On placebos, placebo responses and placebo responders- A Review of psychological, psychopharmacological and psychophysiological factors, II- psychopharmacological and psychophysiological factorsDR Doongaji, VN Vahia, MPE Bharucha
Department of Psychological Medicine, K.E.M. Hospital and Seth G. S.Medical College, Bombay 400 012., India
This article is in continuation with the previous publication. The article reviews various psychopharmacological factors which determine the ultimate success of therapy, even when the drug administered does not have any specific efficacy in conditions under treatment. The importance of usually ignored variables of the psychophysiological state of an individual undergoing therapy, and its relevance to placebo genesis is discussed.
Any attempt to understand the pharmacology of placebo reveals a peculiar paradox. Classical pharmacology studies the effects of an active chemical on the functioning of a living organism. The pharmacological study of placebo usually deals with those effects of an inert substance which are independent of its chemical action. The action of placebo is not restricted to subjective symptoms like pain, dizziness, giddiness etc., but is seen also with various symptoms which have objective parameters. Cleghorn  et al studied the effects of placebo on adrenocortical secretions. They observed that injections of placebo can produce a response similar to ACTH i.e., changes in the eosinophil and lymphocyte count, sodium and potassium levels, and 17ketosteroid excretion.
Most situations involving the administration of a drug consist of the following variables. 
1. The Drug Itself
Some of the interacting variables about the drug itself need special consideration. A drug can be administered in a variety of ways to impress the consumer. The drug can be a bland white unattractive tablet, or a pretty coloured capsule. It can be given as a syrup, or as a vile liquid, or it can be administered parenterally in the form of injections which are impressive but painful . 
The potency of placebo effect is evident most dramatically by the side effects produced by its administration. These are as frequent and as unpleasant as those which can occur with pharmacologically active drugs. Placebos have been known to mimic the effects of true pharmacological substances. ,,,, Complaints like drowsiness, staggering gait, blurring of vision, urinary frequency, ringing in the ears, dermatitis medicamentosa, anaphylactic shock and a variety of other side effects are reported in the literature. Lasagna et al , showed that dose response relationships, peak effects, cumulative action, and holdover effect; can occur during the course of placebo therapy. Cases of habituation and addiction to placebo have also been reported in the literature. ,, However the lethal dose of placebo is as yet unknown.
The following side effects have been reported on administration of placebo. 
Other side effects of placebo reported in the literature include. 
It is therefore clear that the scientific evaluation of any new treatment procedure or new drug should include a study of placebo interaction. DuBois  once said `Any young neophyte can introduce a new drug. It requires a man of large experience and considerable reputation to destroy an old one'. The statement emphasises the fact that in a clinical study, before any cure is attributed to a new drug, various variables associated with the treatment procedure (e.g. the effects of hospitalisation, attention by the physician, rest, change in the diet etc.) must all be accounted for.
2. Doctor-Patient Relationship 
Patients generally look upon the treating physician as their saviour. This often leads to an unrealistic relationship between the therapist and the patient. Since the majority of patients develop a `positive transference' relationship, a proper manipulation of such a situation can be the basis for deliberate and effective use of placebo therapy.
In any therapeutic endevour the implication has always been that certain variables in the patient were responsible for any undue complication during therapy. Recent studies have shown that the psychology and the social standing of the treating physician may contribute substantially to the treatment effects.
The term `iatroplacebogenesis'  is used to describe the study of placebo effects produced by physicians. Iatroplacebogenesis can be direct or indirect.
(a) Direct iatroplacebogenesis
This is a condition produced by the direct effect of the physician's attitude towards (i) the patient, (ii) the treatment, and (iii) the results of the treatment. 
(i) Physician's attitude towards the patient
While studying the clients' reactions to initial interviews, Pollansky and Kounin  concluded that the physician's personal interest and not his competence was the main determinant of placebogenesis. The patients' liking for the doctor, the doctor's interest, empathy, sympathy, neutrality or disinterest, hostility or rejection-al. contributed to the results.
The therapist's favourable feelings to the patient were related to the therapist': expectation of improvement. The patient'; emotional attachment to the therapist also influenced the obtained improvement.  Interest in the patient also varied with the amount of interest expressed by the patient towards the therapy, and the therapist's understanding of the patient's behaviour. 
Shapiro stated that psychologists who were interested in their patients or subjects were usually more persuasive and elicited better conditioning, learning, higher intelligence scores, and better Rorschach records. 
