The Irrelevance of Animal Experimentation
By Murry Cohen, M.D.
Psychiatry is, to be sure, a branch of medicine. Furthermore, it has undergone a process of increased medicalization in recent years. One would think, therefore, that to the extent that animal experimentation serves as the scientific underpinning for medicine in general — a highly dubious assumption — the same would apply to psychiatry.
Regardless of the validity of this notion for medicine — and there are powerful arguments to the contrary1 — it can be argued that it applies less to psychiatry because of psychiatry's concern with the human mind. After all, other fields of medicine focus on percepts: biologically discrete entities such as organs or processes: heart, lungs, skin, blood, neoplasm, allergy, fertility, etc. Psychiatry is the only medical discipline which attempts to treat a conceptual abstraction — the human mind does not actually exist as a biological percept — as if it really existed perceptually.
This need to change psychology and psychiatry into disciplines considered to be commensurate with the rest of medicine has jokingly been referred to as "physics envy," understandable by those who remain psychoanalytically-oriented. As an example of this, witness the translation of Freud's allusion to "die Seele," or the soul, in his psychoanalytic formulations, into its English version, the "psychic apparatus."
"Psychic apparatus" certainly appears to be a notion more susceptible to conventional scientific scrutiny involving animal models; however, the amount of reductionism required to convert the concept of the soul to one of a machine-like "psychic apparatus" undermines whatever validity Freud's original concept might have had. Therefore, the application of animal testing to such a scientific paradigm creates a situation whereby the results are no longer biologically valid. The object to be studied is altered to fit the method used to study it, and the results are valueless. Methods to be used to study human behavior which are scientifically valid include clinical and epidemiological research, techniques which are to human psychology what post-mortem examination is to human pathology. More about this later.
The myth of objectivity in biomedical research has been previously exposed and written about extensively.2,3,4,5 This absence of objectivity particularly applies to psychiatric and psychological research. Objectivity is thwarted by many factors. One is unconscious, subjective, and, often, implicit societal influence. Alvin Toffler writes about science in general that it "is not an independent variable. It is an open system embedded in society and linked to it by very dense feedback loops. It is powerfully influenced by its external environment...."6 Harvard paleontologist Stephen Jay Gould has written:
Another impediment to objectivity is personal interest. Broad and Wade noted in l982 that "careerism" eroded the quality of scientific research, since 90 percent of all scientists who had ever lived were living then, writing for at least 8,000 journals in medicine alone.8 To further support this contention, two sociologists found that about half of all scientific papers were never cited at all during the year after they had been published, and they concluded that the majority of scientists publish work that has "little or...even zero impact on the forward march of knowledge."9
Writing in The Pharos, the journal of the Alpha Omega Alpha Honorary Medical Society, Harvard Medical School graduate Dr. Josef Fischer, Chairman of the Department of Surgery at the University of Cincinnati Medical Center, wrote in l987: "scientists do research not for altruistic reasons but to further their careers..."10 In a letter to The Lancet in l986, Dr. E.J.H. Moore wrote: "The pressure on young doctors to publish and the availability of laboratory animals have made professional advancement the main reason for doing animal experimentation."11
A third problem contributing to non-objective scientific conclusions is the obscuring nature of the jargon often used to describe them. This reductionistic, devivified, mechanistic, and often euphemistic manner of communicating biological phenomena often adversely affects experimenters' thinking and results in the belief on the part of scientist and public alike that more is being stated of value than is actually the case. Nield wrote in "Why Scientists Keep It Obscure" that "technical language can fool laypersons into concluding importance when none exists."12
