schneiderman's source for this comment is an editorial by Wallace Sampson, published in the Scientific Review of Alternative Medicine, in which Dr. Sampson accuses the 1997 NIH consensus panel on acupuncture of a "lack of critical thinking," "social advocacy," financial conflict of interest, and "pseudoscientific thinking."64
During the Pitts Memorial Lectureship, the rhetorical use of this resource limit against CAM was further illustrated when an audience member suggested, as an example of private funding possibilities, that the copper industry be approached to finance a study of the use of copper bracelets to prevent or treat arthritis. His point was not advocacy for the practice, but was based on the observation that a great many people use copper bracelets this way, so a study would be valuable and should not be hard to design. Another conference participant responded that such a study would be a foolish waste of money, since copper bracelets don't work. If the copper industry wants to do something for health, he said, they should contribute to a childhood immunization program or something else that does work.
As these two examples show, the resource limit argument works in concert with the "they won't pay attention to data" claim, and the assumption that some CAM can be accurately assessed without any study at all.
THE AUTONOMY LIMITS
The resource limit leads to a second kind of limit, the limit on autonomy in establishing the research agenda. Granted, scientists have unique knowledge that entitles them to special consideration in determining the direction of research. However, factors both internal and external to contemporary science restrict their autonomy.
Internal limits on scientific autonomy
Internally, the social structure of science gives current theories disproportionate influence, especially in the face of anomalous observational claims as found in CAM. Of course there is room for argument in how much influence is "disproportionate." Thomas Kuhn, as noted above, considered the conservative tendency of science a good thing during "normal science," and the natural inclination to accept current theories is an important aspect of that conservative bent. The clearest example of the disproportionate influence of conventional theory with regard to CAM is found when those theories prevent serious inquiry into widespread observational claims. It was one thing for Western scientists to be reluctant to accept the theories of traditional Chinese medicine rooted in Qi and yin and yang, concepts foreign to modern Western science. It was another for this reluctance to result in almost a century of unwillingness to look seriously at the observational claims of millions of physicians, acupuncturists, and patients in Asia and many other parts of the world.
Because of the scarce resource issue, it is reasonable for current theory to enter into the prioritization of observations. But intransigence in the face of widespread anomalous observations, often from reputable sources, marks the influence of current theory as disproportionate. In this case it prevents even the possibility of science-based rejection, producing instead mere dogmatism. In such situations the theories resisted are typically based on the anomalous observations. Rejecting the observations and the theories separately and without serious argument prevents either from being seen in their most favorable light.
This is what Paul Feyerabend calls "the consistency condition," arguing that it is "unreasonable because it preserves the older theory, not the better theory.... It eliminates a theory or a hypothesis not because it disagrees with the facts; it eliminates it because it disagrees with another theory."65 This limiting effect of existing theory constrains scientific autonomy by restricting hypothetical options. This influence is a kind of bias, but in many circumstances it is a justifiable bias like the bias in favor of rigorous methods and objective evidence. But in situations of controversy, this bias inevitably gives rise to claims of unfairness.
This appearance of prejudice is also one reason for some of the external curbs that have been placed on the autonomy of scientific medicine, such as in 1992 when Congress mandated what is now known as the National Center for Complementary and Alternative Medicine (NCCAM) at NIH. The scientific community did not seek such an office. The normal approach of science assumed that if observations and practices in CAM had merit, they would rise to the top through the same processes by which other scientific ideas come to be studied. But many voters thought that the unconventional treatments they believed to be warranted by their experience were not getting a fair hearing in the conventional scientific arena. The congressional mandate that established NCCAM represented the intrusion of the lay public and legislators without scientific training into the scientific peer-review controlled process of prioritizing research. Like the earlier congressionally mandated study of alternative cancer treatments by Congress's own Office of Technology Assessment,66 society outside of science was asserting its right to influence scientific decision-making. Whether one likes this intrusion or not, and regardless of one's opinion of the new center at NIH, this is clearly an external curb on scientific autonomy.
Dr. Schneiderman, in his Cambridge Quarterly of Healthcare Ethics article, states that CAM is largely theory-driven, whereas regular medicine rejects dogmatic theories and simply follows where the empirical data lead.67 This is true in normal science, to a considerable extent. However, when paradigms clash, the situation changes. At these times, the scientific paradigm itself becomes a limit on the autonomy of scientists. The following example illustrates this point.
