[Page 1]

RESEARCH IN PSYCHOTHERAPY AND CHEMOTHERAPY:

Research Problems, and the Relationship between Psychological and Physiological Variables

Eugene T. Gendlin, Ph. D.

University of Chicago

Dr. Ewalt's impressive study has found that a combination of psychotherapy and drugs resulted in significantly more improvement with schizophrenic patients than psychotherapy alone.

In the report given to this conference nothing is said about the effects of drugs alone. However, since the hospitals are still full, despite an almost routine use of drugs, and since his chronic patients had been hospitalized at least three years before the study began, it is clear that drugs alone don't cure, although they may have effects. Five patients in the study were placed on placebos after taking the drug for more than a year. Their worsening was quite noticeable, as was their improvement when drugs were returned to them three months later.

The study thus implies that both psychotherapy alone and drugs alone effect improvements, but both together show significantly more. This finding should give pause to all those who would champion the use of either one exclusively.

I

In this discussion paper I want to give my main attention to the implications of this finding, which can be summarized in the word "both." But first I want to discuss the finding about psychotherapy alone, and some problems of psychotherapy research.

It is important to emphasize that Ewalt's study found psychotherapy alone leading to measurable and significant improvement. Psychotherapy research has been on the defensive for some years, more for socio-cultural than scientific reasons. Although it is a cardinal principle of scientific method that nothing at all can be concluded from no findings, there has been a tendency [Page 2] to do so anyway, as though it were in fact more scientific to emphasize where no findings were gotten. The news of no findings tended to outrun the news of findings. The old rule that nothing can be concluded from the null hypothesis, is based on the fact that no findings are very easy to obtain! One need only do a bad study, and one is sure to turn up no findings. Any well established effect can be measured poorly with inappropriate methods, or in a place where it happens not to be, and one can turn out any number of studies that fail to find it. It might be pouring rain, but if you measure it indoors with a thermometer, you might not find a measurable effect of it.

Can it possibly be imagined that some of our psychotherapy research methods and instruments might be poor or inappropriate? Oh, and how! Most of our outcome measures were designed for other purposes, and for stability, i.e., items sensitive to change were systematically eliminated in the making of most of our tests. Outcome measures have been shown to disagree and have failed to correlate with each other on the same cases (Fiske, '71). Thus many outcome measures are known to be inappropriate and poor.

Can it be imagined that psychotherapy research has attempted to measure psychotherapy in a place where it happened not to be? Oh, and how! Dr. Ewalt mentions the fact that most studies of psychotherapy have employed inexperienced therapists, usually psychiatric residents. In Ewalt's study the therapists of the chronic patients were experienced and analytically trained, but the therapists of the acute patients were again residents. These are not only inexperienced, having just arrived in the psychotherapy field after many years of medical school, they are in a threatened position in which their therapies are part of their in-training performance. Can we imagine that psychotherapy was actually not present, as a phenomenon, when it was attempted by these therapists? We can indeed.

[Page 3]

But let us reflect; how did it happen that research often studies psychotherapy when it isn't there? What sort of research methodology could lead to such a pass? Don't we at least know that we are measuring psychotherapy where it is? Isn't psychotherapy defined, albeit roughly, but well enough, so that we can test for its outcome? How is the "experimental" group usually defined? The usual operational definition of "psychotherapy" is that two people stay in a room together for an hour, and one of them claims that he is attempting psychotherapy. Now, this definition is not only redundant, since it uses the word "psychotherapy" to define psychotherapy, but, worse, it uses only the word. We have no check as to what is happening in the room!

Is it imaginable that psychotherapy, in whatever way we may roughly think of it, might not be happening in the room? Indeed, that is very imaginable. In measuring psychotherapy's outcome we want to measure, not how often it is made to happen, but if it has good outcomes when it does happen! If it does have good outcomes when it does happen, that will justify a greater social effort to train and offer psychotherapy, and a greater effort to discover just how and when it is made to happen, so that in the future we will know how to make it happen more often and regularly. We therefore must not confuse the question of the frequency with which psychotherapy is made to happen, with the question of its outcomes when it does happen.

Unless there is some way to define whether psychotherapy has happened or not, just testing for outcomes means little. Unless there are process measures indicating that the therapist and the patient actually engaged in certain kinds of interaction, why test outcome?

But the outcomes of a group of cases with experienced therapists intending to give therapy (if therapy is indeed something "given") ought to be measurable [Page 4] for the group as a whole. Such a result would be enormously weaker since it would be mixed with those cases in which no therapy arose. Thus we cut the power of any results of therapy down to half or less than half by this mixture of therapy and no therapy cases, when we lack a way to recognize our phenomenon before we test its results.