(ii) Physician's attitude towards treatment
The Physician's interest in the treatment was found to be associated with the ready acceptance of treatment, decreased number of drop-cuts, and a more successful outcome. Various authors have concluded that the interest of the therapist was the crucial determinant in the reported results of surgery in dogs, gastric acid secretions, metabolic changes, laboratory procedures and galvanic skin responses .  It determined the success of psychotherapy, , behaviour therapy , psychochemotherapy,  placebo effects  and the successes of shamans and quacks and other indigenous therapies. 
A number of studies on psychochemotherapy, psychotherapy and hypnosis have confirmed the view that the therapist's belief, enthusiasm, commitment, interest, and expectations (either positive or negative) about the treatment were the nonspecific determinants in all therapies. 
(iii) Attitude towards results of treatment
The interest of the investigator resulted in data distortion not only caused by random observer effects, but also by intentional or unintentional non-random observer bias. 
In one experiment  on rat learning, two sets of investigators were told that their subjects were selectively bred for either high or low levels of intelligence. In actual fact the rats in both groups were genetically similar. The results showed significantly better learning from rats who were supposed to be of high intelligence (in accordance with the prior instructions given to the investigator).
It was also inferred that the more biased investigators tended to make larger errors in the direction of their preconceived notions.
(b) Indirect iatroplacebogenesis
Shapiro has stated that indirect iatroplacebogenesis is a subtle mechanism which is used extensively and which has not been considered adequately in the literature. 
Interest in the theory and method of treatment on the part of the treating physician as opposed to personal interest in the patient may also lead to placebogenesis. This happened when the patient misinterpreted the physician's interest in his treatment as interest in his person. Placebo effects were also augmented when the physician was prestigious, and dedicated to his theory,  and when the therapy was elaborate, detailed, expensive, time-consuming, fashionable, esoteric and dangerous.
Iatroplacebogenesis was reported to be particularly important in psychological therapies and situational variables where staff attitudes, type of patient populations, treatment procedures and miscellaneous factors could have influenced the results.
(3) The Patient
In conventional medicine, credit or blame (for undue effectiveness or undue ineffectiveness of therapy) is attributed to variables in the patient. Many of the possible variables in a patient which ultimately influence the drug effects have already been discussed. Some of the other variables are as follows:
(a) Personality traits and patterns
Attempts to correlate an individual's personality and placebo effect have not been successful.
Various authors have found multiple dependant variables. Placebo responders have been shown  to be compliant, religious, hypochondriac,  anxious, less educated, and freqnent users of cathartics.  Lasagna, , Beecher, ,, Frank,  Gleidman et al,  Shapiro , and others found that anxious and depresed patients were frequent placebo responders. Dependency, neuroticism, extraversion, etc., were the other variables which were considered to be important. 
Shapiro  reviewed the available literature on placebo responders and concluded that no specific traits seem to characterise placebo responders. Gordon  reinforced Shapiro's statement with his own statement that `the relationship between personality traits and placebo reaction that is found in one situation may not appear or may even be reversed in another situation.'
This was very clearly demonstrated in an important study by Knowles and Lucas. 
The authors carried out a series of experiments under `individual' and `group' conditions. All the subjects were asked to report on the `side effects' of a new 'psychotropic drug' which was in fact a small white lactose tablet. Patients assigned at random to `individual' conditions assessed the pill while seated alone in a room; while those assigned to `group' conditions sat in groups of three. All subjects were scored for neuroticism, emotional instability and extraversion on the Maudsly Personality Questionnaire. 
In two experiments carried out under `group' conditions there was a significantly positive correlation between neuroticism and the number of placebo side effects reported. When similar comparisons were made in two experiments carried out under `individual' conditions, the relationship between neuroticism and side effects was almost zero. This observation confirmed the postulate that under group conditions of drug administration, the more suggestible neurotic was likely to `catch' side effects from other people around him. 
In the same experiment, Knowles and Laucast  demonstrated a complex relationship between extraversion and frequency of side effects to placebo. In one `individual' experiment, there was a significant correlation between extraversion scores and placebo response. This did not appear in any of the two `group' experiments. In the second `individual' experiment the opposite results were found and the introverted subjects showed more placebo effects. The reason given for this was that in the first `individual' as well as `group' experiments, student nurses were used as subjects, whereas the second `individual' experiment was done with theological students. The experimenters suggested that the theological students, having little background knowledge of drugs, reported minor effects which were disregarded by the more experienced nurses. Introverts would be much more meticulous in doing so, and hence the greater placebo response in these personalities.
Thus no significant correlationship exists between personality patterns and traits and placebo effect, except that anxious people respond better to placebo.