Finally, scientific fraud is yet another rather overt obstacle to objectivity in scientific inquiry. It may seem odd to relate these two factors, however much fraud is unconsciously motivated, resulting in subtle manipulations of the variables to prove one's hypothesis. Indeed, the very system whereby scientific inquiry flourishes has been called into question.13,14,15 Stewart and Feder, after studying l09 publications authored by a prominent Harvard researcher together with 47 other scientists, concluded that "...new light (has been shed) on the system that ensures the integrity of the scientific literature. Certain lapses may be more frequent than is commonly believed."13 Science has reported that "two researchers from NIH allege that the scientific literature may be full of minor errors...."14 Dr. Irwin Bross has written repeatedly on scientific fraud and will be introducing an expanded view of it at this conference.15,16
Why this emphasis on impaired objectivity and fraud? Psychiatric and psychological research in general, and animal research in particular, are prone to these destructive influences. Research pertaining to the mind is inherently "soft" when measured against those postulates of Robert Koch designed to use animal models of disease to study, isolate, and identify single causes of disease.17 Even if this method did apply to animal-based, non-psychiatric biomedical research — a dubious supposition — when applied to psychiatric research, medical historian Dr. Brandon Reines finds no evidence that "animal models of psychopathology have contributed anything,"17 and Swedish psychiatrist Dr. Per Dalen also is skeptical, believing that the animal model approach "does not yield much."18
Societal influences in psychiatric research are obvious. Because we are concerned with diseases of the mind, because the mind is the repository of societal values, and because so many symptoms and behavior patterns can be considered pathological only as they represent deviance from social norms, it is natural for social norms to skew experimental results. Psychiatric research and practice are highly vulnerable to these confounding factors. For example, psychotherapists must vigilantly attempt to exclude personal value judgments from the psychotherapy they dispense.
The effect of personal interest is particularly felt in psychiatric and psychological research because these researchers, unlike their more biologically-grounded brethren, must fight for their places in academia in fields less embraced as "real" sciences when compared to other medical specialties. Since financial support derives primarily from government research grants, and since the ability to attract such support can profoundly affect one's career, performing research, regardless of its need or validity, becomes a priority. Dr. Fischer has complained that "research is in trouble because the research goals of those who set the priorities of the biomedical research establishment have become distinct from those of the general population."10
The question of jargon is uniquely relevant to psychiatric and psychological research, resulting in the term "psychobabble." This pseudoscientific language can appear to place psychiatry and psychology research on a par with other medical research, which, in turn, attempts (unsuccessfully) to compete for the "hard" scientific validity of natural science research. To speak of "drug cravings" in animal addicts made physiologically dependent on drugs, to speak of animals forced to drink alcohol as "being high,"19 to consider an "opponent-process theory of acquired motivation" as having anything to do with the etiology of drug addiction,20 to refer to monkeys taught learned helplessness as being "depressed,"21 to conclude that primates undergoing deprivation are "schizophrenic," and to label animals traumatized in laboratories as "neurotic,"22 is to disdain epistemology, indulge in self-serving fantasy, demean proper scientific exploration, obfuscate the significance of empirical facts, and manipulate, mislead, and delude everyone.
