Placebo Studies Research Paper

In clinical trials, patients are the subjects of all treatment modes, including drugs, radiation, surgery, psychotherapy. This research paper describes how trials have been conducted and the results interpreted. Clinical trials are termed ‘open label’ when patients and researchers know that the active drug is being given; ‘single-blind’ when the patient does not know what he or she is receiving; and ‘double-blind’ when neither the patient nor the research team knows who is receiving what. ‘Randomization’ denotes the assignment of experimental subjects through a table of random numbers, dice, or other statistical tools. ‘Placebo’ was first defined for medical use in 1785 as a substance that was pharmacologically inert, given to reinforce a patient’s expectation to get well. Glantz’s definition of placebo terms the placebo effect ‘a response attributable to therapy per se, as opposed to the therapy’s specific effects’ (Glantz 1997). Current usage differentiates active placebo that has pharmacologic effect from inert placebo.

Thomas Kuhn, the historian of science, defines normal science as the prevalent system of beliefs and practice. He likens practitioners to chess players, who test alternative moves but do not challenge the rules of the game. The game is challenged through discovery or a new hypothesis. Practitioners of normal science combat the skeptics, solve the problem, postpone solution, or attempt to reconstruct the discussion from new fundamentals. Kuhn quotes Max Planck: ‘A new scientific truth can often prevail only when its opponents die and the new generation grows up.’

Kuhn underlines that resistance to change evolves not from wrong-headedness but from the impossibility of measuring or comparing different standards. Scientists and laymen could not accept Einstein’s theory of relativity that proposed that space was curved, because Newtonian physics had proven that space was flat, homogeneous, and unaffected by the presence of matter. In the 1960s and 1970s depression was thought to be caused by psychological factors until challenged by evidence that a functional deficit of neurotransmitters, norepinephrine, dopamine, or serotonin or abnormality of their receptors is the cause. While opposition to this remains, most researchers now are refining particulars of neurotransmitter metabolism.

The acceptance of the double-blind method and placebo controls are among the most important developments in science and illustrate how new discoveries drive new theories that are eventually accepted in place of incompatible, older theories.

1. History Of Placebo

Placebo concept and the word derives from the Hebrew ‘I will walk before,’ interpreted to mean ‘to please.’ It gained a secular connotation in the fourteenth century, to become derisive meaning one who ‘servily echoes another’ (Shapiro and Shapiro 1997). The first medical definition of placebo appeared in a medical dictionary in 1785. In the late eighteenth and early nineteenth century, it was defined as medication given to please. Eventually, it meant ‘a common method or medicine’ and referred to treatments as well as medicaments not yet classified as inert. References to placebo as medications were retained in the medical dictionaries until the 1950s.

The use of placebos as controls in clinical research emerged from their long history as therapeutic agents. For four millennia, medical practitioners utilized thousands of substances, 0.03 to 0.04 percent of which are pharmacologically active. These physicians did not discriminate among active and inactive concoctions, and it is not known whether active drugs were used in effective doses. In the nineteenth and twentieth centuries, some physicians’ rationalization for the use of placebo was mobilizing hope.

Efforts to introduce placebo controls into research designs had been confounded in the 1930s by the growing low esteem of placebo as treatment on the part of the medical establishment. But because of self-delusion or hypocrisy while placebos were considered quack treatments prescribed by quacks, most community physicians also used placebos. They gave vitamins without acknowledging their lack of activity for the conditions for which they were prescribed, for example. Jerome Frank, the American psychiatrist, has said that psychoanalysis might be the major placebo of the twentieth century.

Reformers from different generations concerned with the practices of their mainstream clinical colleagues advocated altertness to self-deception as to the efficacy of treatments. Roueche said that ‘results from 10 well-controlled cases were superior to haphazard impressions from 1,000 cases’ (Marks 1997).