On November 10, 1999, at a conference on CAM held in Philadelphia, Dr. Marcia Angell, then editor-in-chief of the New England Journal of Medicine, participated in a panel that addressed questions of editorial bias against CAM. Publication bias would be, of course, a major barrier to the development of CAM research.68 Dr. Angell insisted that there was no bias against CAM studies, and she claimed that she would gladly publish any good scientific study on the subject that came to her. She was then asked about her own rejection of the meta-analysis of homeopathy recently published in Lancet.69 After all, it was argued, the methodology of that study had been widely praised. Dr. Angell responded that to be good science a study must offer a "plausible biological mechanism" for the effects reported. Otherwise, the study would not be believable. She said that therapeutic touch, homeopathy, moxibustion, and intercessory prayer are examples of practices that are "preposterous" and "impossible" because they lack a plausible biological mechanism. Additionally she claimed that studies of these practices are only being published for social and political reasons. These remarks extended those that she has made in print. Similar comments were made by some of the other editors on the panel.
Dr. Angell is to be complimented for stating this conventional position with great clarity and force. This view hinges on what I call the theoretical plausibility criterion. Some have attempted to express this criterion in Bayesian terms as prior probability. However, prior probability is generally established empirically rather than theoretically, and the use of a prior of zero, as Edmond Murphy has pointed out, is an abuse of Bayesian reasoning.70 In deciding whether a report is accurate or false, a critic using the prior probability argument:
evaluates the present evidence not (as he should) on its merits but on some multiple of the frequency with which it has previously been reported.... If his prior is zero he will never accept any evidence that A occurs.71
This is a spurious approach in the conventional terrain because it terminates investigation rather than rationally ordering it. The problem is compounded in CAM research where "the frequency with which the event has previously been reported" is one of the most contested issues. If we are prepared to say that intercessory prayers have never produced physical healing or that infinitesimal homeopathic dilutions have never produced a biological effect, it would seem reasonable to say that there is no likelihood that they will start doing so now. There is no point in testing the hypothesis that they have these effects. But such effects actually have been reported frequently. Whether these reports are true is part of the issue under investigation. It is therefore more accurate to discuss this issue in terms of theoretical plausibility than prior probability.
The theoretical plausibility criterion asserts that (1) all valid knowledge will prove to be coherent (that is, to follow logically without inconsistencies or gaps) with some characteristic of established contemporary science (known biological mechanisms in Dr. Angell's instance), and (2) the likelihood that a claim will eventually have this coherent relation to contemporary science can be judged on the basis of present knowledge. Both halves of this assertion are problematic. The assumption of eventual coherence is itself a theory-based prediction that may or may not prove true, and even if it does prove true, that does not mean that present knowledge is sufficient to judge which claims may ultimately meet this test. These ideas support expert paternalism and suggest that a process of free inquiry open to diverse views is unnecessary and counterproductive in science, except within narrow bounds internal to conventional science. It also suggests that the patient's autonomous right to refuse conventional treatment and to use legal alternatives is merely the right to be wrong. This is a view of science as radically transcending cultural and political processes so that the ethical and political mechanisms of a free society are, at best, a distraction and, at worst, a destructive force when they intrude on science. This view has profound ethical as well as epistemological implications, and these are especially serious with regard to medical science. My disagreement with the view is not based in a relativizing view of science. Rather, it follows from a liberal view of the central importance of free speech in discovering truth and error.
In a discussion following my presentation, another conference participant suggested that Dr. Angell might well be right in saying that one should not bother to pursue biologically implausible CAM theories. To support his point, he referred to the publication in Nature, in 1988,72 of a laboratory study that supported the biological effect of infinitesimal dilutions - a central tenet of homeopathy. He said that when Nature sent a panel of experts to investigate the lab, following the publication of the article, they found fraud. Therefore, the article should not have been published. This is the kind of wasted time that Dr. Angell's advice could avoid.
In fact, though, this example demonstrates just the opposite. The "experts" who visited the lab of Dr. Jacques Benveniste consisted of a physicist-turned-journalist (John Maddox, then editor of Nature); James "the Amazing" Randi, a stage magician and member of CSICOP (Committee for the Scientific Investigation of Claims of the Paranormal), a debunking organization associated with Prometheus Books noted above; and a mycologist from NIH who investigated statistical evidence of fraud as an avocation. None of the "investigators" had expertise in the biochemical science being conducted in the laboratory. After their investigation, they published a report in Nature stating that they did not find any evidence of fraud, but that they were convinced the lab staff was collectively deluded in their interpretation of results.73 The results reported as observed by the investiga-tors were equivocal and did not contradict the findings reported in the published study.
