Orne, M. T. Hypnosis in the treatment of smoking. In J. Steinfeld, W. Griffiths, K. P. Ball, & R. M. Taylor (Eds.), Smoking and health. II. Health consequences, education, cessation activities and social action: Proceedings 3rd World Conference on Smoking and Health. (DHEW Publication No. NIH 77-1413). Washington, D.C.: U. S. Department of Health, Education and Welfare, 1977. Pp.489-507.
Evidence concerning the effectiveness of hypnosis as an adjunct to control smoking is reviewed. To provide perspective, the nature of hypnosis is discussed, emphasizing that it is not particularly powerful as a means of controlling behavior but is uniquely effective as a means of altering subjective experience. The methodological issues which must be considered in evaluating smoking cessation outcome studies are briefly outlined, and it is concluded that conservative studies report data only about patients who stop smoking entirely. While such findings are relatively reliable, they do have a tendency to understate therapeutic results since reports of individuals who claim to have greatly reduced their consumption are excluded. Similarly, conservative criteria of success consider all patients not available for followup as failures. Finally, in order to compare results it is crucial to know the nature of the population from which patients are derived and precisely how they were selected for participation in research studies. Reported results vary from clinical reports where neither case selection nor followup are specified, but well over 90% success rates are claimed, to the more typical result of 25% to 30% abstinent after one year. No truly satisfactory studies including adequate control data are available. Therefore, an effort is made to compare the overall results of studies using hypnosis with those involving other treatment procedures. In particular, results obtained with either placebo, recognized as such, or lobeline provide an estimate of the therapeutic response that can be anticipated solely from the placebo effect.
It is concluded that there are at least two major components involved in the hypnotherapeutic treatment of smoking: (a) a specific effect of hypnotic suggestion leading to an immediate non-traumatic cessation -- with a high rate of recidivism after a single session, and (b) non-specific effects which involve the mystique of hypnosis but do not require the patient to be responsive to hypnosis. These effects, best conceptualized as a placebo response, can nonetheless be remarkably effective in bringing about long-term changes in smoking behavior. Ultimately the most effective therapeutic use of hypnosis will depend not only on a better appreciation of what hypnosis can and cannot do but also on the understanding of the myriad of factors -- demographic, interpersonal, intrapersonal, and situational -- which of necessity affect the outcome of all efforts to modify smoking behavior.
ON THE NATURE OF HYPNOSIS
The use of hypnosis as a means to help patients stop smoking invariably arouses considerable interest among the public and considerable controversy within the professions. Thus, questions about the success of the method, the circumstances under which it should or should not be utilized, the extent to which deep hypnosis is required in order for therapeutic suggestions to be
490 ORNE
effective, and the possibility of serious substitute symptom formation are all issues which have received attention in the literature. It is not possible, however, to clarify the problems involved in the use of hypnosis without considering the nature of the hypnotic phenomenon itself. No attempt will be made here to present various points of view; rather, this discussion will be limited to statements which have considerable empirical basis and are generally agreed upon by serious investigators.
Thanks to the portrayal of hypnosis in fiction and on the stage, most individuals conceive of the phenomenon as an extremely powerful way of controlling behavior. However, hypnosis is never induced in a vacuum. I illustrate this fact when I lecture to students by asking one to take off his left shoe, another to exchange her glasses with the individual beside her, yet a third one to give me his wallet, a fourth his watch, and so on. After eliciting compliance with these inappropriate and mildly embarrassing requests, I inquire of these cooperative students whether they had been hypnotized, obviously receiving a negative answer. This allows me to point out that if I had first "hypnotized" them and then asked them to perform these identical tasks, the other members of the audience would have said, "Ah, he has them under his control!" In other words, it would have been erroneously assumed that these subjects were willing to carry out these various odd requests because they had been hypnotized when, in fact, their compliance could as readily be elicited simply by virtue of the teacher-student relationship. In the context of that relationship, students -- rightly or wrongly -- assume that their professor has some legitimate reason for his request and tend to go along with it. Since the limits of this compliance are not generally tested, the students' behavior strikes the observer as incongruous.
Similarly, hypnosis is generally induced either in the context of a doctor-patient relationship or in the context of an experimental setting. Anyone doubting the degree of control of the experimental context has but to try a simple experiment. Ask a group of casual acquaintances whether or not they will do you a favor. Their typical response will be yes. Then ask them to perform five pushups. This request will generally be met with a raised eyebrow and a question, "Why?" (phrased more or less politely, depending upon the social setting). Now take another group of casual acquaintances and again ask them to do you a favor. Once they have agreed to do so, ask them to participate in a simple experiment. Once they have agreed to that, tell them, "Do five pushups." Now the response is simply, "Where?" Note that the identical request, merely placed in the context of an experiment, leads to unquestioned compliance.
Finally, consider the doctor-patient relationship. A moment's reflection will make it clear that patients submit to an incredible variety of painful, boring, humiliating, and degrading procedures when requested to do so by their physician. One may, of course, assert that patients do this because it is necessary for their cure, but it is well to keep in mind that the patient is forced to assume that this is the case, and for the most part he has only the implicit assurance of his physician that the procedure is really needed and likely to be helpful. Physicians can, and usually do, exert a remarkable degree of control over patients' behavior.
491 HYPNOSIS
Considering the high degree of control inherent in each of the circumstances in which hypnosis is generally induced, it is hardly surprising that no unequivocal evidence is presently available which documents that hypnosis will cause individuals to perform actions that they would not do without hypnosis in an analogous context -- provided their behavior is somehow rationalized and made the responsibility of someone else (1,2).
Much of the confusion about the use of hypnosis is based upon the misconception that hypnosis will place the hypnotized individual under the control of the hypnotist. In fact, hypnosis is not a particularly effective means of inducing an individual to carry out an action that he or she is not otherwise prepared to undertake. Further, since simple compliance to requests can be elicited in a myriad of ways, if I tell a patient to raise his arm, he will do so whether he is hypnotized or not; if I tell him to close his eyes, he will do likewise. Therefore, simple compliance cannot serve to define or diagnose the presence of hypnosis.