2. Medical Theories

In primitive supernaturalism, illness was attributed to actions of the gods and treated with magic administered by shamans. Greek medicine combined rational and irrational explanations and treatments, using miracles, divine healing, and natural remedies. Dream interpretation was a precursor of the confession and psychotherapy. Galenic medicine of ancient Rome was predicated on Galen’s principle of opposites: if a drug caused fever, it would cure chills. Galenic theories had an answer for every question and maintained a tenacious hold on medical practice well into the nineteenth century. For example, Galen had observed pus in wounds and erroneously concluded that it was essential to healing rather than a symptom and made no attempts to relieve it. Because of this belief, thousands died of infection.

Theriac, a concoction that was the mainstay of Galenic medicine, had between 33 to 100 substances in it, eventually being prescribed as a remedy for almost every human ailment. Theriac contained opium, which alleviated pain but was addictive. The rest of the ingredients were mainly placebos; many were harmful.

The laying on of hands was a hallmark of Christianity’s return to mystical explanations of illnesses. Charles II of England used therapeutic touch on almost 100,000 persons to cure scrofula, glandular tuberculosis. Skeptics attributed cures to effects of the imagination, incorrect diagnoses to start with, and spontaneous remissions. In the nineteenth century, Claude Bernard asserted that science had to be ‘founded on observation and proved by experience.’ He influenced the development of studies on pathology, physiology, and bacteriology, but pharmacology continued to lag behind. However, criticisms of ‘normal’ medicine and its treatments were voiced, notably by Oliver Wendel Holmes, who urged the retention only of quinine for malaria, mercury for syphilis, colchinicum for gout, iodine for goiter, and ipecac for dysentery. If all medications ‘could be sunk to the bottom of the sea, it would be all the better for mankind and all the worse for the fishes,’ he said in 1860 in a much-quoted statement. Often paradigm shifts are spurred by knowledge acquired elsewhere.

Laboratory science was making major inroads by the end of the nineteenth and early twentieth century, when therapeutic chemical compounds began to be synthesized and replaced some of the naturally occurring medicaments. This was made possible by progress in laboratory sciences, biochemistry, physiology, genetics, and immunology, which are in turn based on chemistry and physics. Animal experiments where variables could be well controlled became standard.

While the concept of the scientific method was beginning to impact medicine in the nineteenth century, theories of therapeutics based on a priori reasoning, were still entertained. Homeopathy was a success because of placebo effects, but also due to its theoretical base that less is more. Miniscule dosing probably protected patients from ill affects of noxious substances dispensed in more robust doses by nonhomeopathic physicians.

3. Beginnings Of The Scientific Method

From antiquity, ‘open label’ experiments were conducted but results were explained by ‘medical ideas (that) were speculative … and medical hypotheses that relied on a priori reasoning (Shapiro and Shapiro 1997). In the seventeenth century, Sir Kenneth Digby, an English Renaissance man, carried out a single blind experiment on a single patient. He attributed the positive results to the popular concept of sympathetic powder (Shapiro and Shapiro 1997).

In 1872, John Tyndall, an English physician, suggested that the community pray for from 3–5 years for patients ill with fatal diseases to see if differences emerged between them and others not prayed for. If or how the suggestion was carried out was not clear.

Towards the end of the seventeenth century through the early eighteenth century, some British physicians called for comparative trials using ‘simple arithmetic,’ i.e., ‘open label’ trials. These reformers were generally Scots, provincials, without access to established centers of medical influence and power, and thus, more open to new ideas. Harry M. Marks, the historian, warns that the historical record of the clinical trial is incomplete and informs that Western European and American scientists influenced each other. He points out that among medical reformers, two parallel viewpoints were emerging: medicine can be a science if it rigorously applies research that is subsequently channeled into clinical practice, and clinical medicine can be as scientific as laboratory sciences if scientific methods are applied to judging medical treatment.

Marks further subdivides the history of the clinical trials in the US into organizational and methodological reform. The first Food and Drug Act was adopted in 1906. The American Medical Association Council on Pharmacy and Chemistry was established in 1912. The standards of the AMA Council on Pharmacy and Chemistry set by expert opinion were adopted in the Food and Drug Act of 1938.

