Lecture 9:Placebo outline Dr.Anna


Published on

1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Lecture 9:Placebo outline Dr.Anna

  1. 1. TOPICS FOR STUDY: Treatment Evaluation & Placebo EffectTEXT: Provided articles and Lecture Powerpoint 1. What is meant by Treatment Evaluation? - Definition: Impact of intervention on outcomes of interest - How can treatments be evaluated? - Through experiments 2. What is meant by the placebo effect? - Definition of placebo (Shapiro) - How is the placebo effect different from other therapies that rely on direct physical or chemical effects on disease processes? 3. Is the placebo effect psychologically or therapeutically inert? - Which factors influence the placebo effect? - Two proposed ideas to explain the effect: expectancy theory and classical conditioning 4. General features of the placebo effect - Long history and a wide range of interventions (herbs, procedures, preparations) - Treats a wide variety of physical and psychological conditions (e.g. pain, depression) - Not always pharmacologically inert - Placebo effect is non-specific - Can occur even with awareness of placebo procedure 5. How does the placebo effect work? - Catalyzing self-healing mechanisms within the patient - Body produces endorphins (endogenous opioids)
  2. 2. 6. What is the Nocebo effect? - Definition7. How to control best for placebo effect in treatment evaluations? - Randomized controlled trial (RCT) - The importance of randomization - Other Issues: Double-Blind Procedure, Informed Consent
  3. 3. DEPARTMENT OF MEDICAL HISTORYDepartment of medical historyPowerful placebo: the dark side of the randomised controlled trialTed J KaptchukThe placebo went through a dramatic metamorphosis inthe years after World War II as the double-blindedrandomised controlled trial (RCT) developed. Until1945 the placebo was a “morally” useful but innocuousmanagement tool without curative or symptomaticconsequences. By the time the double-blind randomisedcontrolled trial took form and began to establish itself,around 1955, the placebo was imbued with powerfultherapeutic effects and its ethical clinical use was moregenerally being questioned. In 10 years, the placebochanged from what was called the “humble humbug” toan entity with occult-like powers that could mimic potentdrugs. It may be that efforts to bring the precision ofscience into the evaluation of efficacy with the RCT hasits own form of confusion and darkness.Pre-RCT placeboWhen pre-World War II paternalistic ethics prevailed,informed consent was not a standard of care.1 Physicianswere generally comfortable with benevolent deception Wellcome Institute Library, Londonand a “polychromatic assortment of sugar pills” wasroutinely swallowed by patients.2,3 In 1903, RichardCabot (1868–1939), the eminent professor at HarvardMedical School, described the placebo’s persuasiveness.He was “brought up, as I suppose every physician is, touse placebo, bread pills, water subcutaneously, and otherdevices . . . How frequently such methods are used varies Richard Clarke Cabota great deal I suppose with individual practitioners, but Idoubt if there is a physician in this room who has not hypnosis, where the power of the imagination wasused them and used them pretty often . . . I used to give accepted.7 If there was a “medical” value for a placebo itthem by the bushels”.4 Cabot’s remarks are confirmed in was as an occasional diagnostic tool to separate imaginarymany other observations of medical practice. For “psychological” symptoms from real problems.8 In 1945,example, in 1807, Thomas Jefferson (1743–1826) an influential article stated that the placebo was apenned a description of what he called the “pious fraud” valuable device “to smooth [the patient’s] path”, whichand noted that, “one of the most successful physicians I “cannot harm and may comfort” patients, especially thehave ever known has assured me that he used more bread “ignorant . . . disappointed and displeased . . . hopeless,pills, drops of coloured water, and powders of hickory [and] incurable case[s]”.9 A 1954 Lancet article, entitledashes, than of all other medicines put together”.5 “The Humble Humbug”, depicted the swan song on this Peculiarly (at least in the current view), the bread pill old-fashioned understanding of the placebo; “a means ofwas generally thought to have no consequences other reinforcing a patient’s confidence in his recovery, whenthan giving patients (especially ignorant and complaining the diagnosis is undoubted and no more effectiveones) peace of mind. A medical dictionary published in treatment is possible; that for some unintelligent or1785 described the placebo as “calculated to amuse for a inadequate patients life is made easier by a bottle oftime, rather than for any other purpose”; a dictionary medicine to comfort their ego; that to refuse a placebo tofrom 1811 depicted it as “given more to please than to a dying incurable patient may be simply cruel; and that tobenefit the patient”; and until the 1950s medical decline to humour an elderly ‘chronic’ brought up on thedictionaries echoed this definition of an inactive and bottle is hardly within the bounds of possibility”.10innocent management contrivance.6 The main objectionsto this prevailing view can be found in the few RCT placebo effectsympathetic discussions of such unorthodox practices as In 1955, the modern biomedical concept of the placeboLancet 1998; 351: 1722–25 makes its first definitive appearance with the publicationCenter for Alternative Medicine Research, Beth Israel Deaconess of Henry Beecher’s (1904–1976), “The PowerfulMedical Center, Harvard Medical School, Boston, Massachusetts Placebo” in the Journal of the American Medical Association02215, USA (T J Kaptchuk OMD) (JAMA)11 Beecher, a distinguished medical researcher at1722 THE LANCET • Vol 351 • June 6, 1998