I noted these points in response to the suggestion that Dr. Angell's approach was reasonable. The reaction of the skeptic who had raised this example, following my account of the episode, suggested that I was quibbling over technicalities by arguing about whether the investigators were experts and the result could be called fraud. In contrast, I consider those two issues central to the skeptic's argument and the overall validity of the conventional criticisms leveled at Benveniste and his team. It is precisely the fact that such an inquiry has been accepted as a devastating critique that concerns me!
The use of this episode, especially its notably nonscientific investigative and editorial component, to support refusal to seriously investigate "impossible" scientific claims actually demonstrates the reverse. It is the intransigent refusal to seriously investigate such claims scientifically, arising from dogmatic insistence on the prior plausibility criterion, that hampers the production of scientific knowledge. The use of sensational journalistic techniques to stigmatize researchers who report unacceptable results further degrades the credibility of science in the eyes of the public.
External limits on scientific autonomy
Contemporary science in the United States has, especially since the Second World War, become enormously dependent on public funds and commercial investment. This dependency creates a vulnerability to political and commercial influence in decision-making by scientists. Scientists often resent such influences from nonscientists as unwarranted intrusions. Some would grant that the setting of priorities, such as recent increases in attention to women's health in medical research, could be appropriately influenced by nonscientists because of funding issues. But decisions about the science used to serve those priorities seem different. However, that difference is vitiated by the difficulty in entirely separating priorities from science per se and the existing commercial influences that guide scientific questions.
For example, it has been clear for a long time that one thing holding back the scientific evaluation of herbal medicines is the inability to patent plants. Pharmaceutical investment must be focused on either patentable procedures for extracting active ingredients or on synthesizing those ingredients. The fiduciary responsibility of pharmaceutical executives to their shareholders requires that this consideration determine the direction of research. The scientific questions concerning natural products, including the possibility that synergistic effects in botanicals produce treatment advantages, get short shrift as a result. In this instance, outside influences have worked against the study and acceptance of a group of CAM practices.
In other cases, external influence on medical research has served to eventually bring productive attention to CAM practices. Several examples can be seen in the natural healing domain. The use of whole grains and the avoidance of refined carbohydrates, the breastfeeding of infants, the use of cranberry juice to prevent and treat urinary tract infections, and the use of live culture yogurt to treat vaginal candidiasis have all been important CAM practices (in naturopathy and the health food movement) for many years. Sylvester Graham began the tradition of advocating for whole grains against refined flour in 1838.74 It is often pointed out that many of Graham's theories were nonsensical. True, but beside the point. Most regular medical theories of that time now look nonsensical. The advocacy for whole grains picked up momentum with the development of Seventh Day Adventist health reforms in the second half of the nineteenth century,75 adding religion to the influences coming to bear on medical research. In the twentieth century, whole grains and a general high fiber diet were the staples of the health food movement, so much so that advocates were widely ridiculed in the medical press for their preoccupation with bowel movements. Not until the mid-1970s, with Denis Burkitt and colleagues' epidemiological observation of the effect of high fiber diets in Africa, did medicine begin to take serious notice.76 Now recent studies such as Salmeron and colleagues on the role of fiber and glycemic load in adult-onset diabetes risk77 have brought conventional dietary recommendations into line with CAM claims that are more than a century old.
Women have been using cranberry juice to prevent and treat urinary tract infections for generations. Not until the 1990s did medical research provide the first randomized, placebo-controlled trial.78 That trial and subsequent studies, including one reported in JAMA this past June, support the CAM practice.79 This use of cranberry juice is now actually in the Merck Manual.80 The use of yogurt has a similar history.81
Breastfeeding was greatly reduced during the twentieth century, a change supported by the advice of physicians. By the 1960s, active medical efforts against breastfeeding were common in American hospitals. The La Leche League, formed in 1958 by a small group of mothers seeking support in breastfeeding their babies, became a well-known symbol of the far out fringe mentality found in health food circles. Not until 1992 did the American Academy of Pediatrics endorse breastfeeding as the best nutrition for American infants. But a national survey of physicians' attitudes and knowledge regarding breastfeeding, published in JAMA in 1995,82 revealed some startling facts. Using a national sample of over 5,000 pediatricians, ob/gyns, and family doctors, the survey found (1) an overwhelming assertion of support for breastfeeding, and (2) that between 40 percent and 60 percent of each group gave factually wrong answers to questions about common problems in breastfeeding - where the wrong answer was "stop breastfeeding."