The emergence of fraud in psychiatric and psychological research poses a real threat because of both the inherently subjective nature of matters pertaining to the human mind and the psyche-soma split. This split represents a dualism that remains so prevalent in the theory and practice of psychiatry and psychology that it ultimately precludes proper and necessary integration of knowledge because, in reality, it is a false dichotomy. As prominent psychiatrist Dr. George Engle, University of Rochester School of Medicine Professor Emeritus, a pioneer in the field of psychosomatic medicine, and developer of the biopsychosocial model of disease23 wrote in The Pharos in an article entitled "Where you think you stand determines what you think you see":
The absence of specificity of "mental illness," the innate uncertainty of its nature, together with the idea that the "mind" is merely an abstraction, explain the total inappropriateness of animal experimentation as a method for arriving at knowledge or truth in psychiatry or psychology. Psychiatrists Matthew Dumont25 and John Beahrs26 have both written about the applicability of quantum physics' Heisenberg Principle of Uncertainty to human behavior. They believe that medicine and psychology must stop short of conventional physics and "...accept ambiguity and unity and...inevitable limitations,"25 that there is "...a fundamental limitation within the domain of mental health against even the possibility that we might acquire knowledge that is totally precise and fully reliable yet comprehensive."26 They both support Severs' call for a new ecologic, holistic model to replace the prevalent atomistic model of disease, a new model stating everything is connected, that it is not possible to simply do one thing, that "disease is the result of the complex interaction of factors relating to the host, the agent, and the environment."27
Dumont believes that psychiatry, more than even the rest of medicine, cannot "think of illness in terms of categories or organ pathology...but rather must reach a fuller understanding of health and its perturbations."25 Beahrs suggests that "scientific psychiatry" has its limits and that one must accept the "psychic uncertainty principle" in mental health. Certainty, in his opinion, is impossible because of the inverse relationship between precision and relevance (or reliability); that is, since certainty requires both, in psychiatry and psychology, the more we see of one, the less of the other.26
It will be demonstrated later that animal experimentation is based on just the opposite premise; namely, that scientific manipulation of controllable variables will result in both precision and relevance, that certainty is possible in medicine in general and in psychiatry and psychology in particular. The notion of uncertainty has been opposed and shunned by western scientific and technological thinking, and the misapplication of animal research is understandable, whatever the disastrous consequences. The opposition to understanding the true nature of health and disease has resulted in the inappropriate natural science-based application of animal experimentation to matters medical and psychiatric.
However, all the animal experimentation in the world cannot and never will obliterate the uncertainty inherent in human psychology and psychopathology and address itself to these new and important concepts of health and disease, despite the implicit, and sometimes explicit, premise that it can. This premise is based on a static, simplistic, dualistic, deterministic, non-holistic, mechanistic, and reductionistic vision, allowing the irrelevance of animal experimentation to psychiatry and psychology to be transparent.
Animal experimentation conforms to the now outdated concept of human beings as mere machines — a complex amalgam of biochemical events, physiological processes, anatomic structures, and pathological malformations or malfunctions. This idea supports the observation of psychiatrist Roy R. Grinker, Sr. of an "intense interest in psychological and sociological factors capable of altering biochemical functions," resulting in "psychiatrists...[talking] endlessly about the total personality...[but] ...gain[ing] little more knowledge than ancient philosophies had."28 Conceptualizing human beings as machines impedes the integration so vital and necessary to truly understand mental disease.
Dr. Michael Simpson echoed these ideas in his criticism of "physico-chemical reductionism," a belief that man is merely an accidentally superb molecular arrangement, and in his suggestion that we remember that "[m]edicine always did pay great attention to the symptoms of illness and its setting, and it stressed psychological and spiritual aspects."29 Since psychiatric illness almost always includes psychological and spiritual factors, Simpson's reminder is especially timely, for these are the very factors almost totally ignored by animal experimentation, which consists of a collection of data based on manipulating biological variables in animals, and then misapplying them to human organisms.
Empathy, so important to psychiatry and psychology, is eroded by the mechanistic view of disease and the undervaluing or even ignoring of psychological and spiritual components.30 Animal experimentation opposes the development and maintenance of empathy in at least two ways. First, when clinicians are exposed to animal experimentation, sensitivity, so important in empathic awareness, is eroded, because clinicians must desensitize themselves to the pain and suffering of other sentient creatures. Second, the body of theory derived from animal experimentation is so bereft of social, psychological or spiritual ingredients that its applicability to humans eliminates the possibility of empathic connection or communication with people.
