July 2000 · Vol. 49, No. 7
The Placebo Response
Recent Research and Implications for Family Medicine
Howard Brody, MD, PhD
The placebo response is commonly invoked as a factor in the therapeutic relationship between the family physician and the patient, but important recent literature can be difficult for family physicians to access. Coordinated interdisciplinary research into the placebo response as it occurs in primary care settings is lacking.
Although there is controversy about the nature and scope of the placebo response, important suggestions are emerging about its psychological mechanisms (expectancy and conditioning) and the biochemical pathways that act as psychosomatic linkages (endorphins, catecholamines and cortisol, psychoneuroimmunology). The available research justifies interventions by family physicians that maximize the placebo response in everyday patient encounters. These include the sustained partnership approach, working with patients on the narratives they construct to explain illness, listening to patients, providing them with satisfactory explanations, expressing care and concern, and enhancing their sense of control. Notable opportunities exist for family medicine investigators to expand the understanding of this phenomenon.
endorphins. (J Fam Pract 2000; 49:649-654)
Since the placebo response was last reviewed for family physicians1 new research has added to our basic knowledge of this phenomenon, but placebo research remains in its infancy. Outcome studies have shown how various physician behaviors and practice styles can enhance the health status of patients. Seldom, however, are these 2 bodies of research carefully integrated to show the clinician how to best use the science of the placebo response to improve patient outcomes. Such integration is possible and would be useful for family physicians.
The science of the placebo
Two recent findings highlight the continued controversy over the placebo response. The apparent importance of the placebo response was recently emphasized by the ethical debate over the use of sham surgery control groups in studies of fetal cell brain implants for intractable Parkinson’s disease.2,3 The need for a sham group and the ethical question of whether exposing subjects to this risk is warranted arises because subjects receiving the sham procedure typically exhibit marked improvements in their Parkinson’s symptoms for up to 6 months and are indistinguishable from patients given the active treatment. This improvement does not seem to be due to either the natural history of the disease or observer bias.
However, the preliminary findings from a carefully conducted meta-analysis4 failed to demonstrate any consistent sizeable benefit from receiving a dummy treatment in a randomized blinded trial when another nontreatment group was included with which the sham-treatment group could be compared. These findings do not exclude the possibility that the psychosocial aspects of the physician-patient encounter may exert independent healing effects, but they do cast doubt on the placebo response as traditionally conceived-the so-called power of the sugar pill.
Although much controversy remains, the bulk of the literature supports the proposition that the placebo response is real and potentially powerful, and some of its underlying mechanisms are becoming better understood.
Defining “placebo” and “placebo response” in a logically coherent fashion is difficult, perhaps impossible.5,6 Terms commonly used in medical definitions, such as “inert” and “nonspecific,” range from useless to misleading. Ideally, a definition would be precise enough to clearly demarcate the phenomenon, yet open enough to include a range of possible empirical hypotheses on cause and mechanisms. Terminologic problems, among other factors, have led some thoughtful family physicians to suggest that we eliminate “placebo effect” from our lexicon.7,8
Assuming that the placebo response falls under the general heading of psychosomatic processes, how wide a range of such processes do we want to admit under the definition? At one extreme, the placebo response can be defined in a very narrow way as bodily changes that occur following the administration of a dummy treatment. This type of placebo response is of little interest to family physicians in daily practice. This definition appeared in a recent review in which the authors claimed that there is no scientific evidence that the placebo response exists.9 Since they insisted on the ubiquity and robustness of psychosomatic processes generally, they leave themselves with a conundrum-why should the mind-body connection be so powerful except where it would work by giving dummy medications?
I argue for a broader definition: The placebo response is a change in the patient’s health or bodily state that is attributable to the symbolic impact of medical treatment or the treatment setting. By this definition, a placebo response would be predicted whenever a conscious patient engages in any form of medical encounter (or self-treatment activity). But the definition explicitly excludes from consideration mind-body changes that are not directly related to medical treatment. This definition highlights why family physicians should be particularly interested in the placebo response. Also, since bodily change can be negative as well as positive (nocebo effects),10 it warns us that if we fail to understand the placebo response, we may unwittingly do harm. For example, a frequently quoted study11 showed that expressing diagnostic uncertainty regarding common nonspecific symptoms can adversely affect symptom resolution.