Suppose you did the same in a drug study. Would you accept it as a good test of the effect of a drug, if a study reported that some of its drug "active" patients took the drug and some not at all, that the dosages were unknown, and that it wasn't known which drug was used? Suppose further there was excellent reason to believe that more than half the "drug" patients did not receive any of the drug at all, but it wasn't known which ones they were.

Suppose further that we test the effects of this "drug" by using measures which had been designed to be stable in the face of many other chemicals, so that the least likelihood existed of finding any change. Would we then not be quite surprised if any findings at all emerged?

It is under these conditions, that psychotherapy research studies have found significant improvement more often than not.

In the research on psychotherapy with schizophrenics which I directed in Wisconsin (Rogers, Gendlin, et. al., '67), when cases were divided according to variables of therapist interview behavior, significant improvement was found for the high subgroup. When cases were divided according to patient in-interview behavior (EXP Scale), significant improvement was found for the patients high on the EXP (Rogers, Gendlin, et. al, '67) (Gendlin et. al., '67) (Gendlin '66) (Mathieu-Coughlin). We found little improvement for the "therapy group" (the "in the room" group) as a whole.

[Page 5]

Throughout the literature, whenever cases have been subdivided, significant differences have nearly always been found. Sometimes outcome is taken as the basis for dividing, success and failure cases are then found significantly different on process variables. Or, process variables are used to divide cases (either therapist behavior or patient behavior or both), and outcomes are then found significantly better both as compared to the low process group, and as compared to controls (van der Veen, '65) (Gendlin et. al., '60).

Because of these reasons, my work in the last decade has been on defining process, with the result that an EXPERIENCING SCALE now exists, applicable to tape-recorded interview material. In a series of studies (summarized in Gendlin et. al., '67), this scale has regularly selected the successful outcomes (by a variety of poor measures) from just a few bits of tape recording. There is, therefore, at least one operational process measure which defines interview behavior that has desirable outcomes. This lead should be followed up. Better, perhaps computerized versions of the EXP variable should be devised. The specific outcomes of the process defined by it should be formulated in psychometric measures specifically designed for them. The specific therapist operations leading to this process should be defined further (they are the three Rogerian conditions, cf. Rogers, Gendlin, et. al., '67), and more recently direct focusing instructions (Gendlin, '69). (And those who have observed other kinds of therapeutic process due to other therapist operations leading to other outcomes can devise specific process and outcome measures for their kinds of psychotherapy.)

Since there is at least one measure of whether psychotherapy is happening in an interview, or not, we need no longer consider this a devoutly wished-for impossible. Rather, it is an already existing essential which must be improved upon, and pursued.

[Page 6]

We must apply "process measures" to therapist and patient interview behavior, to know if psychotherapy is there. We must then test if it has improvement as its outcome, and devise new measures for the improvements we think we find. At least let us recognize that we cannot think of ourselves as testing the outcome of "psychotherapy" as if we knew it to be there from two people being in a room, with one of them having the intention of giving something verbally so called.

Thus it is no great puzzle, really, that at this very conference another study, (May), found that psychotherapy added nothing to the effect of drugs. The variable called "psychotherapy" is undefined.

Dr. Ewalt is highly aware of the problem. He calls psychotherapy a "non-specific" variable. He applied the Strong Inventory of vocational choice (Whitehorn & Betz, '56), and found (again) that the B therapists did better than the As. This shows that by defining the therapist even on this rough A/B measure (the Strong test was developed not for therapy but for vocational choice!), we get a significantly better subgroup. (Even so we don't know if therapy happened in all or just one of the B cases.)

It is time we threw overboard the dead weights with which we have sailed in the past. The word, room, and intention of psychotherapy do not define it. Not only do we not know in which cases psychotherapy occurred, but psychoanalytically-oriented psychotherapy is an utterly different attempt to make therapy process happen, than say Rosen's, Laing's, or Haily's, or other more modern, active and interactional approaches. Nothing is shown by no findings, but also: little can be concluded from findings from one therapy about others. If we do not agree on the interview behaviors of therapist and of the patient that constitute psychotherapy, let us distinguish between them and measure the outcomes of each.

[Page 7]

Granted, psychotherapy process may be relatively rare. But that is no reason for trying to measure its outcome when it isn't happening. If rare, then especially let us not mix the specimens of our phenomenon with lots of cases in which it was not there. Let us collect cases in which naturalistic reports indicate its presence and define both process and outcome. Once we learn what therapist and what patient behaviors have outcomes we wish, then we can work to make these behaviors less rare.