We must worry about this process of desensitization and the erosion of empathy, for the consequences are potentially grave. Since the primary identification of a physician currently is that of a scientist rather than a healer, the desensitization of the scientist through animal experimentation can have dire effects when this detachment infiltrates the persona of the physician. Crighton's critical characterization of his medical school teachers as "scientist-physicians" highlights this relationship.31 Since many basic scientists must maim and kill animals in their laboratories, and since the physician is now identified with the scientist rather than the healer, one can more readily understand Lifton's characterization of Nazi doctors as "healer-killers."32 This is especially so considering the historical fact that the Nazi physicians were prepared for human experimentation by animal experimentation.33 Since medicine and psychiatry are primarily arts bolstered by scientific knowledge, if the basic identity of a physician is that of a "scientist" rather than of a humanistic clinician, then the "scientific detachment" of the scientist would impede the physician's clinical skills and therapeutic efficacy, constituting grave dangers to patients, creating major changes in the medical profession, and threatening potentially devastating effects on the physician's self-image.34
It may appear odd that I speak of animal research data ignoring psychological factors when so much animal research pertains to psychological research. This genre of research is known as behavioral, experimental, or physiological psychology. It was initiated by the bacteriology-trained Karl Lashley, who hoped that the "accustomed...method of animal experimentation...would shed light on human psychology."35 After thirty years, Lashley found no value in all his years of research in human memory.36 Nevertheless, physiological psychology remained a valid discipline whereby experiments in animal physiology would reveal insights into the physiological basis of human psychology.
In l968 psychologist Don Bannister complained that physiological psychology was a myth and that "no form of [it] can be a science or part of a science."37 Bannister's epistemological argument was that:
He listed four reasons:
Bannister concluded that "the conceptual contradictions involved in physiological psychology are such that experiments within its domain cannot add to the coherent development of psychology as a science."37
Sixteen years later, in l984, three physiological psychologists concluded that their field was undergoing an "identity crisis" because "the development of neuroscience has raised questions about the place of physiological psychology in both psychology and neuroscience," and that one must resist the "naive reductionistic assumption that behavioral phenomena can and will simply be explained by reduction to some brain event or events, whatever their specific nature at some future time might be." They blamed reductionism and the false belief that psychological functions are already sufficiently understood to account for the tendency by physiological psychologists to explain things molecularly rather than in terms of behavior, to avoid developing integrative and generalizable theories of behavior, and to regard behavior as unitary events rather than complex phenomena requiring further analysis. They described the conceptual bankruptcy of their field and its practice to "utilize those behavioral techniques that are already in vogue (or) concentrate their energies on mastering the latest techniques developed by colleagues for technical sophistication." They concluded that the proper role of the physiological psychologist should be that of an "integrator and theoretician rather than...scientific practitioner," providing "genuine assistance to the practitioner..."38
So it is suggested that all of the animal research supporting the field of physiological psychology — research which is the mainstay of the field — is a waste; a waste because it fails to elucidate human psychology or behavior for the following reasons:
If animal experimentation should not be employed to gain insight into human behavior and psychopathology, what should? The obvious answer is clearly articulated by Dumont:
Noted child psychiatrist Jerome Bruner would probably agree that human psychiatric and psychological diseases are distinctly human, and are poorly enough understood and biologically sufficiently complex that attempts to produce them in animals are misguided and self-defeating."39
But what about animal studies, other than those performed in the fields of behavior, physiological psychology, and addiction, which purport to further our understanding of psychobiological phenomena? An appreciation of psychoimmunological factors would demonstrate the fallacy of this approach. Immunologic function affects every system of the body. Psychological concomitants to states of stress, hunger, thirst, sleeplessness, hypersexuality, pain, and the like effect immunologic function. Since these very states are those induced by experimental protocols, and, indeed, are inherent in even the best laboratory conditions, they must, and always do, alter experimental results via immunological alteration.40 This is predictable, inevitable, and ubiquitous.
For this reason, laboratory animal research inherently is scientifically flawed. Not only can the results not be generalized to humans, they cannot be applied to other animals, to members of the same species, or, even, to the same individual animal, for the experimental conditions were unnatural and the data collected riddled with confounding variables, any one of which could easily alter the significance of the findings.