(b) Nature of placebo stimulus
It is a common experience that various individuals respond differently to placebos given in different forms, e.g., patients from lower socioeconomic strata generally prefer injections, whereas patients from higher strata respond better to a stimulus given in the form of a psychotherapeutic sitting.  The size of the tablet, absence or presence of sugar coating, rigidity of instructions at the time of administration etc., all contribute to placebogenesis.  Such variables are poorly understood at present and in the opinion of Shapiro, these factors probably explain the discrepancies found in the results in various studies. 
(c) Age, Sex, Intelligence
Various studies which tried to correlate age, sex and intelligence with the placebo response gave contradictory results.
In an animal study, untreated rats placed in cages with amphetamine treated rats, essentially mirrored the behaviour of the treated animals.  Other investigators have found that when a therapeutic procedure was administered in a permissive atmosphere where communication between patients and the doctor was encouraged, the improvement was greater than when the same procedure was performed under nonpermissive conditions where silence was maintained.  When stimulant and depressant drugs were given in various combinations to subjects who were paired off with interpersonal contact permitted during the response period, a subject's response on psychomotor tests was dependent not so much on the drug he received as upon the drug his partner received. 
(4) Persons Administering and Other Environmental Factors
(a) Staff attitudes
The attitudes of the staff, their expectations anad their conflicts can influence placebo reactions.  The reported incidence of placebogenesis in one study decreased from 70% to 25% when the nurses' attitude towards the administered placebo was conveyed to the patients.  In another study, patients on placebo showed greater improvement than those on tranquilizers. The authors of this study  attributed their results to the unfavourable attitude of the nurses towards pharmacoiherapy.
It has been postulated by some ,, that the reasons for the continued decrease of patients from mental asylums since 1955 was not only because of the introduction of major tranquilizers, but also due to changes in the character of attendants, better nursing, greater understanding of the cases by physicians, increase in the number of staff members, greater optimism, and so on. It is difficult to perceive why such a phenomenon should have occurred at this particular period in time.
Changes associated with research activity have been reported to produce improvement in 80% of patients.  Multiple studies have concluded that research, while attempting to control some variables, often introduced other variables which were often overlooked by the experimenters.,,,
Most of the drug studies reported in the literature used volunteers as control groups, and all ,, except one study  had concluded that the basic assumption that volunteers constituted a `normal' group was untenable, as they had a high incidence of social or psychological pathology. The degree of normality or abnormality appeared to depend on the population studied. ,,
(d) Treatment procedures
Patients did not react uniformly to different treatment procedures. Several investigators have reported that 50% of patients do not take their medication or follow instructions about dosages. 
It is probable that prescribing exactly nine drops of liquid medication would be more effective than the conventional ten drops, , and greater placebo effect was observed with injections, various complicated procedures and impressive machines.  It was also reported that chronic psychotic patients when transferred from mental asylums to intensive care units, showed marked improvement in their symptoms even before the proposed lobotomy was performed. 
Psychiatric Clinical Diagnosis and Placebo Responses
Symptoms of anxiety and depression are known to respond to placebo more often than other symptoms seen in various psychiatric illnesses.
It has often been reported that overt, manifest free floating anxiety predisposes to a placebo response, ,,,,,, as stress and anxiety lead to an increased response to suggestion. This asumption has been disproved by the simple observation that patients with hysteria who are supposed to be very suggestible, but who have no manifest anxiety are poor placebo reactors. 
Illnesses which are not associated with significant manifest anxiety e.g., personality disorders, conversion and dissociative hysterical reactions etc., generally show a lesser incidence of placebogenesis. Presence of anxiety, tension, restlessness, and panic are often considered favourable for ultimate improvement. Anxiety is also a favourable prognostic sign in many other therapies like psychotherapy, psychopharmacotherapy, insulin treatment and lobotomy. 
Samuels and Edisen  in their study of 100 psychotics and neurotics (in-patients and out-patients) showed that 10% of the cases responded negatively to `placebo treatment' 25% showed no response, 29% showed mild improvement, 29% showed moderate improvement, whereas marked improvement was seen in 7% of cases.
Various studies are reported in the literature where attempts were made to correlate placebogenesis with some psychiatric symptoms. Placebo effects were frequently reported in patients with symptoms of anxiety and depression , Hankoff et al  Kurland , and Samuels  found no relation between placebo effect and a given diagnostic category. Shapiro  however found a definitely higher intensity and range of such placebo responses in psychotics.
Constancy and type of placebo effects
In the opinion of many workers like Fisher,  Hargreaves,  Rashkis,  Roberts  Samuels,  Shapiro  etc., severity or duration of symptoms of an illness did not directly correlate with the response to placebo. However, Black' in his study has shown that a short duration of illness was more likely to elicit a positive placebo effect.  Even the duration of time during which a placebo effect lasts is disputed. There are reports suggesting that the placebo response is short lived  though some authorities oppose such a views. ,
Many authors have reported that patients who showed positive placebo responses were also likely to show a more pronounced and quicker improvement in the underlying illness.