Breastfeeding, along with the midwifery movement, provides an excellent example of another point where I agree with Dr. Schneiderman. Much of what we call CAM and what fits the NIH definition is not medicine. In fact, part of the impetus for the CAM movement is the medicalization of health in the twentieth century - the tendency to assume that if there is evidence that a practice has health benefits, it must become a part of medicine. There are good arguments that breastfeeding, for example, is better supported by experienced lactation consultants and that the involvement of physicians tends to have a detrimental effect - except for those doctors who have breastfed their own babies. That was another finding of the JAMA study,83 and it underlines another crucial issue - experience has value! The difficulty of using scientific knowledge about breastfeeding to create support that actually helps women breastfeed successfully, and the ancient success of breastfeeding women to inform and help each other, suggest another limit of science. Perhaps in some areas of CAM, it is not a matter of external intrusions on medical science, but rather the excessive influence of medical science in areas of life not properly described as primarily medical. This is not to say that science is unhelpful in this or any domain, but that science is not enough!
Similar points could be made about the use of vitamins and antioxidants. Granted the exaggerated commercial hype and the importance of a knowledgeable approach to dietary planning, it is clear that CAM, especially natural healing and health food ideas, has found scientific support and that the research would not have been performed if not for pressure from consumers. Obviously some nonscience influence is inappropriate and retards science, but some is actually appropriate and helpful. Locating that limit requires the collaboration of scientific and nonscience voices, and we should expect the results to be contentious.
THE SUBJECTIVITY LIMIT
Modern science has a strong preference for objective evidence - that which is "publicly observable." Such evidence has several advantages, the most central of which is that the reliability (the degree of measurement stability across observation points) of public observations can be independently assessed, whereas epistemically private observations (e.g., pain) cannot, and their replicability (the ability to predictably reproduce a phenomenon) can be similarly assessed regarding whether what is reproduced is actually the same phenomenon. The simplest way to confirm that an observation is reliably "public" is to show that it can be registered by a machine, removing the human observer from the loop altogether (almost!). Medicine has moved rapidly in this direction throughout the twentieth century, and this has led to several problems, including excessive reliance on lab tests, lack of attention to history taking and physical diagnosis, and the poor record of American medicine in controlling pain - an irretrievably subjective but indubitable phenomenon.
In CAM, there are many claims that can most easily be tested using subjective report, and if vitalist theories in CAM are correct, these are some of the most crucial and distinctive observations. Subjective report can be dealt with in a rigorous scientific manner, as in the use of visual analog scales to report pain and the use of blinded designs. Nonetheless, limits exist beyond which it seems implausible to call the use of subjective report scientific. At this point, other kinds of systematic rational techniques supplement science. Qualitative methods, phenomenology, ethnography - it should not be necessary to call these "scientific" in order to acknowledge their usefulness, but if they are useful and not science (we should hesitate to call everything that is not science "nonscientific"!), then here we find a limit to science. This limit is related to my final example, the religion limit.
THE RELIGION LIMIT
Since the Enlightenment, there has been an uneasy peace between science and religion. This truce basically follows David Hume's insistence that science operates by rationality and observation, whereas religion operates by faith (understood as inspired belief without rational grounds). In modern science, this distinction has often been taken to mean that religious questions are out of bounds for science and scientific questions are out of bounds for religion. This fits a particular Western view of religion that is willing to exclude "spiritual causality" or else, in the case of the miracles recounted in the Christian Gospels, to relegate such spiritual eruptions to the distant past. However, even as modern theology has sought to operate within that limit, it turns out that most Americans, including the well-educated, do not. There is very widespread belief that prayer and other spiritual activities, both within and outside the framework of formal religion, have effects, and that sometimes those effects are not produced by material means.84 That is, a great many Americans today make empirical claims about their spiritual activities that are anomalous within the materialistic view that had been assumed appropriate to science, even by many religious scientists.