What kind of changes, then, do index the hypnotic phenomenon? Consider another simple experiment. I ask my students to think about the hippopotamus, saying, "Perhaps you have seen a hippopotamus in a local zoo. It is a funny looking beast. Now, having thought about it, simply put it out of your mind and forget it. Try to forget the animal, the hippopotamus. Try very hard to do so. Do it now. Simply forget it." Following instructions of this kind it is the rare individual indeed who is successful in fulfilling the demand placed upon him.
In contrast, however, the deeply hypnotized individual, given precisely the suggestion outlined above, will articulate the peculiar sensation of having known something only a minute before and somehow being unable to think of it. He or she may well insist it is on the tip of the tongue. "Give me a moment. I just can't think of it right now." I am not talking here of an individual who is trying to be nice to me or who is complying with something, but rather of an individual who experiences changes in his or her memory in response to suggestions. A far easier item of the same kind would be to suggest that the right arm is getting light and floating upward. The individual who responds to such a suggestion will note -- often with some surprise --that the arm seems to be floating upward of its own accord. What indexes hypnosis is not the objective fact that the arm is moving upward, but rather that this behavior reflects an alteration in the subjective experience of the hypnotized individual. In sum, hypnosis is generally not especially effective in controlling an individual's behavior but is remarkably effective in being able to bring about alterations in hypnotized individuals' experience or memory.
One final aspect of hypnosis requires clarification. Individuals differ in their ability to respond to hypnosis. This ability is approximately normally distributed (3). It tends to be depressed in individuals with psychiatric difficulties, though there are some very disturbed people nonetheless capable of entering hypnosis and, as would be expected with a normally distributed phenomenon, there are some well adjusted individuals unable to enter hypnosis.
Only a relatively small group of individuals is totally unable to respond to
492 ORNE
hypnosis -- around 10%, depending on the criteria employed. By the same token, only a relatively small percentage of the population is extremely responsive, with most individuals falling inbetween. The 10 to 20% of normal individuals who are highly responsive constitute those individuals who would respond in the way most lay persons think of hypnosis. Only this relatively small group will manifest reasonably complete posthypnotic amnesia and respond to difficult posthypnotic suggestions.
HYPNOTHERAPY AND SMOKING CESSATION
It is hardly meaningful to conceive of hypnotherapy as a unitary treatment in the light of these background observations. Hypnosis is a phenomenon which can be employed in a variety of different ways as an adjunct to a therapeutic process. The manner in which most lay persons expect hypnosis to be used and the way it is described in some of the older literature -- still popular among many of the lay hypnotists -- requires the induction of deep hypnosis. With suitable individuals, it is possible to suggest that they will no longer wish to or be able to smoke; further, that cigarettes will henceforth taste harsh, acrid, bitter, foul, or the like, and even if the patient is tempted to smoke, the experience will be so unpleasant that he or she will certainly not want to do so again. On awakening, such a subject will express a marked disinclination to smoke, and when urged to do so will take a few puffs, rapidly discarding the cigarette and vividly describing the extremely unpleasant consequences he or she has just experienced. Both observer and patient alike will be convinced about the dramatic effectiveness of this procedure. Indeed, it is often used by stage hypnotists, presumably to show "the medical uses of hypnosis."
It should be noted that even such an authoritarian approach by no means focuses exclusively on an attempt to modify behavior directly but rather leans heavily on an attempt to appropriately modify the subject's experience and thereby bring about the desired behavior. In practice, hypnotherapeutic procedures differ in the degree to which they rely upon directly suggesting desired behavior as opposed to eliciting it indirectly by attempting to alter the patient's experience.
Another dimension which differentiates hypnotherapeutic approaches and is probably even more important distinguishes between authoritative suggestions on the one extreme and suggestions which require the patient's active cooperation for their execution. The authoritarian suggestion is usually given in a quasimagical way and can, with suitable subjects, bring about dramatic -- though at times transient -- changes.
The remarkable effectiveness of hypnosis with a suitable subject, as well as some of the limitations of direct posthypnotic suggestions to modify behavior, are illustrated in one of my early attempts to use posthypnotic suggestion to prevent a particularly difficult alcoholic patient from drinking. This patient, in intensive psychotherapy with a colleague, would begin to drink as soon as she experienced any kind of anxiety, thereby making effective psychotherapeutic treatment well-nigh impossible. The patient was known to be highly responsive
493 HYPNOSIS
to hypnosis, and I was asked whether posthypnotic suggestion might be useful to maintain sobriety long enough for treatment to take hold. Since I realized that a posthypnotic prohibition was far less likely to be effective over time than a posthypnotic suggestion to carry out a specific action, I hit upon the maneuver of suggesting in deep hypnosis that she take Antabuse daily. This drug, with which the patient was familiar, generally prevents an individual from drinking for as long as five days after its ingestion, and thus makes it less easy for a sudden impulse to drink to overcome a patient's wish to remain sober. Accordingly, I suggested that every morning she would take Antabuse and then send me a postcard saying, "I have taken my Antabuse today. I will take my Antabuse tomorrow. I feel fine" -- to which she could then add any message she wished. Writing this message was intended to reinforce the posthypnotic suggestion.
The patient, who had been deeply hypnotized, promptly began to send the postcards as suggested. Each day I received a card beginning with the fixed formula, followed by some incidental item of information. This went on for some eighty days when I finally received the most interesting postcard. It said, "I have taken my Antabuse today. I will take my Antabuse tomorrow. I feel fine. I am drunk now." The handwriting served to substantiate this assertion.
When I had the opportunity to reconstruct what had happened, it became clear that the patient had followed the suggestion compulsively. She had, however, stopped taking Antabuse some five days before she sent the last postcard. While the posthypnotic suggestion was by that time not sufficiently effective to induce her to take the medication, the remaining compulsion was sufficiently strong to induce her to send the postcard, and she described how even while intoxicated she had had this strange urge to write out a postcard and send it to me. Unfortunately, the urge to drink proved to be more powerful than the induced compulsion to take Antabuse.