Frequency of Placebo Responses
The first modern review of the placebo effect made the often-repeated claim that an average of approximately one third subjects will respond to placebos.12 It was later shown that this average, even if accurate, masks an extremely wide variation in placebo response rates among studies.13 Roberts and colleagues14 recently reviewed the original uncontrolled trials of therapies introduced with great fanfare in past years and now universally dismissed as medically useless. They found a 70% average rate of “excellent” and “good” responses to those treatments before properly designed double-blind trials were conducted. They concluded that in settings of heightened expectancy, the placebo response rate could be much higher than previously thought.
Attention has recently turned to meta-analyses of the placebo response.9,13,14 One sophisticated analysis of placebo-controlled studies with more than 2 arms suggests that so-called “nonspecific” effects may be substantial and may be either synergistic or antagonistic toward “specific” effects, so the merely “additive” model of the placebo and drug effects, implicit in the design of most randomized controlled trials, is overly simplistic.15
According to a meaning model,16 a positive placebo response is most likely to occur when the meaning of the illness experience is altered in a positive direction. A positive change in meaning occurs when one or more of 3 things happens: The patient feels listened to and receives a satisfactory, coherent explanation of his illness; the patient feels care and concern from those around him; and the patient feels an enhanced sense of mastery and control over his symptoms. Because the meaning that we attach to events in our lives often hinges on the stories we construct about those events, this model helps explain the importance of narrative in medicine.17,18
Support for this model comes from several directions. In a psychology laboratory, a group of subjects told about the potentially healthful effects of brief hand immersion in ice water had better pain tolerance than a group briefed on the dangers of cold immersion, and both groups were significantly different from control subjects given a neutral message.19 In clinical research, the informed consent process is often thwarted by the “therapeutic misconception,” the tendency of subjects to believe that they are getting medical care aimed at their personal medical problems despite being told that their treatment is selected by the research protocol.20 The therapeutic misconception might enhance the placebo response and make it more difficult to show that an active drug is superior to placebo.
Much of the early literature on the placebo response was devoted to identifying a “placebo personality type,” in the vain hope that such subjects could be eliminated from clinical trials so as not to confuse the results with placebo responses. In general, this early literature21,22 concluded that no such personality type exists and that the placebo response is more a contextual situational phenomenon than an enduring personality trait.
One important exception to this general conclusion is highlighted in a superb recent review of the placebo literature. Fisher and Greenberg23 argue that the “acquiescent” personality type is more likely to experience positive placebo responses. At first misunderstood as passivity, the label “acquiescent” is now better understood as an active coping mechanism. This type of person deals with adversity by developing positive relationships with others and using them as a healing resource. This suggests the importance of continuity of care and effective interpersonal relationships in producing positive placebo responses as part of daily practice.
Theories of Placebo Action
Experimental data currently support 2 general theories about the placebo response: expectancy and conditioning.22,24 The first proposes that bodily changes occur to the extent that the subject expects them to; the second, that bodily changes occur when the subject is exposed to a stimulus that has been linked in the past to active processes that produce the change.
A classic study of expectancy25 showed subjects unable to distinguish the pharmacologic properties of amphetamines and chloral hydrate when they had no expectancy cues for the drugs they were receiving. A recent study of conditioning26 showed that young asthmatics, exposed over time to a vanilla aroma at the same time as they used their metered dose inhalers, demonstrated objective bronchodilation when later exposed only to the vanilla.
We may view these theories, taken in conjunction with the meaning model, as complementary not competing. Conditioning theory may explain how the past experiences of the patient affect the meaning attached to present events, while expectancy focuses on how the patient’s thoughts of the future may influence therapeutic outcomes.
Neither expectancy nor conditioning theory explains the process by which the altered meaning in the patient’s mind causes a change in bodily status. The complete answer to the neuroanatomy and biochemistry of the placebo response is unknown. The most interesting research has focused on 3 pathways: endorphins, cathecholamines and cortisol, and psychoneuroimmunology. All are capable of altering bodily symptoms, and all are known to be connected intimately with the individual’s emotional and cognitive state.
An elegant, recent, 12-armed trial conducted in a psychology laboratory27 compared 2 aspects of placebo response to pain: expectancy versus conditioning, and conditioning with morphine versus conditioning with ketorolac. The results showed that placebo analgesic effects produced by expectancy could be reversed by naloxone, suggesting that these might be endorphin mediated. Similarly, placebo conditioning with morphine could be reversed by naloxone, but the placebo response was not naloxone-reversible when ketorolac was used as the initial stimulus. The authors concluded that the biochemical pathway for placebo conditioning might depend on the biochemical pathway activated by the active drug.