There have been many scientific studies of the health effects of being religious.85 Usually these have been epidemiological in design. Although these have their critics,86 they are not obviously outside the bounds of science on methodological grounds. Some studies of intercessory prayer published in well-known medical journals have shown that scientific methods can be adapted to some questions that appear to be intrinsically religious.87
So what precisely are the limits on scientific inquiry into religious and spiritual issues? It seems clear that there must be such limits, but it is not clear where they are. Perhaps the most obvious involves the question of whether prayer effects are, as most Americans who pray seem to believe, the effect of Divine intervention. Randomized controlled trials of intercessory prayer might seem to require that God provide only "usual Divine support" to the control group. However, most investigators studying intercessory prayer do not postulate Divine intervention. Another theological objection is that such studies "test God," and that to do so is wrong. Whether this is a sound objection depends on one's theology, not one's science.
Current debate over this issue is intense, from both the religious and the scientific side. The fact that CAM practices have a strong predilection for implicit spirituality, and that conventional medicine is increasingly open to the importance of spirituality within medical care, makes this issue directly pertinent to the limits of science in both CAM and in medicine generally. Here again there is good reason to say that, if spiritual and religious practices are healthful, it does not necessarily mean that they must become part of medicine. Medicine as well as science has limits.
MEDICINE, THE SCIENCE AND THE ART
The title of Dr. Schneiderman's and my session at the conference was called "Medicine, the Science and the Art." Coupled with my assigned title, "The Limits of Science in Medicine," the panel title suggests a fundamental assumption that what is valid in medicine but is not science is best characterized as "art." I consider this a false dichotomy, though a very commonly asserted one, as in "arts and sciences" in our universities. Art may mean "the arts" in the esthetic sense, or technique, as in the "art of building." There are CAM practices that involve the arts, including painting, music, and poetry. These are intriguing, and rehabilitation medicine has employed arts and crafts in its practice. There is also the mastery of technique required to some extent in all of science and medicine, in which the more deftly an instrument is used (whether a musical or a scientific instrument), the more artful the performance. Surgery is an obvious example where art in this sense is of great importance. Art can also be extended to doctor-patient communication, another kind of mastery of technique. But these examples do not exactly suggest limits of science, even though they are not science in themselves. Does the art of the violin indicate a limit of composition?
The more important - the truly central - limits of science are those that involve the terrain on which science itself operates, the production of knowledge. I have argued that such limits do exist. But this does not necessarily indicate the limits of valid knowledge, because not all of valid knowledge is produced scientifically unless we define science as what rationality and knowledge are. This is an outmoded notion left over from defunct logical positivism, having no more practical utility than its postmodern opposite. If we accept that inflated meaning for science, we either lose the core meaning of the term that is rooted in the modern development of the natural sciences, or we are forced to demote whole domains of important knowledge in such disciplines as history, philosophy, and anthropology, together with the products of life experience, to mere opinion. Logical positivism attempted, unsuccessfully, to equate valid knowledge with science, and the postmodern turn has attempted to demote science to one of many equally valid (or invalid) points of view. Both extremes harm rationality in general and science in particular. The inflation of science, produced by its great cultural value and the desire of other knowledge producers to capture some of its luster, is disastrous to both science and its admirers.
CONCLUSION
Science is necessary to the evaluation of both "regular medicine" and CAM. But it is not sufficient for either. Similarly, the inflation of medicine to mean all valid health knowledge and practice, though an understandable product of its success, is disastrous to medicine. As Donald Seldin complained in his 1981 presidential address to the American Association of Physicians, the medicalization of health makes medicine responsible for "the realization of happiness, inner tranquility, moral nobility, [and] good citizenship."88 The proper limits of medicine are as crucial to evaluating CAM as are the limits of science within medicine. If the passion and good intentions of CAM's most vocal critics would recognize and deal with these limits - not eliminating bias (an impossibility), but acknowledging and controlling it - their mission to help the common good would be better served.
SIDEBARJournal of Law, Medicine & Ethics, 31 (2003): 198-212.
(C) 2003 by the American Society of Law, Medicine & Ethics.
SIDEBARScience is a necessary tool for the evaluation of complementary and alternative medicine, yet the ability of science to evaluate effectively is constrained. The primary relevant limits are not inherent within the methods of science but rather lie within the culture of science, the particular ways that scientific knowledge, theory, and method are configured and arrayed rhetorically, and in the social context where science operations.
REFERENCEREFERENCES
1. D.J. Hufford, The Limits of Science in Medicine, presentation at Alternative Medical Systems: Here to Stay, But on What Terms?, Ninth Annual Thomas A. Pitts Memorial Lectureship, Medical University of South Carolina, Charleston, September 13, 2002.
2. R.B. Costello, ed., American Heritage Dictionary of the English Language, 3rd ed. (Boston: Houghton Mifflin Co., 1997): at 1221.