On the one hand, the patient did remain sober for a time which, for this individual, was remarkably long; on the other, the effectiveness of the therapeutic aspect of the suggestion ceased unpredictably and abruptly. I feel convinced that even for those patients who are highly responsive to hypnosis, direct suggestion administered in a fashion which makes the patient a passive responder is generally inappropriate. Not only will many patients find such an authoritarian technique repugnant, but it also tends to be ineffective over a long period. When hypnosis is used in this manner, the patient need take no responsibility for a therapeutic outcome such as discontinuing smoking. On the contrary, this task is seen as the responsibility of the hypnotist and his suggestions. As a consequence, the patient tends to find the posthypnotic suggestion ego-alien, and over a period of days --occasionally weeks or even months -- he or she is generally able to overcome the effect of the suggestion and will triumphantly resume smoking. While it is possible to reinforce the posthypnotic suggestion by seeing the patient repeatedly, and in an occasional highly responsive individual to maintain the effect of the suggestion by this means, the general approach cannot be recommended. It is neither particularly effective nor, in an overall sense, helpful to the patient. Relieving the patient of his or her responsibility for smoking behav-
494 ORNE
ior will tend to facilitate inappropriate dependency, and in such a context may even provoke substitute symptoms (4).
In sharp contrast to the kind of approach discussed above, all modern techniques of medical hypnosis are employed in the context of a treatment program that incorporates a variety of procedures designed to decrease the patient's desire to smoke. Hypnosis is presented to patients as an ability which they can learn to utilize in their efforts to achieve what they would like to achieve. Thus it is emphasized that it is the patient's wish -- not the doctor's wish -- to stop smoking. Special care is taken to make certain that the patient takes the responsibility for the decision to stop smoking and to make it clear that if a patient chooses to resume smoking, it is not the physician but the patient who has failed. When hypnosis is used in such a context, it should be recognized that a great many other motivational procedures likely to affect smoking cessation will become intertwined with the hypnotherapeutic approach.
EVALUATION OF HYPNOTHERAPY: A. METHODOLOGICAL ISSUES
The evaluation of hypnotherapy in the treatment of smoking is beset with methodological problems, some of which will be considered briefly before we turn to the studies themselves.
The Criterion Measures by Which Outcomes Are Evaluated
Studies of smoking in particular tend to share problems in the criteria of success that are employed. Most typically, patients are asked how much they smoke before, at the termination of, and some time after, treatment. These reports form the basic data for almost all the research. It is not obvious, but nonetheless true, that simply requiring a patient to monitor his or her own behavior tends to exert a considerable -- and often profound -- effect on that behavior. Precisely what and how an individual is asked to report about his or her smoking behavior may also exert a differential effect on cigarette consumption (5).
Asking a patient to monitor the details of his or her own behavior may affect what he or she consumes. (This is most clearly seen when overweight patients, requested to keep track of food consumption but not to alter it, promptly begin to lose weight.) Under some circumstances the reported consumption may be modified without necessarily altering the actual consumption. Thus, when a meaningful relationship exists between a patient and a therapist seeking to help the patient decrease cigarette consumption and a report is solicited by the therapist concerning the patient's cigarette consumption, there is a tendency for the patient to under-report the amount smoked. Such a tendency, perhaps because of the patient's wish to please the therapist, makes the therapist feel good, or because the patient feels improvement is apparently expected and he or she wants to look good, or because of a fear of disappointing the therapist, or any combination of these and related motives will serve to distort a patient's report. As a consequence, a patient's unverified report of a decrease in cigarette consumption is notoriously unreliable. In contrast, however, when patients report that they no longer smoke and have not had any cigarettes for the past year, the report is
495 HYPNOSIS
far more likely to be reliable since it would not merely be bending the truth but a knowing, overt, barefaced lie to provide such a report while still smoking (6).
Any investigator confronting the task of assessing the outcome of smoking interventions must be aware of the dilemma of reporting the data. If he bases his results on reported decrements in cigarette consumption, it is very likely that the findings inadvertently seriously over-inflate the strength of the results. Equally troublesome is the fact that this same procedure applied to two different treatments will not necessarily lead to a constant degree of error. For example, in comparing some psychotherapeutic interventions (involving considerable doctor-patient contact) with a drug treatment (minimal doctor-patient contact), there is likely to be a far greater reported decrement in smoking -- given a strong doctor-patient relationship -- as opposed to a possibly much smaller difference in cigarette consumption. For reasons such as these, conservative studies report data only about those patients who stop smoking entirely. Reported in this way, results are considerably more reliable, but are likely to understate the findings since investigators are forced to ignore the reports of individuals who claim to have significantly reduced their cigarette consumption, some of which are undoubtedly true.
Problems of Attrition
A similar problem exists in how investigators choose to report findings when a significant percentage of patients is lost to followup. It becomes particularly important to know how and why only some patients are included in the followup. If, for example, a short questionnaire is sent out a year after treatment is terminated and half the patients respond, of which half indicate that they are still abstinent, how should the results be reported? Clearly, it would not seem appropriate to assume that those patients who responded are representative of the total sample. Therefore, one could not justify a statement that 50% of the treated patients continued to be abstinent after one year since it would be likely that a far higher percentage of those patients not responding to the questionnaire had resumed smoking than would be the case from among those patients who did respond. The most conservative manner of reporting would be to say 25% of the patients treated are not smoking after one year, considering all patients lost to followup as failures -- a procedure generally adopted in reporting surgical results. Other problems are how patients are solicited and selected for the study and how they are assigned to the various groups in the research. The question of appropriate controls is also very often a difficult one; this involves the matter of placebo effects. This topic will be treated in a special section later in this paper.
In the studies reported below, the details of methodological difficulties will not be emphasized except when it is essential that they be mentioned to clarify specific findings. In general, however, reports of treatment with hypnotherapy are vague in specifying patient selection and often they do not compare their results with those of control groups. All too often the precise manner in which outcome measures are employed and results calculated vary greatly, and followup
496 ORNE
procedures are not clearly specified. The reported results vary greatly. It should be clear, however, that most of the differences in reported results could as easily be accounted for by differences in subject selection, followup procedures, and criterion measures as by differences in treatment approach. Extreme caution seems appropriate in evaluating this literature.