A clinical example of the importance of setting is the routinely found difference, between experimental and general use, in a drug's performance when used in physicians' offices. Despite identical drug and dose, similar prescribers, and patients of the same species with the same diagnosis, the setting differs. This is comparable to the difference between an animal's natural setting and its artificial laboratory environment. This difference produces major pharmacologic surprises, for, as psychiatrist Norman Zinberg emphasized, the effect of setting on drug actions cannot be underestimated.41
Witness the safety of nomifensine (Merital) when prescribed in the context of office experimental use, but the emergence of a lethal hemolytic anemia, resulting it its being withdrawn after causing several deaths, when prescribed for general use in the office. Whereas it is difficult to detect idiosyncratic side effects of a drug in a pre-clinical setting, the search would be more thorough and comprehensive were it not for the false sense of security provided by animal screening.
Underestimating the role of setting is quite dangerous and has resulted in multiple human catastrophes, all directly related to the belief in, emphasis on, and security derived from animal testing. Belief in the (false) protective value of this screening procedure has quite possibly prevented the implementation of more appropriate post-marketing monitoring techniques. It would be wiser to consider the recommendation of well-known clinical pharmacologist Dr. Kenneth Melmon, Chairman of the Joint Commission on Prescription Drug use in l980, to monitor side-effects of drugs already on the market (a plan rejected because of the $l0 million necessary to launch the program.)42 Dr. Brandon Reines also advocates public health research consisting of computerized post-marketing surveillance: "Rather than continuing with excessive requirements for pre-market testing, the regulatory agencies should maximize the potential for 'clinical serendipity'"43
For the above reasons, animal models of psychiatric diseases are scientifically invalid. But even were it not true that laboratory conditions themselves destroy accuracy in biological research, we must confront the fact that animal models of human psychiatric disease do not exist.5 As far as we know, animals do not suffer from schizophrenia, affective disorder, substance use disorder,19 anorexia nervosa, bulimia, borderline personality disorder, obsessive-compulsive disorder, panic disorders, hypochondriasis, psychogenic pain disorder, and the like. Attempts to create these conditions in animals are intrinsically doomed to failure given our lack of understanding of the etiology and complex biological structure of these conditions.
Why, then, does animal research persist, both in psychiatry/psychology and other fields? There are seven major answers:44
Despite these factors, animal research will not lead the growth and development of psychiatry or psychology into the 2lst century, and therefore it has no place as an experimental underpinning of these fields. Human behavior can be understood best by clinical and epidemiological trials. Addiction phenomena can be studied by working with persons already afflicted. New drugs can be developed and tested by a combination of mathematical modelling,45 in-vitro screening,46 careful human clinical testing (already in place), and computerized post-marketing surveillance.43 Neurobiological phenomena can be understood by autopsy and biopsy study, as well as by implementation of recent innovative scanning techniques, including computerized axial tomography, positron-emission tomography, and magnetic resonance imaging.
We are at a crossroads. It is clear that henceforth animal experimentation can only lead psychiatry and psychology away from their proper purview and mandate. The continuation into the future of this invalid method of seeking knowledge, finding facts, and identifying truths can only impede the growth and development of psychiatry and psychology, prevent their expansion, and hinder and retard the process whereby new and innovative techniques can be discovered to lead these honorable disciplines forward to new understanding in the 2lst century.
Question: Seligman did learned helplessness studies and drew conclusions which he extrapolated to human data. Can you comment on those studies? Seligman was once my colleague at the University of Pennsylvania and I remember having several discussions with him in which he maintained vehemently that the learned helplessness work he did with animals was directly transferable to human beings and he cited a number of clinical applications. Can you comment on that?