  4. 4. DEPARTMENT OF MEDICAL HISTORYHarvard Medical School, summarised and World War II the evaluation of new therapeutics wasmathematically presented a perspective that had been made by recognised leaders of the medical professiondeveloping in a few elite biomedical research centres whose judgments were based on clinical impressions, andsince 1946.12 Beecher used a proto-meta-analytic method on rare occasions, poorly controlled evidence.21 In anto aggregate the percentage of patients satisfactorily effort to impose an objective and scientific discipline ontorelieved by a placebo across 15 clinical trials. Of 1082 the extraordinary postwar expansion of medical research,patients, 35·2 (SD 2·2%) experienced therapeutic the components of the double-blind RCT were adoptedbenefits. This number, which Beecher called “average and coalesced in the years after the war.22 The majorsignificant effectiveness” did not measure the exact features of these innovations included blind assessmentmagnitude of improvement, (usually meaning a placeboalthough that is how most control), random assignment topeople have subsequently comparable groups, andinterpreted his paper. Beecher inferential statistics as aargued forcefully that placebos surrogate for determinism.23alleviated beyond psychological The postwar placebo effectproblems by citing evidence resulted from an almost sleight-that “these powerful placebo of-hand shift in the placebo’seffects . . . can produce gross operational meaning in the newphysical change”, which RCT model. Instead of an inert“include objective changes at sham given to individualthe end organ which may patients, the placebo becameexceed those attributable to the emblem for all the healingpotent pharmacological occurring in the disguised “no-action”.11 In an important treatment” arm of an RCT. Therevision of old orthodoxy, “placebo effect” encompassedBeecher considered this placebo all “nonspecific effects” that dideffect to operate regardless of not depend on the treatment inthe intelligence of a person. the active arm. The “powerful Beecher explicitly assumed an A green placebo? placebo” became a hodge-podgeadditive model of placebo of non-linear, difficult toeffects, “The placebo effect of active drugs is masked by quantify, remnants collected under the rubric of thetheir active effects . . . The total ‘drug’ effect is equal to dummy control of an RCT. Anything that threatened theits ‘active’ effect plus its placebo effect” (quotes in the fastidious detection of a predictable cause and effectoriginal).11 This premise took for granted that the active outcome was conveniently disposed of in a repositorydrug response results partly from a placebo effect and labelled the “placebo effect”. This new concept ofthat the placebo effect buried in the active arm is identical placebo was much larger both in meaning and powerto the placebo effect of the dummy treatment.13 The than its predecessor. It incorporated many contributors toplacebo was a single and stable “power” that behaved in a health outcomes such as natural history, routine medicalconsistent manner. The new placebo effect, the and nursing care, and the “art” of medicine that had onceoxymoron-like enigma of an effect produced by been clearly distinct from the deception of an inert pill.something that is inert came to haunt biomedicine. Stillto this day, extensive examination leaves scientists and Powerful placebo and acceptance of RCTphilosophers to conclude that “the placebo concept as Medical proponents of the RCT were under pressure topresently used cannot be defined in a logically consistent convince their colleagues of the RCT’s value.23 Fewway and leads to contradictions”.14 Conflicting physicians wanted randomly to assign treatment toexplanations—expectation, faith, classic conditioning, patients, forgo the individualisation of therapy, andanxiety relief, symbolic processes, patient-doctor withhold promising new therapies. Austin Bradford Hillrelationship, self-perceptions—vie for acceptance. The (1897–1991), the designer of the first randomised trial,placebo effect has attributes of a neo-mesmeric energy. It many years later confessed that he “deliberately left outis radically protean: placebo responders seem to react or the words ‘randomisation’ and ‘random samplingnot capriciously,15 and for the same disease the placebo numbers’ at the time because . . . I might have scaredeffect can vary between 0% to 100%, when compared them [collaborating physicians] off”.24 Physicians resistedwith identical drugs.16,17 The theurgic placebo can even be treating patients as so “many bricks in a column” and themore effective18,19 or, as noted by Beecher, completely “elimination of the responsibility of the doctor to get thereverse the outcomes of known powerful pharmacological individual back to health”.25 In the same issue