EVALUATION OF HYPNOTHERAPY: B. STUDIES
Graff, Hammett, Bash, Fackler, Yanovski, and Goldman (7) compared the effects of four treatments: hypnosis, group therapy, prescription of lobeline, and prescription of chlordiazepoxide. In this study, 111 potential subjects were recruited by newspaper and radio advertising. Of these, however, 74 dropped out after the initial session when told about a $25 fee. These 74 were considered controls. The remaining 37 patients were divided among four groups; however, only 24 patients completed the ten weeks of treatment. The hypnosis and group sessions were one hour each while the chemotherapy sessions were less than ten minutes each. At the end of the ten-week period the success rates were: hypnosis 100%, group therapy 55%, chlordiazepoxide 33%, and lobeline 2.9%. At a three-month followup the number of patients still totally abstinent were hypnosis 88%, group therapy 44%, chlordiazepoxide 22%, and lobeline 0%. It should be kept in mind, of course, that the groups themselves were extremely small. Furthermore, the use of dropouts as controls cannot be justified. Though hypnosis appears to be the most effective treatment, the authors point out that this group was treated preferentially, receiving a great deal more individualized attention. Furthermore, the amount of selection bias which resulted in only 24 patients (of 111 individuals who volunteered for the study) completing treatment makes any conclusions based on this work extremely tenuous.
Edwards (8) compared 40 male subjects, of whom half were treated with hypnotherapy and the other half with lobeline. No subjects were lost in this study and all individuals who originally went to the anti-smoking clinic were accepted. In that study treatment consisted of an initial 30-minute session followed by three followup sessions of 15 minutes each. It is reported that both the hypnotherapy and lobeline control groups decreased the amount of smoking, but neither group did so beyond the results that might be expected from placebo.
Of the results of studies using hypnosis without a comparison group, those reported by von Dedenroth (9, 10) are most dramatic. In his earlier paper he reported on 50 cases, of which 48 had stopped smoking. In 1968 he reported on an additional 1000 cases that were treated after 1962, of which 940 are reported to have stopped smoking. While these numbers are extremely dramatic, he unfortunately fails to specify the nature of his followup, and it is difficult to know precisely what kind of data would allow him to assert with certainty that of his patients, "none have relapsed nor have any of those in later studies who stopped smoking resumed the practice." Though the numbers seem too good to be true, the approach itself is interesting and involves using hypnosis in the context of a variety of behaviorally oriented therapeutic maneuvers designed to break the association between smoking and some of its behavioral antecedents. The treat-
497 HYPNOSIS
ment procedure involves four sessions. In the first session patients are asked about their favorite brands of cigarettes, and are then told that they must switch to another brand and never smoke their favorite brand again. During a later session, patients are instructed not to smoke during meals or for an hour after meals. In a variety of ways, then, efforts are made to alter the contingencies which had in the past helped maintain smoking behavior. Hypnosis is used as a means of administering suggestions designed to reinforce patients' attitudes, and self-hypnosis is also taught in this context. Not only are some of the therapeutic maneuvers interesting and ingenious, but therapists themselves convey an almost missionary zeal in their efforts to treat tobacco mania. The importance of such variables, though difficult to evaluate, is likely to be considerable.
A somewhat different four-session approach is utilized in a study reported by Hall and Crasilneck (12). For this study 75 consecutive adults seeking to stop smoking were selected. However, a small number (not specified) of individuals who were totally unresponsive to hypnosis were excluded from the study. One year after treatment 64 of these patients reported not smoking.
Kline (13) reports the use of hypnotherapy in a very novel fashion. He selected individuals who had previously been unsuccessful in giving up smoking despite professional help, in many instances including hypnotherapy. Subjects were first evaluated clinically, and if they seemed somewhat responsive to hypnosis were included. Subjects were then assigned to hypnotherapy groups limited to 10 individuals who were treated in a group session lasting for 12 hours. Hypnosis was a part of this session and again a number of approaches were employed. All patients were required to abstain from smoking for the 24 hours preceding their session. Since it was felt crucial for individuals to learn to deal with cigarettes in their environment, cigarettes were freely present during the entire session and patients not only could touch and see cigarettes but individuals were intentionally smoking in the presence of the patients at various times. The experience was designed to be intensive and to become highly meaningful to each of the participants. Group discussions were interspersed with hypnotheraputic parts of the session. Six groups of 10 patients each were studied. After one year 100% of the participants were reported to have responded to followup. Of these, 88% had quit smoking.
Almost equally dramatic were the results reported by Nuland and Field (14) who compared two types of treatment, one fairly standard procedure in an earlier study with a limited and fixed numberof sessions, and the other a highly individualized procedure which leaned heavily on a variety of therapeutic approaches, including feedback under hypnosis of the patients' own reasons for quitting, continued telephone contact with the patients, meditative techniques during hypnosis, and as many sessions as seemed necessary. Whereas the results of the standard procedures on followup had been approximately 25% still not smoking the individualized technique resulted in a significantly greater number of cures. In the second study involving 84 subjects, 90% responded to followup and of these 60% had stopped smoking.
Another well controlled study by Orr (15) involved as many sessions as required. The subject population was 198 patients between the ages of 15 and 60
498 ORNE
who were connected with the Atomic Energy Research Establishment at Harwell, England. Those individuals who did not attend more than two sessions were considered dropouts, and the final study group consisted of 258 subjects. Treatment involved 15-minute sessions but the patients were able to come as often as they wished; the average number of visits was six. Suggestions were given that the patient would be free of the desire to smoke. On a six-month followup inquiry, 96% of the patients responded and of those responding 50% had stopped smoking entirely. Though these results are quite promising, if one considered the total number of subjects who originally volunteered, only about 30% of those who applied stopped smoking.
Though not supported by quantitative data, Hartland (16), Kroger and Libott (17), Stein (18) and others reviewed by Johnston and Donoghue (19) apply somewhat varying approaches in the hypnotherapeutic treatment of smoking. It should again be emphasized, however, that while these various approaches all utilize hypnosis or self-hypnosis as a means of reinforcing the patient's desire to stop smoking and employ suggestions of relaxation to help him or her get over the difficult periods of tension simetimes seen during early periods of abstinence, these procedures invariably involve a great many other aspects.