II

What are the implications of the finding that psychotherapy and drugs together are significantly more effective than psychotherapy alone? Two implications follow:

One) We must give up the parochial loyalty we may hold to either alone. It does not behoove scientists or healers to cling too long even to deeply felt hunches. We may have come into the field feeling deeply that psychological factors affect the body and account for much physical illness. Or we may have come, feeling initially that psychological factors always involve somatic ones. However deeply we feel one of these, and however much progress there may be in one of them, we must recognize that both are true.

Two) But it isn't enough simply to assert both. The split between physiology and psychology cannot be healed simply by denying that split and asserting in general that the organism is a unity. After such a helpless and fuzzy assertion of unity, we mostly go on as we were, which means that we go our separate ways again. What is a more trenchant way of thinking? What is an operational way of pursuing this unity of mind and body, which is again brought out by Ewalt's finding of "both"?

[Page 8]

Let me cite a few other findings which also point in the same direction:

When scientists first studied LSD about a decade ago, they regularly found it resulting in highly unpleasant paranoid conditions, and they therefore called the drug "psychotomimetic." It produced only "bad trips." But, more recently, when young people used the drug, they found that it could produce "good trips" as well. Why did they discover what devoted, open-minded scientists had missed? Because psychologists studied the LSD drug subject alone in an empty whitewashed room (to avoid confounding factors). The experimenter would not even be present. Instead, to avoid confounding the drug's effects with his own presence and variability, the experimenter observed the subject through a peephole. Under these circumstances the drug produced paranoid experiences. Today we know that LSD, although one and the same chemical, applied to one and the same human body, produces very different results depending upon the interactional conditions. When taken among friends, with music, it can produce results people report as desirable.

Nor is this relation between chemicals and interpersonal activities true only of humans. Even rats show the effects of amphetamine, for instance, only when they are put together with other rats. Then only does their blood pressure rise with amphetamine. When left alone, and studied as individuals, amphetamine has no measurable effect (Moore, '64).

The implications are several:

We must overcome the illusion that we are studying individuals (or effects on individuals). In different interactions the "same" individual body is differently ongoing. The individual's ongoing living is always interactional, it is always with others in context. We are not studying the patient, but the patient-in-hospital [Page 9] interaction (of a certain kind), or patient-in-specific-therapy interaction, or patient-in-the-relationship-said-to-be-psychotherapy-but-it-isn't, or whatever the ongoing interactional life processes happen to be. They need defining, operationally, measurably.

We must give up the reductive presumption that there is such a studiable variable as "body A (say the human body), when put together with chemical substance B." Defined only this way, the variable does not have definable results on the behavioral level. (It may have definable results in some physiological variables such as blood concentrations, but not in others such as blood pressure.) The interpersonal, interactional context and process must be defined too.

On the behavioral level "body plus chemical" is indeterminate. Far from puzzling over how to put interpersonal interaction together with chemotherapy, the fact is that the chemotherapy variable cannot be defined alone. It requires specifying the interactional circumstances, before it can have uniform research outcomes.

If we want to know the effects of a drug, we must add the question, under what specific interactional and life circumstances? Only then can the question receive a scientific answer. Until this factor is specified, drug researchers are likely to produce contradictory findings.

The body is clearly involved in all physiological processes. "How to put the body together with psychology?" How did this problem even arise? The problem arose, not because the body is in any way hard to relate to psychology and interaction, but because physiology ("-ology" is the study of) uses a theoretical model of cutting into structural units. This quite successful theoretical model has given us a great deal of medical progress and we needn't throw it out, [Page 10] but we must see beyond it. It is a way of studying. The living body isn't the physiological model's "body." Their "body" is studied as you might study a machine, an automobile, say, which can be kept perfectly healthy without running. Say we keep it on four bricks in a dry garage. The automobile will be fine there, and will run again when you wish. The living body, and all the more, the human body, must function and interact, or it dies. Its physical processes are possible only interactionally, and it is physically different in different interactions. When chemo- is combined with therapy to form chemotherapy, then this fact must be taken into account with some further thinking and research specifications concerning the interactional processes which are ongoing when a drug is being evaluated.

The routine administration of drugs in hospitals today has made wards quieter and better to live in for both patients and staff, but has not emptied the hospitals. It cannot provide patients with ways to live well. This cannot be done in hospitals, nor can patients do that as lone individuals, no matter how much drugs improve them. The living functioning requires really effective "milieu therapy," and more, it requires helping the patient establish a new life outside, and a new life style, finding people with whom he can live healthily if indeed, drugs have made it so that he now could. No one can live healthily in a hospital.