of JAMAdrugs.20 as Beecher’s paper, another research team concluded, What happened in the 10 years between 1945 and “we seriously doubted whether the double blindfold1955, when the placebo shifted from insipid decoy to a technique was a valid method of distinguishing betweenmischievous genie that could trick the most discerning the effectiveness of analgesic agents”.26clinician? For elite researchers, the moral imperative for scientific method in therapeutic evaluation was critical. TheyRCT and creation of the modern placebo realised that the imprecision of standard methods was aWorld War II has been called the great divide in medical hazard for the health of nations. (The general public wasresearch and certainly, for the placebo, this is true. In mostly unaware of any problem until the much laterfact, the “powerful placebo” was born in the vortex of thalidomide tragedy). For these reformers, “the powerfulone of medicine’s most momentous transitions. Before placebo effect” became a major argument used toTHE LANCET • Vol 351 • June 6, 1998 1723
  5. 5. DEPARTMENT OF MEDICAL HISTORYpersuade the medical profession to accept the placebo- blind RCT became the instrument to prove its own self-controlled RCT. The greater the placebo’s power the created value system. This shift from emphasisingmore necessity there was for the masked RCT itself. outcomes to the purity of the means directly parallelsBecause of this ominous threat, which could distort the developments in medical ethics where “informedjudgment of even the most unbiased and conscientious consent” replaced “beneficence” as the pinnacle of theresearcher, one needed to adopt the scientific device of value system.placebo control which alone allowed a separation of real Presuppositions embedded in the new concept offrom false effects. An enhanced “placebo effect” came to placebo also helped implement new methods of usingserve a valuable scientific and rhetorical function of frequency statistics to make causal inferences. The criticalpersuading colleagues of the necessity of the RCT. assumption here was that the placebo effect was a A need to show the placebo’s power was monolithic effect which was present to the same degreeunderstandable. Understanding the phenomena itself was and same direction in both the treatment and dummyof much less consequence. How the active treatment arms. (Anomalies such as placebo with a larger effectworked was seen as important. The dummy side of a trial than the real drug or a placebo that could reversereceived inadequate attention, poor methodology, and a pharmacological activity, were conveniently overlooked aspriori assumptions. The archetypal example of this is was the possibility of verum and placebo beingBeecher’s original study with its many flaws, most of differentially effected by the context of the RCT or ofwhich have gone unnoticed until recently.27,28 For interacting.) For the emerging RCT model, the treatmentexample, Beecher did not mention, although he and dummy arm of trials were assumed to receive equalundoubtedly knew, that much of what was labelled a and independent amounts of this force; one could simplypowerful placebo effect was actually regression to the subtract the amount of placebo effect to determine themean, natural history, or concurrent interventions.27–29 presence (or absence) of specific drug effect. TheThere had already been some well-designed experiments possibility that the placebo effect could act differentiallywith two controls (a placebo arm and a “no-treatment in the two arms was discounted. This “assumption ofno-placebo” arm) which demonstrated no placebo additivity . . . enable[d] one to infer that the variabilitieseffect.30 These were not included in the study. Beecher within the treatments should have a . . . [random]retrospectively interpreted the two pre-World War II distribution.”35 Without the premise of a single placebotrials used in his study as having a placebo effect when effect, commonly used statistical procedures would bethe original investigators have confidently reported the confounded.results observed in the sham arm as being due to The placebo had value only as a comparison marker,spontaneous recovery31 or spontaneous variations.32 Also, the magnitude of its absolute power has been anin the calculation of the quantitative effect of the placebo, incidental question. Studies attempting mathematically toBeecher deliberately did not include in his calculation the quantify across trials the magnitude of the placebo effect,numbers of patients who got worse with placebo, like Beecher’s, can be counted almost on one hand.17,27,36–39although this effect was reported in several trials Beecher When their results conflict, as they do, they have beenanalysed.33 Inclusion of this group would have dissolved ignored or tolerated. Beecher’s vintage numbers are stillsignificant amounts of the placebo effect into equally routinely misquoted; they are preferable to any challengedistributed normal variability.27 In fact, the entire point of of their written assumptions.Beechers exercise to establish the “powerful placebo”was to demonstrate persuasively “that ‘clinical RCT and powerful placebo at 50 yearsimpression’ is hardly a dependable source of information The powerful placebo is a modern entity constructed inwithout the essential safeguards of the