Most of the techniques discussed by Dr. Rose earlier in this session in the context of physician intervention and patient health counseling are utilized singly or jointly by one or the other clinician's approach to hypnotherapy. Similarly, the various behavior therapeutic maneuvers outlined by Dr. Lichtenstein are utilized more or less systematically in the context of what different authors describe as the hypnotherapeutic technique. Even the use of pharmacologic adjuncts as discussed by Dr. Jarvik has been subsumed by some individuals in the hypnotherapeutic management of smoking e.g., Kroger and Libott (17).
To clarify the role which hypnosis plays in a given program of smoking cessation, it would be essential to compare a particular treatment program including hypnosis with the identical treatment program without hypnosis. Unfortunately, no such studies are available. Perhaps the one which best approximates this approach is by Francisco (20). In an unpublished thesis on the modification of smoking behavior, he compares three approaches: group hypnosis, group discussion, and group relaxation with each other and a no-treatment control group. Twelve matched subjects are assigned to each of the four groups and abstinence is used as a criterion measure. Followup is carried out at one week, two months, and six months by telephone. The treatments were deliberately chosen to contain overlapping components. In each instance it was stressed that the technique was intended as an aid to help the patient quit smoking but was not a cure. In all instances attempts were made to help individuals identify their difficulties in stopping. Some individualized attention was given to each patient.
The relaxation treatment largely subsumed most of what was included in the group discussion, specifically adding the relaxation component in a way that is sometimes used as a preparation for hypnosis. The hypnosis treatment subsumed all aspects of the relaxation procedure but further added a formal hypnotic induction procedure.
499 HYPNOSIS
In general the results showed that there were significant differences in outcome between treatment and no-treatment groups, both initially and at two months followup, but these differences declined to an insignificant level after six months. There were no significant differences between treatment methods, though the discussion and hypnosis groups did slightly better than the relaxation group.
While this study is interesting and seeks to avoid some of the methodological difficulties, it may well be even the transient improvement in the treatment groups could be attributed to disappointment of the subject in the no-treatment control group -- although limiting the study to confirmed smokers and using the criterion of total abstinence may have helped to minimize such an effect. On the other hand, there were only 12 subjects in each group. From the point of view of evaluating the effect of hypnosis, one needs to keep in mind that only a relatively small percentage of these individuals would have been deeply hypnotized. While a comparison of unselected subjects treated by hypnosis allows inferences to be drawn about the effect of hypnotic induction on unselected patients, it cannot tell us how effective hypnosis might be as an adjunct to treatment with those patients who are particularly responsive to hypnosis.
Francisco's study helps point up the tendency of most studies to define hypnosis by what is done to the patient despite the fact that in response to the identical procedure some patients will be hypnotized while others will not. If one wishes to answer the question as to how effective hypnosis is for those who have the ability to respond, an entirely different design is needed. The comparison is between patients highly responsive to hypnosis and patients unable to respond to hypnosis, both of whom have been exposed to a hypnotic induction procedure. Since the patients who cannot respond to hypnosis would not be expected to obtain whatever benefit is associated with actually being hypnotized they serve as an appropriate control group for the highly hypnotizable individuals. If differences are found in such a design they might, of course, be fully explained by the personality trait of being hypnotizable as such rather than to the effect of being hypnotized. For this reason one would also wish to make the same comparison with groups of patients with known hypnotizability who are not treated by hypnosis. No studies based upon such a model are presently available. However, some hints can be gathered by examining clinical data of patients treated with hypnosis where some independent assessments of hypnotizability are available, as, for example, in Spiegel's work to be discussed later.
Specific and Non-Specific Components of Hypnotic Intervention
As one strives to evaluate the results of hypnotherapy, especially if one hopes to contrast it with alternative procedures, the task becomes extremely difficult. The important differences are likely to be in some of the overtones of treatment. For example, physician counseling leans heavily on the mystique of medicine and the doctor-patient relationship. Further, some of the most dramatic effects are reported when counseling is undertaken with patients immediately after a myocardial infarction. Certainly such an event can provide a unique motivation riot available under usual circumstances. (This does not, of course, explain
500 ORNE
the remarkably high incidence of continued abstinence long after the initial crisis has passed, nor the difference in results obtained by particular practitioners.) Again, behavior therapists distinguish themselves not merely in their approach but even more in the systematic manner in which the techniques are applied and the results quantified. However, beyond this common thread, there is little that unifies the various different behavioral approaches. Similarly, hypnotherapy does have some attributes which partly distinguish it from other procedures. There is an emphasis on the use of posthypnotic suggestion and systematic self-suggestion as a means of increasing self-control. Furthermore, while modern hypnotherapists characteristically emphasize the need for patients to assume responsibility for abstinence, they nonetheless tend also to utilize rapid development of a meaningful interpersonal relationship that is facilitated through hypnosis. Accordingly, some clinicians go to great effort to convey their concern and interest in the patient's struggle to abstain from smoking, and require their patients to provide feedback about the outcome of treatment. This is often effectively managed by telephone contact as well as by appropriately spaced visits.
It is fair to say, however, that the differences between the various therapeutic approaches tend to be far smaller than is generally recognized. Certainly the differences between some behavior therapeutic approaches, some physician counseling efforts, and some forms of hypnotherapeutic intervention are more in the manner in which the practitioner describes these activities to himself or herself than in the manner in which he or she interacts with patients. The simple fact is that the statement that smoking is being treated by hypnotherapy tells next to nothing about the details of the particular approach which is being employed.
In evaluating the outcome of uncontrolled studies reporting different cessation rates, it is appropriate to ask about the spontaneous rate of cessation. According to Horn (21), there is a 4 to 5% spontaneous annual rate of stopping smoking. However, such an overall spontaneous rate is likely to be considerably less than that seen among individuals who are eager enough to seek and accept help in their efforts to discontinue smoking. Differential motivation leading to selection bias, and a host of other factors generally subsumed under the concept of placebo response, should yield a significantly greater cessation rate among individuals seeking any form of therapy.