3. L.J. Schneiderman, Alternative Medicine or Alternatives to Medicine, presentation at Alternative Medical Systems: Here to Stay, Eut on What Terms?, Ninth Annual Thomas A. Pitts Memorial Lectureship, Medical University of South Carolina, Charleston, September 13, 2002.
4. D.J. Hufford, "Complementary & Alternative Medicine and Cultural Diversity: Ethics And Epistemology Converge," in D. Callahan, ed., The Role of Complementary and Alternative Medicine: Accommodating Pluralism (Washington, D.C.: Georgetown University Press, 2002): 15-35, at 24 (quoting Arnold Relman).
5. L.J. Schneiderman, "Alternative Medicine or Alternatives to Medicine? A Physician's Perspective," Cambridge Quarterly of Healthcare Ethics, 9 (2000): 83-97.
6. L.J. Schneiderman, "Medical Ethics and Alternative Medicine," Scientific Review of Alternative Medicine, 2, no. 1 (1998): 63-66.
7. See Schneiderman, supra note 5, at 95.
8. See id. at 87.
9. P.B. Fontanarosa and G.D. Lundberg, "Alternative Medicine Meets Science," JAMA, 280 (1998): 1618-19, at 1618.
10. D.M. Eisenberg et al., "Unconventional Medicine in the United States: Prevalence, Costs, and Patterns of Use," N. Engl. J. Med., 328, no. 4 (1993): 246-52; D.M. Eisenberg et al., "Trends in Alternative Medicine Use in the United States, 1990-1997: Results of a Follow-Up National Survey," JAMA, 280, no. 18 (1998): 1569-75.
11. B. O'Connor et al. (Panel on Definition and Description, CAM Research Conference, April 1995), "Defining and Describing Complementary and Alternative Medicine," Alternative Therapies, 3, no. 2 (1997): 49-57.
12. D.B. Mowrey and D.E. Clayson, "Motion Sickness, Ginger, and Psychophysics," Lancet, 1, no. 8273 (1982): 655-57.
13. J.T. Chibnall, J.M. Jeral and M.A. Cerullo, "Experiments in Distant Intercessory Prayer: God, Science, and the Lesson of Massah," Archives of Internal Medicine, 161, no. 21 (2001): 2529-36.
14. T. Kuhn, The Structure of Scientific Revolutions, 2nd ed. (Chicago: University of Chicago Press, 1970).
15. Id. at 66-91.
16. Schneiderman, supra note 5, at 85.
17. Id.
18. See Hufford supra note 4.
19. H.K. Berthold, T. Sudhop, and K. von Bergman, "Effect of a Garlic Oil Preparation on Serum Eipoproteins and Cholesterol Metabolism: A Randomized Controlled Trial," JAMA, 279, no. 23 (1998): 1900-02.
20. M. Curtius, "Touting Garlic's Glories. Despite Study, Gilroy's Devotion to Aromatic Crop Is Unshaken," Los Angeles Times, June 21, 1998, at A3.
21. The authors concluded that "there is no evidence to recommend garlic therapy for lowering serum lipid levels." Berthold, Sudhop, and von Bergman, supra note 19, at 1902.
22. In Berthold and colleagues' response to a critical letter in the next issue of JAMA, they state, "Even though the conclusions of our study apply only to the preparation we used, we believe that convincing evidence of lipid-lowering effects of any garlic preparation is still lacking." L.D. Lawson, "Effect of Garlic on Serum Eipids" (with reply from Berthold, Sudhop, and von Bergman), JAMA, 280, no. 18 (1998): 1568. Although in the reply by Berthold, Sudhop, and von Bergman they offer brief comments regarding other studies that they feel support their conclusion, and offer a reason for having selected the remedy they used in their study, the fact remains that in their conclusion to their published paper they stated their broad conclusion regarding garlic and serum lipids as if the statement were a reasonable inference from their study. Further, they do not note the steam distillation as a possibly controversial element.
23. Schneiderman, supra note 5, at 85.
24. Id.
25. K. Linde et al., "Are the Clinical Effects of Homeopathy Placebo Effects? A Meta-Analysis of Placebo-Controlled Trials," Lancet, 350 (1997): 834-43.
26. J. Kleijnen, R Knipschild, and G. ter Riet, "Clinical Trials of Homeopathy," British Medical Journal, 302 (1991): 316-23.