double unknown the shadow of the RCT. In the current RCT era, atechnique, the use of placebos also as unknowns, legitimate therapy must demonstrate an effect greaterrandomisation of administration . . . and mathematical than a decoy disguised as a real intervention. Yet, undervalidation of any supposed differences”.11 Accurate the rhetorical label of powerful placebo lies many richportrayal of the placebo effect was of less importance than contributions to health care. These include: nature takinginvoking it as a threat to scientific evaluation, the its course; regression to the mean; routine medical andelimination of which would be accomplished by the nursing care; regimens such as rest, diet, exercise, anddouble-blind RCT. The “new” placebo became both the relaxation; easing of anxiety by diagnosis and treatment;raison d’être for, and the sacrificial victim of, the masked the patient-doctor relationship; classic conditioning andRCT. learnt behaviours; the expectation of relief and the imagination; and the will and belief of both patient andRCT and placebo: the light and dark of a practitioner.partnership The placebo effect also includes another often-Elite medical reformers created a symbiosis between the overlooked consequence of research activity. It isRCT and the powerful placebo effect. A new gold modified by consequences due to the context of the RCTstandard was constructed to fit the new technical itself.40 Issues such as the method of recruiting patients,procedures. Until the RCT, medical therapy became manner of giving informed consent, procedures forlegitimate because of beneficial outcomes; after the RCT, blinding, vehicle of delivery (colour of pills, pills vsa medical intervention was only scientifically acceptable if injection), provider characteristics, provider verbalit was superior to placebo.34 No longer was it sufficient for attitudes, and physical setting of the environment havea therapy to work: it had to be better than placebo. For been insufficiently studied.the first time in history (outside of religious healing But each of the components of the placebo effect hasscuffles), method became more important than consequences in healing and may have a differentialoutcome.34 In a self-authenticating manner, the double- impact on each arm of an RCT. All the variables can1724 THE LANCET • Vol 351 • June 6, 1998
  6. 6. DEPARTMENT OF MEDICAL HISTORYcreate countless variations in outcomes that have 14 Gøtzsche PC. Is there logic in the placebo? Lancet 1994; 344: 925–26.undoubtedly contributed to the haphazard trail that the 15 Liberman RP. The elusive placebo reactor. Neuropsychopharm 1967; 5:placebo effect has traced. It is therefore essential, 557–66.whenever ethically and financially feasible, that research 16 Joyce CRB. Placebos and other comparative treatments. Br J Clinin medicine begins to disentangle the “non-specific” and Pharmacol 1982; 13: 313–18.“art-of-medicine” aspects of healing and therapeutic 17 Moerman DE. General medical effectiveness and human biology: placebo effects in the treatment of ulcer disease. Med Anthropol 1983;evaluation. Besides comparing a real intervention and 14: 3, 13-16.placebo, the inclusion of a third no-treatment no-placebo 18 Beecher HK, Keats AS, Mosteller F, Lasagna L. The effectiveness ofarm would be helpful to distinguish a perception of a placebo “reactors” and “non-reactors”. J Pharmacol Exp Ther 1953; 109: 393–400.“placebo” effect from the ordinary natural history of a 19 Dinnerstein AJ, Lowenthal M, Blitz B. The interaction of drugs withcondition.41 Each “non-specific” effect needs to be placebos in the control of pain and anxiety. Perspect Biol Med 1996; 10:disentangled, carefully varied, and systematically studied 103–117.under controlled conditions.40 Preliminary evidence 20 Wolf S. Effects of suggestion and conditioning on the action of chemical agents in human subjects—the pharmacology of placebos.concerning the examination of non-specific effects and J Clin Invest 1950; 29: 100–09.contextual effects of an RCT has been valuable, 21 Marks HM. The progress of experiment: science and therapeuticprovocative, and contributed to a more refined reform in the United State, 1900–1990. Cambridge: Cambridgeunderstanding of the internal and external validity of University Press, 1997. 22 Lilienfeld AM. Ceteris paribus: the evolution of the clinical trial.trials.42 Bull Hist Med 1982; 56: 1–18. Obviously, the double-blind RCT has meant a 23 Kaptchuk TJ. Intentional ignorance: a history of blind assessment intremendous improvement in research and subsequent medicine. Bull Hist Med 1998; in press.medical care. In the beginning, the RCT needed a 24 Hill AB. Suspended judgment: memories of the British streptomycin trial in tuberculosis. The first randomized clinical trial. Control Clinsimplistic neo-mesmeric placebo as a looming threat. It is Trials 1990; 11: 77–79.undoubtedly time that the “powerful placebo” be 25 Hill AB. The clinical trial. N Eng J Med 1952; 274:4 113–19.examined in all its myriad facets, otherwise medicine will 26 Batterman RC, Grossman AJ. J Am Med Assoc 1955; 159:17 1619–22.always have a limited perception of healing. At age 50, 27 McDonald CJ, Mazzuca SA, McCabe GP. How much of the placebothe RCT is ready to go through a mid-life crisis and face “effect” is really statistical regression? Stat Med 1983; 2: 417–27. 