Traditionally, the placebo response has been conceptualized as the effect of pill-taking, and has been measured by evaluating the therapeutic consequences of a pharmaceutically inert pill. One study of the placebo response, that done by Schwartz and Dubitzky (22, 23), investigated the effect of an inert pill which is offered to patients as a drug helpful in reducing the urge to smoke. In a complex study, over 8000 male members of a prepaid health plan were sent a questionnaire which inquired about their smoking behavior. Subjects who indicated some desire to stop smoking and reported smoking more than ten cigarettes a day were solicited for participation in this study. Thirty-six males each were assigned to seven different treatment groups and two control groups. The treatment group of interest here was given placebo and a prescription which allowed them to renew their "medication" at two-week intervals. All subjects were treated for
501 HYPNOSIS
eight weeks. Two essentially untreated comparison groups were told, in one case before and in the other case after the completion of intake procedures, that the program was filled and they could not be included.
Of the placebo group, 28% are reported successful at the end of treatment as opposed to 11.1 % for each of the control groups. One year later, 2.5% of the placebo group were still abstinent, but, interestingly, the number of individuals who stopped smoking had risen in the first control group to 17% and in the second to 19% (possibly by obtaining some other unspecified kind of help). The criterion of success in this study was that patients reported reduction of cigarette consumption from 85 to 100%. While placebo in combination with counseling was reported more successful, Schwartz and Dubitzky comment on the remarkable success rate with placebo alone, which they feel deserves further consideration in view of its simplicity and low cost.
In contrast to this apparently dramatic effect of placebo is a finding from a study on lobeline. The British Tuberculosis Association (24) reported on the comparison of the effects of lobeline and placebo in 101 subjects. Each subject received either lobeline or placebo and was instructed to take four tablets daily and to keep track of the number of pills ingested and cigarettes consumed weekly for two six-week periods in a double blind study. Only 12% of the lobeline group had stopped smoking at the end of treatment. In a one-year followup, 7% of the lobeline group and 11% of the placebo group were still abstinent. The results reported in this study are in line with the usual observation that lobeline has little if any effectiveness in the treatment of smoking. From the point of view of this paper, these "drug effects" may be the best estimate of a placebo response in another population as it would hardly be realistic to expect a greater response to sugar pills than to lobeline.
A comparison of these studies is instructive in shedding light on some of the possible causes for apparently widely differing results with the same treatment -- placebo in this case. Perhaps the most important difference is likely to be the manner of subject selection. Thus in the Schwartz and Dubitzky study (22) individuals were selected for participation on the basis of their questionnaire responses using the criterion that they indicated a wish to stop smoking and that they smoked at least 10 cigarettes a day. Such a criterion inevitably includes individuals whose cigarette comsumption is relatively limited and may also include many who have never made any concerted effort to stop smoking, who might well be able to stop smoking with little difficulty if they decided to try.
The problem of evaluating these results is analogous to that observed in efforts to evaluate the treatment of phobias. Typically, many of the early behavior therapy studies were carried out with normal college students whose phobias were detected by means of a fear questionnaire. Consequently extremely high rates of "cure" were reported, whereas the results in treating phobic patients presenting themselves to an outpatient clinic with identical techniques yielded far less promising results.
It seems reasonable to expect that a population of individuals who had sought to stop smoking for a long time and had failed to obtain help in the past is considerably less likely to be placebo responders than an unselected population
502 ORNE
of individuals who wish to stop smoking but have never actually tried to do so. The high base rate of spontaneous smoking cessation among the untreated control subjects after one year in the Schwartz and Dubitzky study is significantly above the spontaneous cessation rate in a normal population, strongly suggesting unusual selection bias in this sample.
A second difference relates to the criterion of improvement. Schwartz and Dubitzky arbitrarily chose a reported decrease of smoking from 85 to 100% as indicating a successful outcome. The British Tuberculosis Assoication study on the other hand used abstinence as their criterion. It seems unfortunate that all investigators do not at least report the incidence of total abstinence, as well as any arbitrary criterion they might then choose, to permit comparison of studies (especially since the amount of error in a reported decrease in smoking behavior may interact differentially with treatment procedures).
It should be emphasized that while the concept of placebo derives from psychopharmacology, it has long since been appropriately generalized to any treatment procedure. One may appropriately speak of the placebo effect of a physical examination, or a diet regimen, or of any other real or imaginary treatment. The more dramatic the treatment procedure (e.g., various forms of physical therapy, manipulation, acupuncture, biofeedback, and so on), the more important it is to have appropriate controls for placebo effects. Similarly, psychotherapeutic treatment has placebo components which are independent of the specific help that may derive from psychotherapy (for a review see Shapiro (25)). Similarly, work in our laboratory has shown that hypnosis too may produce significant placebo effects that are independent and uncorrelated with the specific effects of hypnotic suggestions (26,27,28).
It has long been recognized that even individuals who are essentially unhypnotizable may derive considerable therapeutic benefit from appropriately phrased suggestions. (29). Further, if one selects individuals who are in fact unresponsive to hypnotic suggestion and persuades them that hypnosis may nonetheless be helpful in raising their pain threshold, one is able to document a highly significant effect of hypnotic suggestion on this threshold even though these individuals are not otherwise responsive to hypnosis. This placebo effect of hypnosis correlates highly with the placebo effect following the ingestion of a pill, thought by the subject to be an analgesic, tested at another time in the same population. On the other hand, highly hypnotizable individuals show an even greater increase in pain threshold from hypnotic suggestions, and the response in this group is totally unrelated to their placebo response. (Incidentally, the magnitude of the placebo response itself is uncorrelated with hynotizability.)
In assessing the effect of hypnotic treatment on smoking, it seems desirable to distinguish between the specific effects of hypnotic suggestions as opposed to non-specific effects of the hypnotic procedure which can best be conceived of as a placebo response. Judging from past research, the non-specific placebo components of hypnosis are likely to be independent of the effects due to specific suggestions seen in hypnotizable individuals --these two independent effects tend to be additive.
