Jacobson: What led you to become a psychoanalyst?
Michels: To a psychoanalyst, of course, that is not a simple question. As an undergraduate, I was first exposed to psychoanalysis when I read Freud—several papers in a social science course. I don't remember which ones. The social science faculty at the University of Chicago included a large cluster around David Riesman. They were interested in an applied, sociological derivative of psychoanalysis. Their ideas were interesting, but I was much more fascinated by philosophy, physics, and mathematics. I was thinking of a career in that direction. When in the end I decided on biology, my interest was in neuroscience. That seemed the only way to find out about the mind, and about thinking. I saw neurobiology as a sort of reductionist infrastructure of philosophy. I started medical school planning a career in neurology and neuroscience research.
I remember two things pertinent to psychoanalysis from this period. One was from my freshman year in college—I was fifteen. I became friends with a kid on the floor who was from New York, and he invited a group of us to stay with his family during spring break. His parents seemed like nice people. Both were physicians, very busy in their practices. I soon learned that every hour for ten minutes we had to stay out of the front hall. My friend, who became a lifelong one, was Ernest Hartmann, and his parents were Heinz and Dora Hartmann, though I didn't know much about what they did at the time. The second thing I remember was from medical school. When bored by some of my preclinical courses, I would browse in the library. I remember coming across the Columbia Psychoanalytic Center catalog and reading about Sandor Rado and the history of the Center. I was intrigued, primarily by the explicit interest shown in the tie between psychoanalysis and neuroscience, which of course was Rado's influence. But my medical school showed relatively little interest in psychoanalysis, though we were exposed to a couple of analysts on the faculty. I remember Roy Whitman, now a training analyst in Cincinnati, who was then head of one of the hospitals where I worked, and who later became a friend.
When I finished medical school, I took a rotating internship, undecided whether to pursue residency training in neurology or in psychiatry. While still in medical school I had arranged a postresidency research fellowship in neurobiology, in Seymour Kety's group at NIMH, but toward the end of my internship I moved toward psychiatry, planning a career in brain research. I took my residency at Columbia, rejecting other programs because they seemed too narrowly psychoanalytic. During my first year I managed to spend three months doing neurochemistry research, and during my second year I took three
months of neurology. At that point, through clinical exposure to supervisors, I was becoming more interested in psychoanalysis. When I finished my residency, I went to NIMH for two years, taking a leave of absence from Columbia, where I was already a candidate. I arranged to take electives at the Washington Psychoanalytic Institute and to treat a patient with supervision while there. So my commitment to an analytic career developed during the course of my residency.
Jacobson: What were the personal determinants?
Michels: I was the second of two children—I have a sister two years older. My birth was associated with a great family myth. My mother was told after my sister's birth that she could bear no more children. When she developed the symptoms of her pregnancy with me, she was told she had some disorder, a hydatidiform mole, I think, and needed surgery. She was being prepped for the surgery when she felt movement. The obstetrician didn't agree, but she was convinced she was right and sought consultation, whereupon it became clear that she was pregnant. My mother was forty when I was born. Because of concerns about her health, I was delivered by cesarean hysterectomy, and after the birth my mother became depressed. I think there may be some special attunement to, and a feeling of responsibility for, the mood of a mother who has certain needs and affects, and this can predispose one to an interest in our field. At least, it seems this may have been the case for me.
I also found the intellectual curiosity about how the mind works—which informed my interest in philosophy during college—to be a continuing, powerful theme in my interest in psychoanalysis, more powerful than any meliorative desire or therapeutic zeal. When other children wanted to be firemen or policemen, I wanted to be a mailman so I could read everybody's mail. That kind of voyeuristic curiosity informs my interest.
Jacobson: And what determined your research interests?
Michels: My basic research interest was in understanding the process of thinking by studying the brain. Then I had three different mentors in biological research. The first, in medical school, was Harold Koenig, a neurologist. I was interested in the effect of learning on nucleic acid metabolism. We knew that nucleic acids had something to do with information, which gave them an exciting potential. The work looked at how the acquisition of reflexes by rabbits influenced the turnover of nucleic acids. In residency my mentor was Heinrich Waelsch. With him and Doris Clouet, I worked on a model of the induction of enzymes by conditioned experiences, studying tryptophan pyrrolase, an enzyme that can be induced, and seeing if I could condition its induction by coupling an experience with the biochemical stimulus for the induction. My third, and most serious, biological research experience was at NIMH, where, as I mentioned, I worked in Seymour Kety's laboratory. There my mentor was Lou Sokoloff. Sokoloff is most famous in psychiatry for his work on regional blood flow, but our work was on thyroxine and its mechanism of action. I was immensely proud of a paper I published in Science which focused on the differential effects of thyroxine on protein synthesis in the brain and other organs, which seemed to be developmentally relevant.