28 Kienle GS. Der sogenannte placeboeffectk: illusion, fakten realität.its dark side. Stuttgart: Schattauer, 1995. 29 Kienle GS, Kiene H. Placebo effect and placebo concept: a criticalAcknowledgments methodological and conceptual analysis of reports on the magnitude ofResearch for this paper was partly supported by grants from the John E the placebo effect. Altern Ther 1996; 2: 39–54.Fetzer Institute, NIH grant U24 AR3441, and the Kenneth J 30 Tyler DB. The influence of a placebo, body position and medicationGermeshausen Foundation. The author also wishes to thank David of motion sickness. Am J Phys 1946; 146: 458–66.Eisenberg, Howard Brody, Daniel Moerman, Adriane Fugh-Berman, 31 Diehl HS. Medicinal treatment of the common cold. J Am Med AssocHerman Engelbart, and David Stone for advice, discussion and feedback, 1993; 101: 2042–49.and thank Marcia Rich and Janet Walzer for editorial assistance. 32 Gold H, Kwit NT, Otto H. The xanthines (theobromine and aminophylline) in the treatment of cardiac pain. J Am Med Assoc 1937; 108 (26): 2173–79.References 33 Gay LN, Carliner PE. The prevention and treatment of motion1 Faden RR, Beauchamp TL. A history and theory of informed consent. sickness. Bull Johns Hopkins Hospit 1949; 84: 470–87. Oxford: Oxford University Press, 1986. 34 Sullivan MD. Placebo controls and epistemic control in orthodox2 Anon. Placebos. Med Record 1885; 27: 576–577. medicine. J Med Philos 1993; 18: 213–31.3 Rosenberg C. The practice of medicine in New York a century ago. 35 Kempthorne O. Why randomize? J Stat Plan Inf 1977; 1: 1–25. Bull Hist Med 1967; 41: 223–53. 36 Evans FJ. The placebo response in pain reduction. Adv Neurology4 Cabot RC. The use of truth and falsehood in medicine: an 1974; 4: 289–96. experimental study. Am Med 1903; 5: 344–49. 37 Roberts AH, Kewman DG, Mercier L, Hovell M. The power of5 Jefferson T. The writings of Thomas Jefferson. v. 9 PL Ford (ed). nonspecific effects in healing: implications for psychosocial and New York: G P Putnam’s, 1898. biological treatments. Clin Psychol Rev 1993; 12: 375–91.6 Shapiro AK. Semantics of the placebo. Psych Quart 1968; 42: 653–95. 38 Porter DR, Capell HA. The ‘natural’ history of active rhematoid7 Tuke DH. Illustrations of the influence of the mind upon the body in arthritis over 3–6 months: an analysis of patients enrolled into trials of health and disease designed to elucidate the action of the imagination. potential disease-modifying anti-rhematic drugs, and treated with Philadelphia: Henry C Lea’s, 1884 [1872]. placebo. Br J Rheumatol 1993; 32: 463–66.8 Shideman DE, Beckman H. Comments on treatment. Wisconsin Med J 39 McQuay H, Carroll D, Moore A. Variations in the placebo effect in 1958; 57: 456. randomized controlled trials of analgesics: all is blind as it seems. Pain 1995; 65: 331–35.9 Pepper OHP. A note on the placebo. Am J Pharm 1945; 117: 409–12. 40 Kleijnen J, de Craen AJM, van Everdingen J, Krol L. Placebo effects10 Anon. The humble humbug. Lancet 1954; ii: 321. in double-blind clinical trials: a review of interactions with11 Beecher HK. The powerful placebo. J Am Med Assoc 1955; 159:17 medications. Lancet 1994; 344: 1347–49. 1602–1606. 41 Ernst E, Resch KL. Concept of true and perceived placebo effects.12 Conference on therapy. The use of placebos in therapy. N Y J Med BMJ 1995; 311: 551–53. 1946; 17: 722–27. 42 Kaptchuk TJ, Eisenberg DM. Legitimate evidence: randomized or13 Spilker B. Guide to clinical trials. New York: Raven Press, 1991. otherwise. Advances 1997; 13: 48–51.THE LANCET • Vol 351 • June 6, 1998 1725
  7. 7. Rajagopal The placebo effect special article Psychiatric Bulletin (20 0 6), 30, 185^18 8 S U N D A R A R A J A N R A J A G O PA L The placebo effectThe placebo effect is a fascinating phenomenon in clinical superficial procedures (e.g. skin incision, burr hole) arepractice. Studies have shown that there is a significant performed without the actual surgery.placebo effect in a wide range of medical conditions Placebo equivalents are also employed in comple-including psychiatric disorders. This article looks at the mentary medicine. For example, sham acupuncturebackground of the placebo effect, defines the common consists of needles placed at non-acupuncture points. Aterms used, describes the various hypotheses that have recent study (Linde et al, 2005) showed that realbeen put forward to explain this seemingly inexplicable acupuncture was no more effective than sham acupunc-phenomenon and also covers the issue of using placebos ture in reducing migraine headaches, although bothin research trials, highlighting the important ethical interventions produced benefits compared with a waitingdilemmas involved. Throughout, specific emphasis is given list control.to psychiatry. Why does the placebo effect occur?BackgroundThe term placebo is derived from the Latin verb ‘placare’, Natural remission theory‘to please’. The American anaesthetist Henry K. Beecher This states that the improvement that occurs with the(1955) coined the term ‘placebo effect’. He reported that, administration of placebo is coincidental and would haveon average, about a third of patients with a range of occurred even without it. This theory explains the bene-conditions improved when they were given placebos. This ficial effects of placebo in short-lived conditions likesubsequently led