Answer: It's just like any other animal model. The work that the questioner is referring to was conducted by a psychologist named Martin Seligman at the University of Pennsylvania. He taught animals learned helplessness and claimed that it was a suitable model for clinical depression found in human beings. It's true that when you examine an animal who has been taught learned helplessness, it looks like a depressed animal. Of course, you can't ascertain whether or not the animal really is depressed because you can't question it along these lines. This is generally true with many of the animal models currently in use. Very many of them look quite similar to the clinical syndrome in human beings. But it's not sufficient to merely look like it. It must be identical. The pathophysiology, the anatomy, the biochemistry, and all other biological parameters must be identical. If this is not the case, if there's a single variable that's not identical, there's no way to confidently generalize the work of animals, as Dr. Buyukmihci said, to human beings. It just can't be done, and even less so in psychiatry and psychology, because these are self-oriented disciplines which require talking to people and finding out how they feel. And, as much as some of us hold animals in high regard, we can't really, at this point, ask them how they feel, although there are some people working in that field as well. So, I believe that learned helplessness is not a model for human depression, and that Selgiman may have himself retracted it as such. It certainly isn't currently used in clinical practice. Along the same lines, Roger Ulrich retracted his mouse aggression model as a model for human aggression. And the currently performed monkey maternal deprivation studies of Harlow and his disciples also serve as a poor model for human maternal deprivation. The point is that it is not sufficient for a condition in animals to very closely resemble a human illness. It has to be identical and, by definition, it can't be. Therefore, animal models are, by definition, conceptually flawed.
Question: You just mentioned Harlow. Do you know if the work of his successors is still going on?
Answer: I'll defer to Don Barnes to answer that question.
Don Barnes: I was at Madison, Wisconsin two weeks ago and a great deal of that work is going on, a good deal more than I would have thought, but I'll talk a little about that tomorrow. We're going to have some overlap.
Answer: Another interesting aspect of this is that clinical psychiatrists by the name of Bowlby and Spitz studied maternal deprivation in humans in the 40's and the 50's before Harlow began his work in the early 60's. Everything learned by Harlow, his disciple Suomi, and others was known prior. The other interesting thing is that Bowlby, himself, two years ago, published a review article in the American Journal of Psychiatry. Bowlby is a British psychiatrist and probably the world's leading expert in maternal deprivation. The article didn't contain a single citation of Harlow's work or any of the non-human primate maternal deprivation work of his disciples. Therefore, we can conclude that, so far as understanding human behavior is concerned, all of the Harlow work is of no value.
Question: Well, I'm convinced of that and others are too, but it makes it more sad that it still goes on.
Answer: It goes on because it's readily funded and because of the points that I, Dr. Buyukmihci, and others have made and will make; that is, it's the known. There's a whole body of information already in place, with many, many variables already having been manipulated. Researchers are curious about what would happen if the temperature were changed somewhat, or if a bit of sleep deprivation were introduced, or if the humidity were different. These so-called basic scientists, with this limitless curiosity, this insatiable hunger for knowledge, are drawn to investigate that sort of thing, regardless of its pertinence to humans. And because of the peer review process, because these grants are reviewed by people who previously were interested in the same thing, and, indeed, performed that kind of work themselves, the work goes on and on.
Question: Are you saying that there is no way that we can learn about animal behavior or about human behavior as a result of physiological changes in the animal? For instance, I know as a veterinarian that a bitch responds differently under the influence of progesterone than under the influence of estrogen. Does that not teach us something about our own behavior as well? I mean, if we just take a look at it as a biological phenomenon, are there not things that we can learn?
Answer: Well, to answer the first part of your question, it may or it may not teach us something about human behavior. The only way to tell is to repeat the experiment altering progesterone and estrogen levels in human beings. We know that diethelstilbesterol (DES) did not produce vaginal carcinoma in the offspring of animals in whom it was tested, but it does produce vaginal carcinoma in daughters and granddaughters of women who took it as a fertility drug. This is the problem with animal experimentation. You never know in advance whether or not the results will apply to humans. What we could say is that, if the animal being tested was in a natural state, and its responses were not skewed by laboratory-induced variables, we might learn something scientific about the animal. But whether or not what we learn has anything to do with human beings can only be determined by testing on humans. But if the results may or may not pertain to humans, then why continue the animal tests? That's what this conference is all about. Why not find other methodologies that offer a greater degree of certainty?