In 1958, the AMA Council on drugs authorized the publication of a paper by Walter Modell and Raymond Houde and ‘approved the use of the doubleblind method, controlled comparisons, randomization, (and) statistics …’ (Shapiro and Shapiro 1997). By 1970 the Food and Drug Administration was empowered to monitor the safety and effectiveness of new drugs based on adequate, well-controlled investigations using appropriate statistical methods, the result of the reforms in methodology that influenced organizational reform.

In the US, Marks credits successive generations of challengers to the methodological status quo, nineteenth-century therapeutic nihilists, and also followers of Claude Bernard’s rational therapeutics that stressed knowledge of etiology before treatment. The therapeutic reformers of the twentieth century included Torald Sollmann, George H. Simmons, Thomas C. Chalmers, and Morris Fishbein. Simmons, editor of Journal of the American Medical Association, attacked manufacturers for flooding the market with exaggerated claims for unproven nostrums. Fishbein, another editor of JAMA, ‘unrelentingly described the fads and follies of healing cults and medical follies (Shapiro and Shapiro 1997). Chalmers urged in 1975 that the first patient put on any new drug be part of a clinical trial.

Charles S. Pierce and Joseph Jastrow are credited with the first use of randomization in 1885. Rivers in 1908 used the double-blind design and placebo (without using the terminology) in an experiment testing the influence of alcohol and other drugs on fatigue. He noted, ‘The action of suggestion can never be excluded from any medical treatment’ (Shapiro and Shapiro 1997). In 1917, John E. Coover first used the terms ‘controlled experiment’ and ‘randomization’ (Shapiro and Shapiro 1997), which only became standard through Sir Ronald Fisher’s book, The Design of Experiments in 1935. The number of experimental studies increased steadily through the work of, among others, Louis Pasteur, Robert Koch, and Rudolph Virchow. Torald Sollmann, professor and chair of the Department of Pharmacology at the Medical School of Western Reserve University, headed for 55 years the AMA Council on Pharmacy and Chemistry. Sollmann urged that it was ‘essential that clinical experimentation should follow the canons of other scientific experimentation’ (Shapiro and Shapiro 1997). He coined the term ‘comparative method.’ Publications of the Council made reference to ‘the blind test’ and published what was probably the first study of the ‘double-blind procedure.’ In 1930, Sollmann recommended the use of the ‘blind test,’ used the term ‘placebo’ for the first time in research, and dropped the term ‘comparative method.’ But his news had little impact on practice, and the AMA Council on Pharmacy and Chemistry continued to rely on reviews by its member experts rather than on controlled studies.

The experts were researchers of unquestionable skill and integrity, but they rarely had experimental data and focused on confronting the tactics of manufacturers, their aggressiveness and exaggeration of claims, and on manufacturer-supplied laboratory evidence. Their influence led to a ‘seemingly deluded faith in experts,’ from 1900–1930 i.e., themselves and colleagues of the same ilk (Marks 1997).

A major milestone was a single-blind study in 1932, when Harry Gold tested the use of xanthine against placebo (lactose) for cardiac pain. He realized that physicians were asking the patients leading questions and prejudicing answers, and thereafter, tried to blind the physicians, as well. In 1935, Hediger and Gold compared two forms of ether in a ‘blind test’ and legitimized the use of placebos. Kwit summarized their insights: ‘Towards the end of the study we realized that the doctor must not know what he gives, the patient … what he receives, and the questioner must not know what was given’ (Shapiro and Shapiro 1997).

Like Sollmann earlier, Gold in the 1930s failed to have an immediate effect on medical practice. However, throughout the next two decades, researchers flocked to work with him and the new methodology, and he supervised many double-blind, placebo con- trolled studies. In a 1950s interview with Shapiro, Gold said, ‘The placebo and the double-blind were companions after our 1937 study.’

A 1948 joint study by the US Public Health Serviceand the British Research Council is generally considered the first clinical trial. It tested the effectiveness of streptomycin in the treatment of tuberculosis. Patient selection was randomized. Evaluators of the patient X-rays did not know which patients had received streptomycin and which placebo. The treatment was not blind, because streptomycin-treated patients received up to four injections per day. Since the controls were bed-rest only patients, it would have been unethical to subject them to injections of placebo.