to the development of placebo- common cold, headache, etc, but does not satisfactorilycontrolled trials, whereby a new drug is said to have explain why even patients with chronic conditions such assignificant benefit only if it shows superiority over hypertension or schizophrenia show improvement withplacebo. The placebo effect has also been a source of placebo.recent interesting debate in psychiatry with some An allied hypothesis is the ‘regression to the mean’claiming that a considerable proportion of benefit from theory. Regression to the mean is a statistical concept;antidepressant medication derives from the placebo according to this, if an initial test result is extreme andeffect (Kirsch & Sapirstein, 1998), whereas others if the test is repeated, statistically there is a greater like-(Leutcher et al, 2002) have stressed that response to lihood for the second result to be closer to the meanplacebo and to antidepressants involves distinct biological than for it to be more extreme than the first result.mechanisms. Usually only patients who are significantly unwell (e.g. depression score above a certain cut-off point) are eligible to enter a trial. Hence, at follow-up they are moreDefinitions likely to show an improvement (depression score beingIn general, a placebo is an inert substance that has no closer to the mean than the first score) than a deteriora-inherent pharmacological activity. It looks, smells and tion, owing to regression to the mean (McDonald et al,tastes like the active drug with which it is compared. An 1983).‘active placebo’ is one that has its own inherent effectsbut none for the condition that it is being given for(e.g. use of atropine as the control drug in trials of Classical (Pavlovian) conditioningtricyclic antidepressants). A placebo need not always be In the original experiment of Pavlov, the dog salivated atpharmacological. It could be procedural, for example, the sound of the bell even without any food, as it hadsham electroconvulsive therapy (ECT), where the patient previously been conditioned to expect food by pairingis anaesthetised but not given ECT. Surgical placebo is a the bell with food. Food is the unconditioned stimulus,procedure where the patient is anaesthetised and salivation owing to food is the unconditioned response; 185
  8. 8. Rajagopal The placebo effect the bell is the conditioned stimulus and salivation owing particularly dopamine, in placebo effects on mood and to the bell is the conditioned response. behaviour. special In a similar manner, patients who have had past article experience of getting better with active medication may be conditioned to anticipate improvement by any subse- Pattern of placebo improvement quent prescription, including placebo. Using the classical Among psychiatric disorders, the placebo effect has been conditioning analogy, the active medication is the most extensively studied in depression. ‘Pattern analyses’ unconditioned stimulus, improvement owing to active have shown that the improvement as a result of placebo medication is the unconditioned response, the placebo is in depression tends to be abrupt, occurs early in treat- the conditioned stimulus, and improvement owing to ment and is less likely to persist (Quitkin et al, 1991), placebo is the conditioned response. whereas improvement in response to antidepressants tends to be gradual, occurs later and is more likely to Other psychological factors persist. Even among patients apparently responding to the active drug, if the pattern of improvement is consis- Patient expectations are important in determining the tent with a placebo response (i.e. abrupt and early), the placebo effect. Treatments that are perceived as being improvement tends to be short-lived. more powerful tend to have a stronger placebo effect Stewart et al (1998) investigated whether they could than those that are perceived to be less so. Thus, placebo predict relapse of depression from the initial pattern of injections have more effect than oral placebos, capsules response. Patients who had responded to treatment with are perceived as being stronger than tablets, bright- fluoxetine for 12^14 weeks were then randomly allocated coloured placebos are more effective than light-coloured to continuation/maintenance treatment for 50 weeks ones larger placebos have more effect than smaller ones, with either placebo or fluoxetine. Those patients who had and two placebos have more effect than one. Also, the shown a placebo pattern of improvement during the status of the treating professional is directly related to initial fluoxetine phase relapsed in a similar manner the placebo effect. The same compound has been found whether they continued on fluoxetine or were switched to be more powerful if it is branded than when it is to placebo, but patients who had shown a true drug unbranded (Branthwaite & Cooper, 1981). pattern of improvement relapsed more if they were In a novel study, Benedetti et al (2003) examined the switched to placebo in the maintenance phase. This study impact of the patient’s awareness that they are having a adds strength to the hypothesis that, even among drug certain treatment administered/withdrawn on the responders, only a certain proportion will benefit from outcome. They studied three treatments in three maintenance treatment. groups of patients - intravenous