503 HYPNOSIS
A Single-Session Treatment Program
From the point of view of cost-effectiveness as well as to help clarify some of the important theoretical issues raised above, the work of Spiegel, who has developed a single-session treatment of smoking, is of major significance. His procedure will not be discussed in detail here. However, some unique aspects are particularly relevant to our discussion. He has clearly outlined a specific procedure and has systematically applied it to a large number of private patients seeking help in their efforts to give up smoking. The population is unusual in that a very high percentage of these individuals have tried in a variety of ways to give up smoking without success and view hypnosis as the last resort. Whereas other clinicians go to great lengths to individualize their procedures, Spiegel has developed a technique that is almost completely standardized. The session begins with an assessment of hypnotic responsivity using a clearly specified procedure (30) followed by a carefully developed combination of advice, exhortation, instruction, and suggestions on the management of the impulse to smoke. Of particular importance in clarifying the significance of hypnosis in the context of such treatment is the fact that, despite the assessment of hypnotizability, Spiegel treats patients in exactly the same fashion regardless of their response to hypnosis. Patients are instructed to return a postcard two weeks after the session indicating how they are doing, and a written followup is carried out after one year.
From a therapeutic point of view, Spiegel has felt that a single session represents the maximally efficient treatment approach to smoking. There is no possibility for the patient to shift responsibility to the therapist since it is made clear that the decision to stop smoking is the patient's, and the success or failure of the patient's efforts will have no effect on how the therapist feels but will certainly have much effect on how the patient feels. Since the overall results of a single session involve a 25% rate of discontinuing smoking, the results are sufficiently encouraging to suggest that the procedure be considered for any mass treatment effort. Certainly a 25% cessation rate is respectable indeed -- as good as or better than the overwhelming majority of ongoing cessation efforts whose initial results are often better but are rarely maintained.
Hypnotizability is assessed by a brief but standard technique with all patients, and regardless of their response, they are subsequently given the identical therapeutic suggestions. It therefore becomes possible to examine the differential effect of hypnotic suggestions and the hypnotic context on the hypnotizable individual as opposed to the effect of the same hypnotic context on the unhypnotizable individual. The individual who has little or no response to hypnosis may nonetheless respond to the suggestive or placebo components of the procedure, whereas the hypnotizable individual will utilize his or her hypnotic capacity instead, or as well.
In many ways, the most interesting observation for an understanding of the hypnotherapeutic treatment of smoking is the difference between long-term and short-term responses among patients who differ in hypnotizability. Spiegel finds that patients who are able to respond particularly well to hypnotic suggestions tend to stop smoking immediately and with little difficulty. On the other hand,
504 ORNE
there is a relatively high incidence of individuals who begin again to smoke among these high responders. In contrast, a smaller percentage of patients in the unhypnotizable group stop smoking following treatment; however, most of these individuals continue to be abstinent. Spiegel has suggested that highly hypnotizable patients do indeed respond to hypnotic suggestion and as a consequence stop smoking after hypnotic treatment. Unfortunately, these individuals may also be responsive to other hypnotic-type suggestions occurring spontaneously in their environment which may prompt them to resume smoking, and therefore, over a period of time, the initial large differential response of this group all but disappears. The unhypnotizable individuals, on the other hand, may find some novel aspect of the intellectual appeal meaningful, and they may respond to what we have conceptualized as the placebo component of hypnosis. Those unhypnotizable individuals who do stop smoking in response to hypnotic treatment tend to continue to remain abstinent. The mechanism responsible for such an individual to give up smoking is likely to be quite different from that involved in the highly hypnotizable patient.
Elsewhere, Spiegel (31) has emphasized the importance of individual differences in hypnotizability for finding the therapeutic strategy most appropriate for a given patient. It seems likely that highly hypnotizable individuals who stop smoking may require additional support and contact to help reinforce their responses to the therapeutic suggestion to stop smoking and successfully resist the environmental pressures which urge them to resume the habit. The present work, which has purposely avoided making distinctions among patients, may now make it possible to develop hard criteria by which to select those individuals most likely to profit from additional sessions. Perhaps such sessions might well be conducted in a group setting, thus decreasing cost without necessarily losing effectiveness. Indeed, additional information included in a followup postcard might well provide an additional basis for inviting specific patients to return for followup treatment.
SUMMARY AND CONCLUSIONS
We have tried to emphasize that hypnosis should be conceived of not so much as a potent means of controlling behavior but rather as a uniquely effective method to help individuals alter their subjective experience. In a true sense, hypnosis is a therapeutic technique which can assist a patient to do something he or she wishes to do but is unable to do. However, it is rarely if ever useful as a means of forcing a patient to do something he or she does not choose to do but a physician wishes the patient to do.
There are no controlled studies which allow the adequate assessment of hypnotic procedures versus alternative procedures in the treatment of smoking. On the other hand, there are a number of studies which report cessation rates following hypnotherapeutic treatment. For the most part, neither selection procedures, treatment procedures, nor followup procedures are clearly specified, and in one instance at least the reported results are so good as to strain the credulity of the reader. Despite these shortcomings, however, a review of the work by a
505 HYPNOSIS
number of clinicians clearly indicates that the hypnotherapeutic treatment of smoking must be seriously considered in the context of any large-scale cessation effort. Several individualized procedures show considerable promise and deserve careful, systematic evaluation.
One of the major difficulties confronting any systematic research in this area is the tremendous degree of overlap between behavior therapy, physician counseling, and hypnotherapy. It is all but meaningless to ask: "What is the effect of hypnosis alone?" Instead, we must address the question: "To what degree can hypnosis and the hypnotic context facilitate the effectiveness of a total treatment program?"
The work of Spiegel clearly indicates that significant cessation rates can be obtained with a single session, making this type of hypnotherapeutic intervention economically viable. Finally, an examination of data stemming from his work is in line with our own laboratory findings which suggest that there are at least two major components involved in the hypnotherapeutic treatment: (a) a specific effect of hypnotic suggestions administered to hypnotizable subjects leading to an immediate, non-traumatic cessation of smoking -- with a fairly high rate of recidivism after a single session, and (b) non-specific effects which involve the mystique of hypnosis but do not require the patient to actually be responsive to hypnotic suggestions. This effect, best conceptualized as a placebo response, can nonetheless be remarkably effective in bringing about long-term changes in smoking behavior. The smaller percentage of non-hypnotizable subjects who stop smoking tend to remain abstinent for a long period without additional treatment.