4. New Paradigm I

It was not until the 1950s, when antibiotics were discovered and mechanisms of several metabolic diseases understood that modern medicine began, and clinical research could continue to make inroads. Through the 1960s and 1970s, the scientific method superseded authority and tradition. Treatments had to show sensitivity, specificity, and predictability through statistically sound techniques, randomization, the double-blind method, and placebo controls.

Current clinical standards for research are reflected in the US Preventive Services Task Force recommendations for treatment. Guidelines of 1989 (revised 1996), rank in descending order of reliability and ascending order of bias and misinterpretation randomized controlled trials, nonrandomized controlled trials, cohort studies, case control studies, comparisons between times and places, uncontrolled experiments, descriptive studies, and expert opinion.

5. New Paradigm II

The new scientific canon of clinical research encountered challenges and doubts almost as soon as it became established. Questions of method were raised about blindability in double-blind studies, as well as the precise definition, activity, and validity of placebo. Critical voices questioned ethical aspects of clinical trials that deprived half the patients of a medication that might be effective. The ensuing debate polarized ethicists and physicians.

6. Blindability

Extensive reviews of blindability indicated that studies took its validity for granted. It was suspect in the few double-blind trials that tried to carry out such tests of validity. Basoghu, in review of the literature found that only three of 23 double-blind psychopharmacology studies maintained blindability in patients, physicians, and assessors. Side effects, explanations in informed consent forms, past experience with similar drugs, and improvement with the active drug but not placebo sabotaged blindability. In his own comparison study of blindability of alprazolam, an anti-anxiety agent, with behavior therapy for panic disorder, Basoghu found that patients, treating therapists, and assessors were able to guess correctly, to a statistically significant extent, the patient’s category. Patients treated with the active drug were more accurate than those on placebo, and active drug accuracy was highest at late stage of therapy, when optimal dosing and maximal efficacy had presumably been achieved.

Morin et al. (1995), recognizing that judgments carried out at the end of a study as to what was drug, what placebo could be influenced by treatment response to active drug, compared Temazepam, a hypnotic, and an inactive placebo. They carried out three assessments over 8 weeks. All patients received cognitive or behavioral therapy. Both patient and physician groups showed a statistically significant high degree of accuracy in guessing at 1 and 8 weeks with no better than chance guessing at week 4.

From these and other studies, it appeared that certainty ratings increased with time, that experienced patients showed greater accuracy in penetrating blinding, and that patients on active drugs guessed better than those on placebo. In summary, neither the patients nor the administering physicians were blinded to any significant degree.

7. Placebo

Harry Gold had accepted earlier twentieth-century definitions of placebo, such as Rivers’ that placebo gives an inactive imitation of the substance being tested and Sollmann’s that placebo furnishes the comparative check of the natural course of the disease, and did not explore the nature of placebo and placebo effects. Current theories on the nature of placebo and placebo effects include spontaneous remission due to biological diversity, conditioning, mobilization of endogenous endorphins, and the statistical phenomenon of regression to the mean. In 1955, H. K. Beecher calculated that 35.2 percent improvement in 15 clinical trials involving different diseases was due to placebo.

McDonald and Mazzuca (1983) concede the evidence that patients improve during placebo treatment, but point out that that is not proof that placebo treatment is the reason. The percentage of improvement observed in placebo treated patients is identical to that predicted from a statistical phenomenon known as regression to the mean where both objective and subjective measures of abnormalcy tend to improve and move towards the mean with successive measurements. McDonald and Mazzuca (1983) illustrate this in a review of 30 randomly selected placebo controlled clinical trials. Placebo effects and regression to the mean are both ‘not uniform, constant, or predictable.’ (Regression is by definition a random phenomenon.) McDonald and Mazzuca do not negate the usefulness of placebo. In comparisons of known drug with unknown, placebo controls can clarify ambiguous results. Placebo controls are also useful in conditions that fluctuate.