morphine for post- Hrobjartsson & Gotzsche (2001) conducted a major thoracotomy pain, intravenous diazepam for post- systematic review of placebo-controlled trials involving thoracotomy anxiety and stimulation of the subthalamic 40 clinical conditions, including hypertension, asthma, nucleus for idiopathic Parkinson’s disease. In each group, pain, depression, schizophrenia, anxiety and epilepsy. some patients were informed of the fact that they were They concluded that placebos tended to have no signifi- receiving the treatment (e.g. by a doctor administering cant effects on binary outcomes, and possibly had small the injection) but others were not aware as they received beneficial effects on continuous subjective outcomes and an infusion from an automatic pre-programmed machine. in the treatment of pain. In all the groups, the efficacy of the respective inter- ventions was greater when the patient was aware of the procedure than when they were not. Similarly, being Use of placebos in clinical trials aware that a treatment was being withdrawn worsened It is generally accepted that a double-blind randomised the symptoms much more than when the treatment was controlled trial (RCT) is the best research method to withdrawn without the patient’s knowledge. From a study the efficacy of clinical interventions. psychiatric point of view, neither the hidden administra- However, the use of placebos for conditions for tion nor hidden withdrawal of diazepam had any signifi- which effective treatments are already available raises an cant positive or negative effect respectively but the open important ethical question. Should a new treatment be administration of diazepam improved anxiety symptoms compared with an established treatment or should it only and open withdrawal worsened them. have to demonstrate superiority over placebo in order to be accepted as another effective treatment? Rothman & Michels (1994) have criticised the use of placebo- Role of endogenous opioids controlled trials to test new drugs for conditions with In a systematic review, ter Riet et al (1998) concluded potentially irreversible consequences, such as oncho- that endogenous opioids (e.g. endorphins) play a cerciasis and rheumatoid arthritis, when established significant role in mediating placebo-induced analgesia. treatments for these conditions already exist. Previous studies had shown that placebo-induced Death by suicide is associated with major psychiatric analgesia is partially reversed by administering the opioid disorders such as depression, and the use of placebo- antagonist naloxone (Grevert et al, 1983). There is also controlled trials to test the efficacy of new drugs is growing interest in the role of neurotransmitters, fraught with ethical issues. 186
  9. 9. Rajagopal The placebo effect Another important question involves the masking of to the side-effects investigated during RCTs. Thus, side-the double-blind trials. Margraf et al (1991) reported that effects reported by patients on placebo may be a reflec-a majority of patients in a double-blind study of tion of pre-existing or spontaneously occurring symptoms specialalprazolam v. imipramine v. placebo could correctly guess rather than being placebo-induced. Similarly, RCTs may be articlewhether they were on an active drug or placebo. In overestimating the side-effects (especially the non-addition, the ‘masked’ assessors were even able to specific ones) of active drugs.distinguish between the two active drugs. However, the nocebo effect is not purely psycholo- Informed consent entails the patients being made gical. It has been shown that nocebo hyperalgesia (i.e. anaware that they will be receiving either an active drug or increase in pain as a result of placebo) is mediated byplacebo. Therefore, it may not require more than just cholecystokinin and is abolished by the cholecystokininmonitoring one’s side-effects closely to accurately deter- antagonist proglumide (Benedetti et al, 1997). In amine whether one is on active medication or placebo. systematic review of double-blind RCTs comparing fluox- In addition to ethical issues, RCTs with a placebo etine and placebo, Casper et al (2001) found similar ratescontrol group have other limitations. RCTs only demon- of placebo response in men and women but slightly morestrate statistical significance. If the sample size is very nocebo effects in women.large, even if the difference in clinical outcome betweenthe two groups is small and clinically insignificant, it maybe detected as being significant by the statistical test. Conclusions In any RCT, the placebo is made by the manufacturer Despite half a century having passed since its inclusion inof the active drug. Hence, placebos used in one study will modern medicine, the placebo effect is still poorlybe different in form (size, shape, tablet/capsule, etc.) understood. Beecher’s (1955) original study, whichfrom those used in another study, depending on the form showed an overall average placebo response of 35%, hasof the active drug. This may account for the wide varia- been strongly criticised for major methodological short-tion in placebo response observed for the same condi- comings (Kienle & Kiene, 1997).tion. All discussion regarding placebos is based on the assumption that they are inert. But are they really so? Placebos are generally