Jacobson: What led you away from this path?
Michels: Well, I liked the other things I was doing and I guess it became increasingly clear to me that I couldn't do everything well. More than anything else, it was probably the immediate pleasures of clinical, analytic, and educational work that I preferred to the daily routine of the laboratory. So after NIMH I returned to a very heavy teaching responsibility at Columbia. I had an NIMH grant as a career teacher and resumed my analytic training, as well as responsibility for various parts of the curriculum in the Department of Psychiatry.
Jacobson: You still wear many hats. How do you integrate these different Junctions?
Michels: So who integrates them? I remember a few years back when Charlie Brenner and I had a debate about the optimal career structure for a psychoanalyst. My memory may be distorted, but Charlie's position was basically that the more analysis one did, the better, and my position was that I wouldn't want an analyst who did nothing professionally all day long except analysis—it would be too narrow an experience. I enjoy a pluralistic rather than a homogeneous life. I do analysis. I do psychiatric consultations. I do some psychiatric teaching and supervision. I am responsible for a fairly large medical school, where I do a lot of executive things that go far beyond psychiatry. I also do some editorial work, ranging from the Journal of the American Psychoanalytic Association to the New England Journal of Medicine.
Jacobson: You mentioned that you trained at two different psychoanalytic institutes, chiefly at Columbia, but also at Washington. What was this experience like, and how did you manage it in practical terms, including your own analysis?
Michels: I started my analytic training and my analysis in my last year of residency, knowing from the outset that I had a two-year commitment to the Public Health Service. When I went to Washington, I was already committed to analysis, but knew that I'd be doing laboratory research. I had just finished my residency, and was afraid that I would get deskilled in two years without some clinical activity. Both Columbia and the Washington Institute were extremely supportive. I arranged to audit courses for two years in Washington, joining a class for its entire curriculum. To this day I remain a member of that group, with important friendships with its members. I was also assigned my first case in Washington. Afterward, I returned to Columbia and resumed my analytic training. I should mention an interesting similarity between the two institutes at that point. Each was extricating itself from the shadow of a major figure with special views. Columbia was saying we are not automatic followers of Rado, and Washington was saying we are not automatic followers of Sullivan. It was interesting that almost no Rado was being read at Columbia, and almost no Sullivan at Washington. I had naively thought that I would have a wonderful chance to learn each theory from its “true believers,” not yet understanding the sociology of how institutions change over time. But I learned a lot. I had a wonderful experience in Washington with a supervisor I grew to love. I had purposely selected a patient who seemed appropriate for what I thought of as Washington's strength. I also took a year of child analytic work there and liked that very much. And I felt no trouble fitting back in when I returned to Columbia.
Jacobson: What shifts have occurred in your analytic thinking and approach over time?
Michels: I think that my notion of analysis as a treatment has grown more sharply delineated and circumscribed. I've never been one for analyzing a broad range of patients. I respect other treatments in psychiatry, and have a fair amount of knowledge about them. I am most enthusiastic about using analysis when I am fairly clear that it is indeed the treatment of choice. That has led me in general to think of analysis as a treatment for people with characterological difficulties who have a fairly high level of the capacities necessary for participation in the analytic process and not much vulnerability to its negative potentials. Earlier, I was probably more interested in using analysis to treat various symptomatic conditions. In some ways I think the field has moved in two directions simultaneously—some the way I have, and some trying to broaden the indications.
Jacobson: Given your own view of the “narrowing scope,” how do you reconcile it with the kind of patients candidates tend to treat?
Michels: I think that optimally candidates should have exposure to some patients who are without question appropriate for psychoanalysis. But they should also have exposure to some patients who are at, or even just beyond, those boundaries, because part of learning to do something is learning its limits. I've suggested quite seriously in our Executive Committee that our requirements for graduation include not only a successfully terminated case, but also a failed case. Someone who doesn't know how to manage a case that hasn't had a good outcome isn't an analyst. But I think that candidates should not be asked to treat patients for whom
“I think that optimally candidates should have exposure to some patients who are without question appropriate for psychoanalysis. But they should also have exposure to some patients who are at, or even just beyond, those boundaries, because part of learning to do something is learning its limits.”
analysis is clearly not an appropriate prescription. It seems to me that candidates are sometimes asked to do that because “we can't find other cases.” If patients aren't available, we should select fewer candidates.
Jacobson: Do you feel the tasks of doing analysis and prescribing medication are so divergent that they should always be handled by separate practitioners?