McDonald and Mazzuca note that placebo effects can be created through conditioning, citing the work of Stewart Wolf on the effects of conditioning and suggestion on measurable physiologic events in human beings.

8. Placebo Amplification

Richard Thomson believed that fewer studies would show a significant difference between placebo and drug if an active placebo were used as control. He reviewed 75 studies of tricyclic antidepressants com- pared with active placebo (atropine) or with an inert placebo (lactose). Indeed, as postulated, antidepressant efficacy was ranked as superior in only one of the seven atropine placebo studies available, while the tricyclic antidepressant was ranked superior in 43 of 68 studies controlled by an inert placebo. One explanation might be that atropine might have an antidepressant effect of its own. However, Thomson concluded that placebo amplification, a situation where physiological effects of taking a drug convince subjects already committed to treatment that the treatment is, in fact, effective, was the phenomenon being observed.

Roger P. Greenberg compared a tricyclic antidepressant with a new drug being tested, as well as an inactive placebo. He postulated that in a study focused on a new drug’s effectiveness, there would be less interest in proving the efficacy of the older drug. Indeed, compared with claims in earlier literature, the older drug was much less effective.

Shapiro and Shapiro (1997) reviewed 27 studies of guessing of placebo and drug allocation and found that overall, both patients and physicians made correct identifications in 60 to 80 percent of tests. Eighty percent of drug guesses, but only 50 percent of placebo guesses were correct in their own study.

Fisher and Greenberg (1995) found in a review of world literature of studies of psychotropic drugs prior to 1993 that ‘double-blind is vulnerable to penetration and hence, to bias,’ placing all previous psychopharmacological research in question.

To try to rectify previous research on the efficacy of psychotropic drugs, Fisher and Greenberg suggest greater complexity in research design, the use of active placebo, various crossover designs, ‘conventional drug’ and ‘drug disguised’ groups, dose ranging, and the testing of drugs in opposing contexts, i.e., in an atmosphere of optimism or pessimism. They concede that current regulations preclude a drug disguise scheme for example, and agree that a completely foolproof design is not likely ever to be contrived. Personalities of patients and researchers were a factor in correct guessing of placebo or drug. While overall studies showed agreement among evaluators, rater differences were noted. Patient noncompliance was another confounding variable.

9. Ethics

As early as in the 1960s, ethical considerations about clinical trials involving randomization, double-blind methodology, and placebo controls were raised. Benevolent paternalism had been a prevalent ethical guideline since antiquity. At worst, it led to egregious exploitation of patients. In the late 1950s and 1960s, H. K. Beecher, in several watershed papers, cited instances of unethical and unscientific proceedings by researchers. He recommended that patients must give ‘informed consent.’ The US Public Health Service responded by proposing the establishment of Institutional Review Boards. The National Institutes of Health mandated Internal Review Boards, and the Food and Drug Administration, published in 1970, rules for ‘adequate and well controlled clinical evaluation’ (Shapiro and Shapiro 1997).

Beecher’s papers helped move medicine from its tradition of paternalism to informed consent and patient autonomy. However, many in the medical establishment complained that the patient–physician dyad was being fractured by a mass of lawyers, judges, journalists, philosophers, and legislators. And that inevitably, bureaucracy was stifling the practice of medicine.

Ethicists had entered discussions in the 1970s, questioning placebos, the double-blind procedure, and the randomized clinical trial. Sissela Bok and Beth Simmons took strong stands opposing the use of placebos without informed consent. Others considered the use of placebos to be unethical in practice but permissible in research, to the point of not informing subjects of the research design. Some proposed that it was unethical not to carry out a placebo controlled, double-blind study. Still others concluded that since placebos have shown positive therapeutic effects using a placebo as control does not deprive patients of all treatment.

Questions have also been raised about how in-formed is informed consent in general, and especially if the subjects are cognitively or psychologically impaired? Attempts to resolve some of the issues are also being made. Where an effective treatment exists but a new one is being tried, the potency of the established treatment, as well as the severity of the patient’s condition, are considered. As part of the experimental design, patients are moved from experimental or placebo cohorts to conventional treatment if their conditions deteriorate.