27. B.L. Beyerstein and S. Downie, "Naturopathy," Scientific Review of Alternative Medicine, 2, no. 1 (1998): 20-28.
28. Id. at 24.
29. Id.
30. Id.
31. Id. at n.20, n.25, and n.27.
32. A.J. Vickers et al., "Can Acupuncture Have Specific Effects on Health? A Systematic Review of Acupuncture Anti-Emesis Trials," Journal of the Royal Society of Medicine, 89 (1996): 303-11.
33. Id. at 303.
34. Id.
35. Id.
36. Id. at 310.
37. G. ter Riet, J. Kleijnen, and K Knipschild, "Acupuncture and Chronic Pain: A Criteria-Based Meta-Analysis," Journal of Clinical Epidemiology, 43 (1990): 1191-99.
38. "NIH Consensus Conference: Acupuncture," JAMA, 280 (1998): 1518-24.
39. L. Rosa et al., "A Close Look at Therapeutic Touch," JAMA, 279 (1998): 1005-10.
40. J.G. Turner, The Effect of Therapeutic Touch on Pain & Anxiety in Burn Patients, grant no. 94020 final report, Tri-Service Nursing Research Program (November 14, 1996), cited in Rosa et al., supra note 39, at 1007.
41. Id.
42. Lorenzo's Oil, produced by Universal City Studios, directed by George Miller, 135 min. (1992), film.
43. Schneiderman, supra note 5, at 89.
44. R. Rubin, "Lorenzo's Oil Brings Hope for the Afflicted: New Findings Support Unorthodox Treatment of Rare Brain Disorder," USA Today, October 22, 2002, at 1A, 2A.
45. Id. (quoting Dr. Hugo W. Moser).
46. H.W Moser, "Suspended Judgment: Reactions to the Motion Picture 'Lorenzo's Oil,'" Controlled Clinical Trials, 15 (1994): 161-64.
47. Id. at 162.
48. Id. at 163.
49. Id. at 161.
50. H.W. Moser, Letter to the Editor, "Lorenzo's Oil and Thrombocytopenia in Patients with Adrenoleukodystrophy," N. Engl. J. Med., 328, no. 15 (1993): 1126-27.
51. Id. at 1126.
52. Schneiderman, supra note 5, at 92.
53. J.M. Kauffman, "Alternative Medicine: Watching the Watchdogs at Quackwatch," Journal of Scientific Exploration, 16 (2002): 312-37.
54. Id. at 312.
55 Id. at 321.
56. Id.
57. Id.
58. See, e.g., L.T. Chappell and M. Janson, "EDTA Chelation Therapy in the Treatment of Vascular Disease," Journal of Cardiovascular Nursing, 10 (1996): 78-86; W.C. Douglas, Jr., "Chelation Therapy - Better Than Ever!," Second Opinion, suppl. (March 1995): 1-4, and references therein.
59. W. Sampson, "On the National Institute of Drug Abuse Consensus Conference on Acupuncture," Scientific Review of Alternative Medicine, 2, no. 1 (1998): 54-55.
60. G.G. Moertel et al., "A Pharmacologic and Toxicological Study of Amygdalin," JAMA, 245 (1981): 591-94.
61. Schneiderman, supra note 5, at 91.
62. Id.
63. Id.
64. See Sampson, supra note 59, at 54-55.
65. P. Feyerabend, Against Method, rev. ed. (London: Verso, 1988): at 23-24.
66. R.C. Herdman et al., Unconventional Cancer Treatments, OTA-H-405 (Washington, D.C.: U.S. Government Printing Office, 1990).
67. Schneiderman, supra note 5.
68. Complementary and Alternative Therapies in the Academic Medical Center, a conference sponsored by the University of Pennsylvania School of Medicine, Philadelphia, November 9-12, 1999.
69. See Linde et al., supra note 25.
70. E.A. Murphy, Skepsis, Dogma, and Belief: Uses and Abuses in Medicine (Baltimore: Johns Hopkins University Press, 1981).
71. Id. at 157.
72. E. Davenas et al., "Human Basophil Degranulation Triggered by Very Dilute Antiserum Against IgE," Nature, 333 (1988): 816-18.
73. J. Maddox et al., "'High Dilution' Experiments a Delusion," Nature, 334 (1988): 287-90.
74. J.C. Whorton, "Patient Heal Thyself: Popular Health Reform Movements as Unorthodox Medicine," in N. Gevitz, ed., Other Healers: Unorthodox Medicine in America (Baltimore: Johns Hopkins University Press, 1988): 52-81, at 60.