Future Research Directions
My brief review suggests a number of requirements for advancing knowledge of the placebo response. The greatest advances will come from multidisciplinary research teams set up to integrate qualitative measures of meaning and perception with quantitative measures of biochemical mediators and bodily change.
Intriguing studies conducted in psychology laboratories27 must be replicated and expanded in the less controlled environment of clinical practice.
Creative use of brain imaging techniques, such as positron emission tomography and functional magnetic resonance imaging may be used to identify which neuroanatomical centers’ activation correlates with placebo responses.
Carefully designed meta-analyses may increase our understanding of selected facets of the placebo response.
Integrating Science and Practice
The [ Figure] suggests how the scientific elements reviewed here can be interwoven into a causal account of the placebo response that can guide physician behavior. Although all elements are of interest for research, the meaning model is most helpful for suggesting therapeutic interventions, since it describes those parts of the process that the physician can most easily influence.
This review suggests a number of therapeutic strategies for office practice that many sensitive and astute clinicians are already using. The available evidence suggests that although we do not fully understand why or how these strategies work, they are relatively inexpensive and notably nontoxic, so no barriers exist to their employment.
Although the focus here is on specific techniques, the importance of the background rituals of medical practice, such as listening to the heart with a stethoscope and writing a prescription, should not be forgotten. History and culture have imbued these rituals with multiple layers of meaning, and their simple predictability may bring great comfort to a frightened patient.
Perhaps the best single concept of how to structure one’s interactions to maximize the placebo response is that of the “sustained partnership” between physician and patient.28 Sustained partnership consists of 6 physician characteristics:
Interest in the whole person
Knowing the patient over time
Caring, sensitivity, and empathy
Viewed as reliable and trustworthy by the patient
Adapts medical goals of care to the patient’s needs and values
Encourages the patient to participate fully in health decision making
These 6 elements were selected because empirical studies link them to superior health outcomes in practice settings. Relating to patients in this way leads to measurable improvements in the patient’s health and an ethically sound (as well as cost-effective) practice. Sustained partnership is a nondeceptive mode of practice, so the ethical problem created by the use of sugar pills and other dummy remedies never arises-the placebo effect is present without the placebo.29
The meaning model suggests that physicians will obtain the best therapeutic results if they listen carefully to patients’ accounts of their illness, offer explanations that fit with patients’ world-views, express care and concern, and help patients feel more in control of their symptoms (or of their lives, despite the persistence of their symptoms). This will happen more often if the clinician spends sufficient time with the patient, so it is not surprising that good health outcomes correlate with the length of the primary care visit.30,31 Family physicians must act as advocates for their patients with today’s financial managers by bringing these data forcefully to the managers’ attention.
Of the different elements of meaning, perhaps the most challenging to achieve when treating chronically ill, somatizing, or depressed patients is the heightened sense of mastery and control. Malterud32 has summarized an encounter plan to deal with one difficult population, women presenting with pelvic complaints without detectable organic cause, suggesting 4 basic questions:
What do you most of all want me to do for you today?
What do you yourself think is the reason for your problem?
What do you think I should do about your problem?
What have you found so far to be the best way to manage your problem?
Malterud points out that previous physicians, by approaching these patients in “rule out organic causes” fashion, have firmly cemented the idea of victimhood in the minds of these women. Malterud’s strategy is designed to make each encounter an exercise in seeing oneself as potentially worth listening to, potentially wise about one’s bodily needs and their management, and thus potentially powerful. Physicians can achieve progress with some of the most difficult patients when they curb the urge to supply a diagnosis and therapeutic plan before the patient has a chance to speak.
A more formal strategy has been suggested for family counseling that seems readily applicable to daily practice.33 This entails asking at each visit, “How much is [the symptom or problem] in control of your life, and how much are you in control?” Patients can readily assign a percentage score to this variable. It is not unusual for a suffering patient to state at the first visit in which this question is asked that the problem controls 80% of his life and he feels in control of only 20%. In keeping with the mastery goal, it is critical to allow the patient, if at all possible, to choose the name of his problem. (One of my patients elected to name her complex entanglement of psychosocial stressors her “box of rocks.”) The follow-up instruction is, “Between now and the next visit, see if you can discover things that you can do, on at least some days, to make you feel more in control.” It is again desirable, unless forced, to stifle the advice-giving urge, as that can subtly undermine the patient’s sense of potential mastery.