In 1978, while the rest of the Western scientific community moved towards research criteria of randomization, double-blind, and placebo controls in clinical trials, Germany passed laws substituting proof of efficacy through clinical open trials or expert opinion as qualifications for acceptance and registration of drugs. Controlled studies were no longer considered as essential to proof of efficacy.

In the introduction to the Physician’s Desk Reference for Herbal Medicines, the editors say that in the absence of the ‘rigors of FDA review’ of these preparations, the editors are basing their guidelines on ‘the closest available analog to FDA approved labeling—the findings of the German Regulatory Authority’s watchdog agency commonly called Commission E’ … its conclusions represent the best expert consensus …’

Statisticians contend that uncontrolled studies tend to favor a positive outcome.

10. Summary

Scientific thinking, defined by uncertainty that drives observation, hypothesis formation, and further search, was a late arrival to medicine, which was still laboring under few restrictions, guided by elaborate theories, mysticism, superstition, tradition, and authority.

Scientific thinking met much resistance in the medical establishment, but open label clinical trials, clinical trials using placebo controls, and double-blind methods began and continued into the twentieth century. In the 1950s when it became possible for the first time to cure people, statisticians became essential partners in the research team (Glantz 1997). Methodological and ethical questions began to be raised in subsequent decades about contemporary scientific criteria. Patient activism had also contributed to the erosion of paternalism that afflicted research design, as well as clinical practice.

The future may render clinical trials obsolete. We may well have a database atlas of the brain and body that will enable the researchers to calculate and theorize about effects without the use of human subjects.

The reader is referred to Marks (1997) and Shapiro and Shapiro (1997) for more complete historical surveys, detail, and documentation.

Bibliography:

1. Basogliu M 1997 Double-blindness procedures, rater blindness, and ratings outcome: Observations from a controlled trial. Archives of General Psychiatry 54: 744–8
2. Beecher H K 1955 The powerful placebo. Journal of American Medical Association 159: 1602–6
3. Bystritsky A, Waikar S V 1994 Inert placebo vs. active medication: Patient blindability in clinical pharmacological trials. Journal of Nervous and Mental Disease 182: 485–7
4. Cohen A, Posner J 1995 A Guide to Clinical Drug Research. Kluwer Academic Publishers, Dordrecht, The Netherlands
5. Fisher S, Greenberg R P, Riter J A 1995 Placebo washout is not a meaningful part of antidepressant drug trials. Perpetual and Motor Skills 81: 688–90
6. Frank J D, Frank J B 1993 Persuasion and Healing. Johns Hopkins paperback edn. The Johns Hopkins University Press, Baltimore and London
7. Freston J W 1986 Dose ranging in clinic trials: Rationale and proposed use with placebo positive controls. American Journal of Gastroenterology 81: 307–11
8. Glantz S A 1997 Primer of Biostatistics, 4th edn. McGraw Hill, New York
9. Kuhn T S 1996 The Structure of Scientific Revolutions, 3rd edn. The University of Chicago Press, Chicago
10. Marks H M 1997 The Progress of Experiment: Science and Therapeutic Reform in the United States, 1900–1990. Cambridge University Press, Cambridge, MA
11. McDonald C J, Mazzuco S A 1983 How much of the placebo effect is really statistical regression? Statistics in Medicine 2: 417–27
12. Morin C, Calecchi C, Brink D 1995 How ‘blind’ are doubleblind placebo-controlled trials of benzodiazepine hypnotics? Sleep 18: 240–5
13. Niklson I A 1997 Factors that influence the outcome of placebocontrolled antidepressant clinical trials. Psychopharmacology Bulletin 33: 41–51
14. Report of the US Preventive Services Task Force 1996 Guide to Clinical Preventive Services, 2nd edn. Williams and Wilkins, Baltimore, MD
15. Shapiro A K, Shapiro E 1997 The Powerful Placebo: From Ancient Priest to Modern Physician. The Johns Hopkins University Press, Baltimore, MD
16. Thomson R 1982 Side effects and placebo amplification. British Journal of Psychiatry 140: 64–8