Some authors have pointed out inconsistency in placebo reactors. Various studies usually report a greater incidence of positive as compared to negative reactors. , It has also been shown that patients may not show placebo responses in one study but may respond to a placebo in another situation. , This phenomenon supports the opinion of some authorities that the placebo response is seldom uniform, constant or predictable. ,,
Factors like spontaneous remissions, chance emergence of short term symptoms, changes in social structure, changes in the environment etc. are also likely to predispose to placebogenesis. 
A close psychophysiological supervision may help to distinguish those effects which are truly produced by an active drug from those produced from the explicit or implicit response predisposition on the part of the subject,  and/or the experiment.  The phenomenon of placebogenesis has two facets; it is observed that just as a 'drug effect' may be produced when no drug is administered, on some occasions `no drug effect' may occur when an adequate dose of an active drug is administered.  Phychophysiological studies on placebos are of two types. 
Implicit and explicit responses need to be considered in any experimental situation.  In either treatment situation viz, drug or placebo, implicit sets or expectations exist e.g., an anxious subject change; even a basal study into a reactive one whereas an experienced habitual volunteer changes a typically reactive study into a basal study.  Such effects can be separated in a clinical study by repeated monitoring of physiological parameters Placebo effects tend to become complicated when other factors like drug conditioning are introduced into the experimental design. Beecher  found that the average effectiveness of placebo in relieving pathological pain (tissue pathology) was 35% while the effectiveness of placebo on experimentally induced pain e.g., by using tourniquet, heat etc. was only 3.2%. This difference in effectiveness was to be attributable to the placebo being more effective in the presence of anxiety and apprehension, which often accompanied pathological pain as opposed to pain in experimentally induced situations.
Sternbach  studied the role of explicit responses. In his study he gave three sets of instructions to three groups of subjects He instructed the subjects that one kind of pill would relax the stomach, that the second was a placebo pill, and that the third pill was a stimulant to the stomach The three pills were plastic coated magnets which helped to record the G.I. tract activity. In four out of six cases the activity of the gastro-intestinal system was found to vary with the experimental suggestions.
Lasagna et al  have emphasised the subjects participating in drug trials shout be initially studied to assess their predisposition to implicit and explicit responses.
According to Stroebel  a drug can be defined `as an externally supplied, solid, liquid or gaseous substance that is eaten, injected into, inhaled or absorbed by a living organism.' Such substances may be biologically viable (bacteria, viruses etc.) and may or may not result in any noticeable change in behaviour or physiological functioning. Drug effects may be direct or idirect (primary or secondary) depending on whether a response change was specifically attributable to the substance itself or whether the presence of the substance altered the role of other agents in exerting its effect. Thus, in addition to agents usually defined as drugs, such as aspirin, adrenalin or digitalis, the usual respiratory gases, and any contaminants they may contain, foods, beverages and even inert substances such as carbon granules in the atmosphere which may or may not be radioactive must be included.'
Such a broad definition obviously changes any study on psychophysiology into a study of drug effects, as many variables such as diet, smoking, alcohol consumption etc., are not carefully controlled. Some of the other factors that need to be controlled are room temperature, humidity etc. Smoking causes a peculiar problem since rigid control on frequency of smoking is itself likely to provoke anxiety in a heavy smoker.
One also finds multiple reports in the literature of studies on schizophrenic patients where an act of omission in controlling various constituents of their diet led to erroneous conclusions regarding physiological changes occurring in schizophrenia or about the effectiveness of a drug in this condition. ,,,
It is then obvious that a close psychophysiological supervision is most essential before any therapeutic benefit can be at-tributed to a pharmacologically active substance or any other treatment modality, as various environmental and dietetic variables alter the psychophysiological activity of an individual to a significant extent. To add to this, a placebo can also produce physiological changes simulating drug effect.
Thus, in the absence of recognised physiological changes that can be attributed to a placebo effect or inability to predict a placebo responder by ascertaining his psychophysiological activity before a study is undertaken, the mere knowledge of an individual's psychophysiological status does not help an experimenter in confirming or predicting the presence, absence, or probability of placebogenesis.
To conclude, the whole subject of placebogenesis can be summarised in the words of Pepper  who wrote that 'The giving of placebo-when, how and what seems to be a function of the physician, which like certain of the functions of the body, is not to be mentioned in polite society'.
The authors thank Dr. C. K. Deshpande, Dean, K.E.M. Hospital & G.S. Medical College, for the permission to publish this paper.