During the next visit, the patient may report that he has discovered something that works and that he feels 25% in control of his life and the problem controls only 75%. After praising the patient for this accomplishment, your follow-up question might be, “Do you think it would be possible to try that technique even more often in the future?” or “Do you think, now that you have done such a good job of finding one thing that works, that you might think of another?” Patients may want to keep a formal written diary of their progress. By using this approach, the patient may eventually be able to report that his sense of control has risen to more than 50%. Usually that will produce a substantially improved quality of life.34,35
A further implication of the meaning model is the value of explicit “story work” with patients with chronic complex problems. Since most people attach meaning to important events in their lives by implicitly or explicitly constructing narratives about the events, story work allows the physician and patient to work together with the issue of meaning. Steps in working with the patient include:
Ask for your patient’s story about the illness or problem, including its root causes, its present effect on his life, and how he imagines the rest of life unfolding with the problem. This may be done with take-home written assignments between visits to conserve office time.
Analyze the meaning of the story with the patient, looking especially at the negatives that seem to produce unnecessary suffering, such as guilt over the presumed cause of the illness or lack of improvement in the future.
Encourage the patient to begin imaginatively constructing alternative elements to the story, especially happier endings. These are much more powerful when generated by the patient and less effective when the physician suggests them.
Work with the patient to identify changes in day-to-day activities that might reflect and further cement the reality of the revised story in the patient’s life.
A similar approach is typically employed for chronic pain patients, but it is valuable to make the story work explicit.36 At the start of treatment, these patients typically see only 2 possible stories: a future in which their pain magically disappears completely or one of unremitting suffering. It may take much supportive work to facilitate the construction of an alternative story in which the patient can imagine the possibility of feeling less pain and achieving important life goals despite never being completely cured. Often the patients have been so overwhelmed by the pain experience, and the search for a cure, they have simply lost sight of those other goals; the main element of story work is getting them back in touch with those goals. Once the patient has successfully constructed that alternative story, other elements of management (long-term medication use, physical therapy, vocational rehabilitation, depression management, and so forth) often quickly fall into place.
From a research standpoint, many of the proposed interventions are unfortunately sloppy. One would wish to validate an empirical hypothesis such as the meaning model by distinguishing whether a better health outcome is caused by an alteration in meaning or by some other factor. The chronic pain patient who constructs a better story of ongoing life with pain is likely to be much more compliant with other therapeutic recommendations. Did the improved outcome result from the change in meaning, from greater adherence to effective modalities such as physical therapy, or both? Although progress on the research front is expected, it may well come more slowly than we would like.
But this detriment for research purposes is a plus for therapeutics. Clinicians often do not care whether a patient is feeling better because of the pill prescribed, the care and compassion shown, greater adherence, or all of them combined, as long as improvement occurs and the patient is more satisfied. The ease with which elements of the placebo response become comingled with other therapeutic endeavors makes it a more useful tool.
Because of the therapeutic usefulness of the placebo response and the way that it is intertwined with other elements of everyday practice it would seem that family practice investigators would be especially skilled at addressing the research difficulties. This review should encourage more attention to the placebo response and promote a higher priority to its research among primary care investigators and multidisciplinary research teams.
This article was partially based on some preliminary work supported by the Fetzer Institute, Kalamazoo, Michigan. The author is grateful for the suggestions of several anonymous peer reviewers for The Journal of Family Practice.
- Berg AO.
Placebos: a brief review for family physicians. J Fam Pract 1977;5:97–100.
- Freeman TB,
Leaverton PE, et al. Use of placebo surgery in controlled trials of a cellular-based therapy for Parkinson’s disease. N Engl J Med 1999;341:988–92.
- Macklin R.
The ethical problems with sham surgery in clinical research. N Engl J Med 1999;341:992–96.
- Hróbjartsson A.
Placebo intervention: what is its clinical effect? Presented at the conference, Complementary and alternative medicine: the scientific and pluralistic challenge, Boston, Mass, December 10, 1999.
- Gøtzsche PC.
Is there logic in the placebo? Lancet 1994;344:925–26.
- Kirsch I.
Unsuccessful redefinitions of the term ‘placebo’. Am Psychol 1986;41:844–45.