Michels: I think the tasks are quite different. The problem isn't that someone wouldn't have the skills to do both—many medically trained analysts do. Rather, the problem is, can we perform both functions for a single patient? Does doing one interfere with our doing the other? My view is that it almost always does. The analytic community usually worries about the interference of the prescribing role on the psychoanalytic role. And though there is some interference here, I think it can be dealt with most of the time. I'm more concerned about interference in the opposite direction. A good psychopharmacologist must do things with his patient that analysts try to avoid doing, like making direct inquiries about symptoms, attempting to quantify them, engaging the patient as a collaborator in their quantification, and inquiring actively about side effects. The most common compromise made when analysts attempt both roles is that they try to get away with skimpy psychopharmacology—asking the fewest possible questions and waiting for side effects to come up in the associations, taking a chance on intuition as sufficient help in managing the drug.
Jacobson: Let's turn now to the present and future of psychoanalysis. You've written extensively on the role of contemporary psychoanalysis in medicine, psychiatry, and psychiatric education.
Michels: I'm always worried when someone says “you've written extensively.” It usually means “too much.” Let me just say that in much of the world Freud's concern about being swallowed up by medicine has led to a psychoanalytic community that is only partially medical. By contrast, in the United States the political factors around the growth of psychoanalysis in the 1930s led to a decision that it would be entirely medical, and a concomitant attempt was made to suppress nonmedical analysis. But nonphysicians have made rich contributions to analysis, and the attempt to suppress them was a serious error. At the same time, analysis has important medical roots, and I'd be concerned if analysis lost its ties to psychiatry and medicine. In recent years the trend has been away from medicine, toward a stronger definition of psychoanalysis as a subspecialty of several professions: psychology, social work, medicine, and maybe even some people without a primary profession—for whom psychoanalysis is the only profession. I'm worried particularly about this last tendency. I don't believe that analytic training alone provides sufficient professional socialization for someone to become a true professional—that is, someone who hasn't already achieved a professional identity in some primary field. And though I was an early advocate of enlarging the candidate pool to include people with doctorates in psychology, I will add a comment that might be unpopular. I'm much less convinced that other professions provide a professional socialization adequate to stand as the substrate for analytic training. I feel that analytic training by itself does not qualify one to make diagnoses or to prescribe psychoanalysis; these are essentially psychiatric issues. Training in analysis equips one to conduct an analysis, but not to prescribe it as opposed to alternate psychiatric treatments. That's a politically unpopular position, but I think it's true. I also think that there are very important potential interfaces with psychiatry and medicine that might be lost if psychoanalysis moves too far away from them. I think the current concern about the future of health care in the country, the
economic concerns of the mental health community, and the concerns of institutes about finding candidates might lead to a solution that has a short-term attractiveness but a long-term danger of demedicalizing psychoanalysis.
Jacobson: You touched briefly just now on how psychoanalysis is moving away from psychiatry. What about psychiatry moving away from psychoanalysis?
Michels: The first tendency is the more powerful I think. American psychiatry had a very strong analytic orientation in an era when there were very few other paradigms or schools that were scientifically or clinically interesting. In the last twenty years or so, however, other areas of psychiatry have become much stronger. For the most part this has had minimal impact on patients who are appropriate for psychoanalysis. On the other hand, it has wonderful implications for all kinds of other patients. This means that the intellectual leadership of American psychiatry is much more pluralistic than it was twenty-five years ago, and this is good. But some psychoanalysts view the trend as “antianalytic,” and have moved away from psychiatry because of it. Another problem is that academic psychiatry, like many other medical specialties, has become much more oriented around full-time academic scholarship. And of all the various fields of psychiatry, the one subspecialty that has had the most difficulty establishing itself under the usual academic criteria has been psychoanalysis. It doesn't have the usual kinds of research, or peer-reviewed appraisal of new knowledge, or criteria for rejecting obsolete theories. So it's been difficult in most medical schools for analysts to get the recognition they feel they deserve. Still, I think the major factor in any estrangement has been the analytic community's unwillingness to remain within psychiatry, with psychoanalysis regarded as one major paradigm among several.
Jacobson: Do you think that four-to-five-times-a-week psychoanalysis will survive?
Michels: I think there are many clinical situations in which frequency of sessions makes a difference. Personally I'm quite convinced that there are situations in which a higher frequency is better, but we should study this and collect data. I'm optimistic, though, that analysis as we know it will survive.
Jacobson: How do you account for the fact that so few people today are seeking it out?
Michels: It's clear that we haven't convinced the world that in certain instances it's the treatment of choice. And we certainly haven't convinced the people who make the decisions—referring physicians, patients, those who pay the bills—that it's sufficiently better to warrant the additional cost. It seems fairly clear to me that the reason analysis is in danger is primarily economic.
Jacobson: What is your guess of what the profession will be like several decades from now?