It is neither an accident, nor a contradiction, that the family and community emphasis in our field today is contemporaneous with the emphasis on chemotherapy. The two not only go well together, but they must!

Let us therefore radically alter our concepts, so that, instead of being surprised that an interactional process like psychotherapy should produce its best results with drugs, let us instead, when we hear about psychotherapy, ask: "But what [Page 11] body-life regime were the subjects engaged in?" And, similarly, when we hear of a drug being administered, let us, as if we had only heard the half of it, ask: "Yes, but what interactional living processes were the subjects engaged in at that time?"

Let us change the concept "body," "soma," or "organism" to include not only the physio-chemical structures the physiologist studies but also the oxygen and other people's reactions which the living of this body entails, and which differentially change it in both short and long term ways.

Let us similarly change our concept of "psychotherapy" or "psyche" or "subject" to include not only what we wrongly think of as psychic "contents," but also as the concrete bodily tissue-and-nerve-changing that living involves as it goes on. Then psychology will cease to be ghostly, and physiology will cease to be mechanical.

One ongoing process inherently unites both the psychic and physical statics of the other models with which we have studied psychology (psychic "contents") and physiology.

Psychotherapy research alone cannot proceed until it specifies the interaction process (or processes) called Psychotherapy. Nor can chemotherapy research contribute much without doing exactly the same thing!, whether subjects are in therapy and this is to be specified, or whether they are in the hospital generally. Obviously, if Ewalt, and even more, May, find drugs effective, this must involve some factor not generally abroad in all the filled-up hospitals now giving out drugs in copious quantities.

It is clear that in some hospitals (under some interactions conditions both in and outside of the hospital) these randomly given drugs are now effectively helping many patients, but it is equally clear that all too often they don't. [Page 12] And if we transcend the models of psychic and physical static structures, and think of one ongoing interaction life process, this isn't surprising!

Why are drugs so much more effective in some studies than in others and in hospitals generally? Under what interactional circumstances, and in what treatment contexts, are they given? If we knew this, we would know what to add to current hospital routines, to enable the drugs to work!

Let us specify the usual hospital interactional context of the usual drug regime: drugs given routinely, to every patient, according to mass regimes, more often place a blanket over an individual's feelings, beneath which there is no improvement. With overdoses, the patient cannot even work anything through alone, and complains constantly of being tired. Left sitting by himself in some chair in the day room, addressed only to be called to order, punished for infractions of pseudo-military disciplines, unable to talk to the doctor who only walks briskly down the hall every other day, the patient has little functioning available. Scores of professionals troop through the ward, but make little effort to relate with him (nurses, attendants, social workers, gray ladies bringing cookies, nurse's aides, RT workers interested in making him jump and OT workers interested in making him make belts). Milieu meeting discusses the time the TV is turned off at night, or analyze the bad behavior of Mrs. Culbertson in breaking the TV. He may never quite find the way to express himself so as to make anyone want to listen, or to live in personal closeness, if he did indeed still feel that people can relate to him. What effects do drugs have in the course of this type of interactive living? (Even then, as we said earlier about the finding on psychotherapy, even then the result of drugs is on the whole capable of showing up as a finding of improvement.)

Again, we must specify, we cannot only state the general hope that chemotherapeutic [Page 13] discoveries will be used "judiciously." We have to specify for underpaid and overworked hospital staff just how the drugs are to be used, and on whom, and with what type of interactions, and when. We must change the structure of hospitals and their release procedures. Otherwise the drugs enable no functioning and only deprive patients who get no other service even of the ability to work their disturbances through themselves, and remit "spontaneously" (as that is called).

On the other hand, together with the crucial factors of psychotherapy, out-patient help, a new group of people with whom to relate and a new place to live and work, the drug may very well add another perfectly crucial factor!

At the University of Chicago, the psychology students have set up their own service network. When a young person, often from out of town, needs aid, a team of them (not only one) meet the person and try to do something about his needs. These needs may be therapeutic, they may also be for help in finding a place to stay, or with fixing a car, or to come off a drug "trip." New friends and other contacts are often set up for the individual. Referrals to any of a long list of other resources are often made. If the individual chooses to go to the hospital, he is accompanied and visited. In the case of one young man, after some weeks, the hospital seemed to him his only choice. He had been hospitalized twice before, dreaded it, but saw no way out of the strong suicidal and homicidal urges which had reoccurred. Holding the hospital off for one more day, he was sent to an outpatient drug program, where a psychiatrist prescribed a drug. The "psychotherapy" (more exactly, the community aid) of the students, and the drug, were able to do together what neither factor alone had done before. The individual is now well, happy, and active in a new social and work life. The students endeavored to help in any way and with any resource, they were not limited to interactional or chemical aids. It seems to me, supervising this work, that if the young [Page 14] man had been given the drug in the hospital, it could not, there, have resulted in the new and better living process which immediately commenced when he took the drug in a positive interactional context in which better living was possible. He had taken drugs in hospitals before, with known results which had been both poor and dreaded.