referred to as ‘sugar pills’; sugar isSide-effects of placebos not chemically inert. Similarly, the tablet coating or the capsule covering are not inert. Hence, the possibility thatWhen a placebo produces prominent side-effects it is the ‘inert’ chemical in the placebo may be relevant to theknown as a ‘nocebo’. The term ‘nocebo effect’ encom- condition being studied should not be dismissed.passes the negative consequences resulting from the Whatever the reasons for the placebo effect, theadministration of a placebo. In placebo-controlled studies most important message for clinicians is that just becauseof psychotropic drugs, the placebos tend to cause a someone responds to a placebo does not mean that thesimilar range of side-effects as the active drugs but initial ailment for which they sought help was false.usually with a much lower incidence rate. Non-specificside-effects, such as headache and nausea, tend to bemore common than more specific ones such as acutedystonia or QT prolongation. ‘Placebo sag’ refers to the Declaration of interestattenuation of the placebo effect with repeated use None.(Peck & Coleman, 1991). There are historical reports ofplacebo dependence (Vinar, 1969). The nocebo effect clearly illustrates the role of Referencespatient expectations in perceived side-effects. Usually BEECHER, H. K. (1955) The powerful differences in placebo responses ofpatients included in trials of psychotropic medication have placebo. JAMA, 159,1602-1606. patients withmajor depressive disorder.already received previous treatment with active medica- Biological Psychiatry, 15,158-160. BENEDETTI, F., AMANZIO, M.,tion in the past, as most major psychiatric disorders tend CASADIO, C., et al (1997) Blockade of GREVERT, P., ALBERT, L. H. &to follow a chronic course. Hence, even if they are given nocebo hyperalgesia by the GOLDSTEIN, A. (1983) Partial cholecystokinin antagonist proglumide. antagonism of placebo analgesia byplacebo this time, they may anticipate side-effects similar Pain, 71,135-140. naloxone. Pain, 16,129-143.to those that they experienced when they were receivingtreatment with the active drug. Also, patients may be BENEDETTI, F., MAGGI, G., LOPIANO, L., et al (2003) Open versus hidden medical HROBJARTSSON, A. & GOTZSCHE, P. C.influenced by the list of side-effects experienced by their treatments: the patients’ knowledge (2001) Is the placebo powerless? Anfriends or relatives who have received such treatment in about a therapy affects the therapy analysis of clinical trials comparing outcome. Prevention & Treatment, 6. placebo with no treatment. Newthe past, and by the list of potential side-effects EnglandJournal of Medicine, 344, http://content.apa.org/journals/pre/described by the researchers before obtaining informed 6/1/1 1594-1602.consent. BRANTHWAITE, A. & COOPER, P. (1981) KIENLE, G. S. & KIENE, H. (1997) The Just as doubts have been cast on the beneficial Analgesic effects of branding in powerful placebo effect: fact oreffects of the placebo, so have questions been raised treatment of headaches. BMJ, 282, fiction? Journal of Clinicalabout the nocebo effect. Even healthy people who are 1576-1578. Epidemiology, 50,1311-1318.not taking any medication have been shown to have a CASPER, R. C.,TOLLEFSON, G. D. & KIRSCH, I. & SAPIRSTEIN, G. (1998)high prevalence of a range of symptoms which are similar NILSSON, M. E. (2001) No gender Listening to Prozac but hearing 187
  10. 10. Rajagopal The placebo effect placebo: a meta-analysis of studies? Journal of Consulting and during placebo treatment. American with fluoxetine or placebo. Archives antidepressant medication. Clinical Psychology, 59,184-187. Journal of Psychiatry, 148,193-196. of General Psychiatry, 55, 334-343. Prevention & Treatment, 1, http:// McDONALD, C. J., MAZZUCA, S. A. & ROTHMAN, K. J. & MICHELS, K. B. Ter RIET, G., De CRAEN, A. J., special content.apa.org/journals/pre/1/1/2 McCABE, G. P., Jr (1983) How much of (1994) The continuing unethical use of De BOER, A., et al (1998) Is placebo article LEUCHTER, A. F., COOK, I. A.,WITTE, E. the placebo‘effect’ is really statistical placebo controls. New England Journal analgesia mediated by endogenous A., et al (2002) Changes in brain regression? Statistical Medicine, 2, of Medicine, 331, 394-398. opioids? A systematic review. Pain, 76, function of depressed subjects during 417-427. 273-275. STEWART, J.W., QUITKIN, F. M., treatment with placebo. American PECK, C. & COLEMAN, G. (1991) McGRATH, P. J. (1998) Use of pattern VINAR, O. (1969) Dependence on a Journal of Psychiatry, 159,122-129. Implications of placebo theory for analysis to predict differential relapse of placebo: a case report. British Journal LINDE, K., STRENG, A., JURGENS, S., clinical research and practice in pain remitted patients with major of Psychiatry, 115,1189-1190. et al (2005) Acupuncture for management. Theoretical Medicine, 12, depression during1year of treatment patients with migraine. JAMA, 293, 247-270. 2118-2125. QUITKIN, F. M., RABKIN, J. G., Sundararajan Rajagopal Consultant Psychiatrist, South London and MARGRAF, J., EHLERS, A., ROTH,W.T., STEWART, J.W., et al (1991) Maudsley NHS Trust, Adamson Centre for Mental Health, StThomas’ Hospital, et al (1991) How ‘blind’are double-blind Heterogeneity of clinical response London SE17EH, e-mail: Sundararajan.Rajagopal@slam.nhs.uk 188