75. Id. at 69.
76. D.R Burkitt, A.R.P. Walker, and N. S. Painter, "Effect of Dietary Fibre on Stools and Transit-Times, and Its Role in the Causation of Disease," Lancet, 2 (1972): 1408-12.
77. J. Salmeron et al., "Dietary Fiber, Glycemic Load, and Risk of NIDDM in Women," JAMA, 277 (1997): 472-77.
78. J. Avorn et al., "Reduction of Bacteriuria and Pyuria After Ingestion of Cranberry Juice," JAMA, 271 (1994): 751-54.
79. A.B. Howell and B. Foxman, "Cranberry Juice and Adhesion of Antibiotic-Resistant Uropathogens," JAMA, 287 (2002): 3082-83.
80. M.H. Beers and R. Berkow, eds., Merck Manual of Diagnosis and Therapy, (Whitehouse Station, New Jersey: Merck Research Laboratories, 1999): at 1888.
81. E. Hilton et al., "Ingestion of Yogurt Containing Lactobacillus acidopbilus as Prophylaxis for Candidal Vaginitis," Annals of Internal Medicine, 116 (1992): 353-57.
82. G. Freed et al., "National Assessment of Physicians' Breast-feeding Knowledge, Attitudes, Training and Experience," JAMA, 273 (1995): 472-76.
83. Id. at 475.
84. Both of the Eisenberg et al. surveys of CAM use (supra note 10) illustrate this, as do many other surveys. A well-designed poll reported in Newsweek (March 31, 1997) found that 87 percent of Americans believe "that God answers prayers" and that 79 percent of Americans "say God answers prayers for healing someone with an incurable disease." One can readily obtain a large body of additional survey data supporting this point by visiting Polling the Nations (a subscription database of all sorts of survey data, at <http://www.orspub.com>), and entering "prayer" as the search.
85. H.G. Koenig, M.E. McCullough, and D.B. Lacson, Handbook of Religion and Health (Oxford: Oxford University Press, 2001).
86. R. Sloan et al., "Religion, Spirituality & Medicine," Lancet, 353 (1999): 664-67; R. Sloan et al., "Should Physicians Prescribe Religious Activities?," N. Engl. J. Med., 342 (2000): 1913-16.
87. W Harris et al., "A Randomized, Controlled Trial of the Effects of Remote, Intercessory Prayer on Outcomes in Patients Admitted to the Coronary Care Unit," Archives of Internal Medicine, 159 (1999): 2273-78; K.Y. Cha, D.P. Wirth, and R. Lobo, "Does Prayer Influence the Success of In Vitro Fertilization-Embryo Transfer? Report of a Masked, Randomized Trial," Journal of Reproductive Medicine, 46 (2001): 781-87.
88. D.W Seldin, "Presidential Address: The Boundaries of Medicine," Transactions of the American Association of Physicians, 94 (1981): lxxv-lxxxvi, at lxxxv.
In addition, make sure to read these articles:
- Security+ Certification: Computer Forensics and Incident Response
- The Butler Did It
- Watch Your Credit To Safeguard Your Small Business
- New SRI International Study Examines Public Attitudes about Science and Technology in Asia...
- Florida Study: Herbal Remedy Use Affected By Culture.
- Using a controlled research pilot to enter the alternative medicine arena.
- International Wellness Conference to Explore Alternative Medicine and Patient...
- Phoenix Well Care Seeks Participants for NIH Study on Chelation and Heart Disease.
- The morbidity and mortality of scientific illiteracy.
- Sojourns joins traditional and alternative medicine
- Alternative Medicine Advisor.
- A.D.A.M. Acquires Leading Wellness and Alternative Medicine Company.
- U.S. Patent allowance awarded to new alternative medicine new technology; includes...
- GRANTS AWARDED FOR RESEARCH IN ALTERNATIVE MEDICINE.
- Straight talking about alternative medicine
Email
Digg It
del.icio.us
Related Resources
Press Releases
- Florida Study: Herbal Remedy Use Affected By Culture.
- Bastyr University Research Scientist Appointed to Institute of Medicine Committee on Complementary and Alternative Medicine.
- Investigation Findings, Corroborated by Empirical Evidence of the Schisandra Berry as Cardio-Tonic in Traditional Chinese Medicine for Hundreds of Years, Support the Use of Schisandrin B in the Prevention...