- Berg AO.
The placebo effect reconsidered. J Fam Pract 1983;17:647–50.
- Tudor Hart J.
Caring effects. Lancet 1996;347:1606–08.
- Kienle GS,
Placebo effect and placebo concept: a critical methodological and conceptual analysis of reports on the magnitude of the placebo effect. Altern Ther 1996;2:39–54.
- Hahn RA.
The nocebo phenomenon: scope and foundations. In: Harrington A, ed. The placebo effect: an interdisciplinary exploration. Cambridge, Mass: Harvard University Press; 1997;56:–76.
- Thomas KB.
General practice consultations: is there any point in being positive? BMJ 1987;294:1200–02.
- Beecher HK.
The powerful placebo. JAMA 1955;159:1602–06.
- Moerman DE.
General medical effectiveness and human biology: placebo effects in the treatment of ulcer disease. Med Anthropol Quart 1983;14:3–16.
- Roberts AH,
The power of nonspecific effects in healing: implications for psychosocial and biological treatments. Clin Psychol Rev 1993;13:375–91.
- Kleijnen J,
de Craen AJM,
Placebo effect in double-blind clinical trials: a review of interactions with medications. Lancet 1994;344:1347–49.
- Brody H.
Placebos and the philosophy of medicine: clinical, conceptual and ethical issues. Chicago, Ill: University of Chicago Press; 1980.
- Brody H.
Stories of sickness. New Haven, Conn: Yale University Press; 1987.
- Greenhalgh T, Hurwitz B, eds.Narrative based medicine: dialogue and discourse in clinical practice. London, England: BMJ Books; 1999.
- Staats P,
Suggestion/placebo effects on pain: negative as well as positive. J Pain Symptom Manage 1998;15:235–43.
- Searight HR,
Remembering and interpreting informed consent: a qualitative study of drug trial participants. J Am Board Fam Pract 1996;9:14–22.
- Shapiro AK,
The powerful placebo: from ancient priest to modern physician. Baltimore, Md: Johns Hopkins University Press; 1997.
- White L, Tursky B, Schwartz GE, eds.Placebo: theory, research, and mechanisms. New York, NY: Guilford Press; 1985.
- Fisher S,
The curse of the placebo: fanciful pursuit of a pure biological therapy. In: Fisher S, Greenberg RP, eds. From placebo to panacea: putting psychiatric drugs to the test. New York, NY: Wiley; 1997;3–56.
- Harrington A, ed.The placebo effect: an interdisciplinary exploration. Cambridge, Mass: Harvard University Press; 1997.
- Lyerly SB,
Drugs and placebos: the effects of instruction upon performance and mood under amphetamine sulphate and chloral hydrate. J Abnorm Soc Psychol 1964;68:321–27.
- Castes M,
Classic conditioning and placebo effects in the bronchodilator response of asthmatic children. Abstract presented at Research Perspectives in Psychoneuroimmunology VIII, Bristol, United Kingdom, April 1-4, 1998.
- Amanzio M,
Neuropharmacological dissection of placebo analgesia: expectation-activated opioid systems versus conditioning-activated specific subsystems. J Neurosci 1999;19:484–94.
- Leopold N,
Sustained partnership in primary care. J Fam Pract 1996;42:129–37.
- Egbert LD,
Reduction of post-operative pain by encouragement and instruction of patients. N Engl J Med 1964;270:825–27.
- Kaplan SH,
Greenfield S, et al. Patient and visit characteristics related to physicians’ participatory decision-making style. Med Care 1995;33:1176–87.
- Kaplan SH,
Gandek B, et al. Characteristics of physicians with participatory decision-making styles. Ann Intern Med 1996;124:497–504.
- Malterud K.
Key questions: a strategy for modifying clinical communication. Transforming tacit skills into a clinical method. Scand J Prim Health Care 1994;12:121–27.
- White M,
Narrative means to therapeutic ends. New York, NY: Norton; 1990.
- Greenfield S,
Expanding patient involvement in care: effects on patient outcomes. Ann Intern Med 1985;102:520–28.
- Kaplan SH,
Assessing the effects of physician-patient interactions on the outcomes of chronic disease. Med Care 1989;27:S110–27.
- Charon R,
Literature and ethical medicine: five cases from common practice. J Med Philos 1996;21:243–65.
The Journal of Family Practice ©2000 Quadrant HealthCom Inc.