The implication is that drug-in-hospital doesn't have the results that drug-in-psychotherapy may have, or that drug-in-new-community-activities may have. I believe that this is also the implication of Ewalt's finding, and is borne out by the many examples I have cited, as well as by other current writers in our field (Mowrer, '71).

Let us therefore give up the false pride of considering "crucial" only the one factor we like most to offer or to study. The very fact that I do only one thing, that I study only psychotherapy and work with patients only in terms of psychotherapy is an artificial result of how we study. It is an "ological" result, if you can forgive a neologism. I was trained to work with an individual because we have so long studied individuals, individual psyches and bodies, instead of interaction. I was not trained even to think about the physiological side of the processes I myself engage in with patients. I was equally untrained to deal with the patient's relatives—a social worker was supposed to do that! In my years of working in a hospital I found that if I was really to get a patient out of the hospital, and then really well, I had to take on the function successively of therapist, ward physician, social worker, superintendent, vocational rehabilitation worker, and follow-up worker (of which we had none. . .)

I often came into difficult relations if not conflicts with these workers, since they wanted the patient to become well first, in psychotherapy alone, (as Ewalt's study measured it), and only then would the superintendent do the [Page 15] work involved in reassuring the angry little town that the patient came from, that he should be allowed out of the hospital. The social worker wanted the patient well before she took the time to work with his relatives for "possible release." I wanted her to do that work while release was not yet "possible." I wanted the Rehabilitation department to give my patient one of their scarce job resources while my patient was only ready to work on conflicts about going out to a job, not yet ready to keep a job. I needed to set up living quarters for patients, so they had some other place than home to think about going to, and to feel about going to. Again they thought in static concepts about these factors, rather than as changing the patient. And again, none of these factors was I trained to work with. I learned that I had been trained too narrowly.

The same conclusion follows between psychotherapy and bodily chemical aspects. Recent work on experiential focusing (Gendlin, '69) (Gendlin & Berlin, '61) (Oberlander, '67), and some other methods, directly involve the bodily processes occurring when psychotherapeutic change occurs during an interview. It is not that psychotherapy must work "alone," only our concepts set up "psycho-" and "physio-" events as separate.

Let us recognize the chemo-psycho gap as an "ological" split due to different models and modes of studying, rather than as two different static phenomena.

Renewed interest on the part of youth, including young psychologists, in Yoga, relaxation meditation, environmental surroundings, diet, and also drugs, has made for a generation of people accustomed to "thinking from here" (as our youngest faculty member said to me recently, pointing to his viscera). These newly re-emerged topics all involve psyche and body together, and neither as static contents or structure of parts. By operationally specifying both bodily and psychological factors of the ongoing interactional process, we will move toward better concepts to delineate this one process.

[Page 16]

One) We are not studying the individual, but the individual-in-interaction.

Two) We are not studying set entities like drug G, body B, or psychic contents, or treatment intention D, but interaction process. Depending upon the interaction process, drug A, and treatment-attempt D will have different results, and body and psyche will be different. Neither body nor psyche are entities, but are engaged in interaction process.

Three) We must expect contradictory findings from (seemingly) the same studies, such as Ewalt's and May's until we define the ongoing interaction during psychotherapy interviews, and during the period in which the effect of a drug is being evaluated.

Four) To be sure, drug and also psychotherapy will have effects "alone" (the other is always also there, although undefined for a given study). It seems likely that the hallucinatory and bizarre aspects of psychosis will yield to chemical treatment, and will turn out to be definable in current physiological terms. But this will not yet deal with the question. We want to restore normal functioning. It may be, that some physical condition analogous to diabetes will require some people to add some drug to their bodies to be able to function normally. (This physical condition may be the long-term result of poor interactional living, or originally physical.) The normal functioning will require psychological and interactional community factors as well, and without them we will still have full hospitals.

Let us work for replicable chemotherapy and psychotherapy research which specifies ongoing interaction, and let us include psychic and bodily factors. Psychotherapy is already being defined in more directly felt, bodily, physiological processes during psychotherapy (Gendlin, '68 & '70). Let us define the effect of drugs in three ways, if we wish replicable behavioral effects: the body to which the drug is being administered (human or rat), the drug itself, [Page 17] and the ongoing mode of living and interaction.

Eugene T. Gendlin