The optimal therapeutic use of hypnosis demands that the clinician recognize what hypnosis can and cannot do. Only in this fashion can he or she help patients recognize their trance capacities and enable them to utilize it to accomplish what they seek to undertake. Employed in this fashion hypnotic techniques can serve a useful purpose in the context of any cessation effort; however, any ultimate in-depth understanding of the hypnotherapeutic treatment of smoking will in the final analysis depend upon a better appreciation of the myriad of factors -- demographic, interpersonal, intrapersonal, and situational -- which inevitably affect the outcome of any and all efforts to modify smoking behavior.
ACKNOWLEDGEMENTS
The substantive research upon which this paper is based was supported in part by Grant No. MH 19156 from the National Institute of Mental Health and in part by the Institue for Experimental Psychiatry.
The author wishes to express his appreciation for the helpful comments of his colleagues, Frederick J. Evans, John F. Kihlstrom, Emily Carota Orne, William M. Waid, and Stuart K. Wilson. Special thanks is due to James Hamos for identifying and collating the relevant literature and Mae Weglarski for putting the final manuscript into format.
506 ORNE
REFERENCES
1. Orne, M.T. Hypnosis, motivation and the ecological validity of the psychological experiment. "Nebraska Symposium on Motivation" pp. 187-265, University of Nebraska Press, Lincoln, Nebraska, 1970.
2. Orne, M.T., and Evans, F. J. Social control in the psychological experiment: Antisocial behavior and hypnosis. J. Person. Soc. Psy. 1:189-200,1965.
3. Hilgard, E.R. "Hypnotic Susceptibility." Harcourt, Brace & World, New York, 1965.
4. Cohen, S.B. Hypnosis and smoking. J. Amer. Med. Assoc. 208:335-337, 1969.
5. McFall, R.M. The effects of self-monitoring on normal smoking behavior. J. Consult. Clin. Psy. 35:135-142, 1970.
6. Orne, M.T. From the subject's point of view, when is behavior private and when is it public: Problems of inference. J. Consult. Clin. Psy. 35:142-147,1970.
7. Graff, H., Hammett, V.B.O., Bash, N., Fackler, W., Yanovski, A., and Goldman, A. Results of four antismoking therapy methods. Penn. Med. J. 69:39-43, 1966.
8. Edwards, G. Hypnosis and lobeline in an anti-smoking clinic. Med. Officer 111:239-243, 1964.
9. von Dedenroth, T.E.A. The use of hypnosis with "tobaccomaniacs." Amer. J. Clin. Hyp. 4:326-331,1964.
10. von Dedenroth, T.E.A. Further help for the "tobaccomaniac." Amer. J. Clin. Hyp. 4:332-336, 1964.
11. von Dedenroth, T.E.A. The use of hypnosis in 1000 cases of "tobaccomaniacs." Amer. J. Clin. Hyp. 10:194-197, 1968.
12. Hall, J.A., and Crasilneck, H.B. Development of a hypnotic technique for treating chronic cigarette smoking. Int. J. Clin. Exp. Hyp. 18:283-289, 1970.
13. Kline, M.V. The use of extended group hypnotherapy sessions in controlling cigarette habituation. Int. J. Clin. Exp. Hyp. 18:270-282, 1970.
14. Nuland, W., and Field, P.B. Smoking and Hypnosis. A systematic clinical approach. Int. J. Clin. Exp. Hyp. 18:290-306,1970.
15. Orr, R.B. Hypnosis helps reluctant smokers. Practitioner 205:204-208, 1970.
16. Hartland, J. "Medical and Dental Hypnosis and Its Clinical Applications." Williams & Wilkins, Baltimore, 1966.
17. Kroger, W.S., and Libott, R.Y. "Thanks, Doctor, I've Stopped Smoking." Charles C. Thomas, Springfield, Illinois, 1967.
18. Stein, C. A displacement and reconditioning technique for compulsive smokers. Int. J. Clin. Exp. Hyp. 12:230-238,1964.
19. Johnston, E., and Donoghue, J.R. Hypnosis and smoking: A review of the literature. Amer. J. Clin. Hyp. 13:265-272, 1971.
20. Francisco, J. Modification of smoking behavior. A comparison of three approaches. Doctoral dissertation Wayne State University, Ann Arbor, Michigan. University Microfilms No. 73-12, 511, 1972.
21. Horn, D. Epidemiology and psychology of cigarette smoking. Chest 59:225-245, 1971.
22. Schwartz, J.L., and Dubitzky, M. The smoking control research project: Purpose, design, and initial results. Psychol. Reports 20:367-376, 1967.
23. Schwartz, J.L., and Dubitzky, M. One-year followup results of a smoking cessation program. Canadian J. Public Health 59:161-165, 1968.
24. British Tuberculosis Association. Smoking deterrent study. Brit. Med. J. 2:486-487, 1970.
25. Shapiro, A.K. Placebo effects in psychotherapy and psychoanalysis. J. Clin. Pharm. 10: 73-78, 1970.
26. Evans, F. J. The placebo response in pain reduction. "Advances in Neurology" Volume 4, "Pain," pp. 289-296 (J.J. Bonica, ed.), Raven Press, New York, 1974.
27. McGlashan, T.H., Evans, F.J., and Orne, M.T. The nature of hypnotic analgesia and placebo response to experimental pain. Psychosom. Med. 31:227-246, 1969.
507 HYPNOSIS
28. Orne, M.T. Pain suppression by hypnosis and related phenomena. "Advances in Neurology," Volume 4, "Pain," pp. 536-572 (J.J. Bonica, ed.), Raven Press, New York, 1974.
29. Orne, M.T. Hypnosis, motivation, and compliance. Amer. J. Psy. 122:721-726, 1966.
30. Spiegel, H., and Bridger, A.A. "Manual for Hypnotic Induction Profile: Eye-Roll Levitation Method." Soni Medica, New York, 1970.
31. Spiegel, H. The grade 5 syndrome: The highly hypnotizable person. Int.
J. Clin. Exp. Hyp. 22:303-319, 1974.