Michels: Already, of course, there's been a significant divergence between psychiatry and psychoanalysis. I haven't seen recent data, but I know that something like 25 percent or more of candidates entering analytic training in the institutes of The American are not M.D.s. There may be some threshold point here at which a vicious cycle will set in, and analysis will become progressively less medical. Entering it will come to be seen as leaving medicine and psychiatry rather than taking up a subspecialty. In some areas of the country at least, we may be approaching that point. So we're in an unstable phase. My guess is that the profession will be heavily shaped by the economics of the health delivery system.
My own perception is that until recently psychoanalysts were eager to say “We're psychiatrists too, and our treatment is a psychiatric treatment.” That argument is becoming harder to maintain now that it is asserted also that psychoanalysis can be done by all kinds of people who
“Psychoanalysis is more like surgery than like psychopharmacology—no two treatments are identical and the practioner's skill and creativity in the process of the intervention may be more important than the precision of the diagnosis prior to treatment.”
aren't physicians. I think the changing membership of the psychoanalytic profession will make it easier for those who have long wanted to draw the boundary for third-party payment between the biomedical and the psychosocial. I foresee this, along with the issue of degree of impairment, making it unlikely that analysis will be covered by health insurance in the same way that treatments for more serious psychiatric illnesses are. This will undoubtedly have an effect on the attractiveness of psychoanalytic careers. Demedicalization will occasion a shift in the profession's economic status, with earnings in psychiatry brought closer to general medical earnings and those in psychoanalysis falling to the level expected in the other mental health professions offering psychosocial treatments.
Jacobson: Let's turn to psychoanalytic research. Partly in response to these economic pressures, there seems to be a developing consensus that psychoanalysts need to do more research—testing theoretical paradigms, studying psychoanalytic process, outcome, and efficacy.
Michels: I don't think there is such a consensus. Many analysts pay lip service to the importance of research because they think it will be useful in their struggle for social and academic recognition, economic survival, insurance reimbursement, and such. A real consensus would mean that they felt research would give them knowledge relevant to their work, and I'm not sure our colleagues believe that. Short of that, however, research is silly, and sham research is more likely to hurt than help. I think many in the analytic community believe that analysis grows by a process that isn't scientific research—rather, it's clinical wisdom being collected and distilled and redistilled and exchanged in a process that is common in many academic disciplines but that isn't, at least in the narrow sense, scientific. Some believe that analysis would be enriched by scientific study that is rigorous and systematic and that includes falsifiable hypotheses, but so far most of the questions they have studied are of little interest to the majority of analysts. Whether interesting problems exist in psychoanalysis that are best approached by a traditional scientific strategy is a question that has not yet been decided. The enthusiasts are convinced on ideological grounds that there must be, and that we should devote resources to addressing them. The fellow travelers feel that it is politically wise to follow this route and therefore support it, but without great enthusiasm, while the critics keep quiet but cast their vote by showing no interest in research results. They do not find it at all plausible that their lives or their clinical practice might be changed by them.
Jacobson: You have also noted the need to develop specifically psychoanalytic instruments for conducting psychoanalytic research.
Michels: Outcome research has been a very hot field in psychopharmacology, and all kinds of standard methodologies have been developed for assessing and measuring diagnosis, baseline condition, change, and final condition. But these measurements are developed around implicit concepts of who the patient is, what the nature of the intervention is, what the problem is, and what the desired outcome is. These concepts are not treatment-neutral; they are biased toward the treatment that is being evaluated. The standard tools used for evaluating a treatment for panic disorder or depressive disorder are not good tools for evaluating an analysis, because the analysands aren't selected for that diagnosis, although they may also have it as a comorbid disorder; the goal of treatment in analysis is not simply a change in these symptom complexes, although that may be a secondary goal. But if you develop your measures around a side issue, the validity of the findings for the treatment you are interested in will be limited. Psychoanalysis is more like surgery
than like psychopharmacology—no two treatments are identical, and the practitioners skill and creativity in the process of the intervention may be more important than the precision of the diagnosis prior to treatment.
Jacobson: How do you view the influence of the current DSM system on psychiatry and psychoanalysis?
Michels: I think DSM-III, III-R, and IV reflect a broad trend in American psychiatry. The psychodynamic paradigm, as the architectonic organizer of all psychiatry, has waned so that now it is only one of several ways in which psychiatrists think about patients and organize treatments. With that shift there has been a rebellion against its dominance in the nosological system, and an attempt at developing what seems to me impossible—a theory-free nosological system. What we have now is a nosological system that is more phenomenologically oriented, and it has taken over the imagination and interest of most of American psychiatry. This has occurred along with a shift of attention away from neurotic problems toward the so-called major psychiatric disorders, their conceptualization and treatment, so that the great advances of the DSM system have been in those disorders in which analysts have little interest. The weakest and least valuable part of DSM is the description of precisely those disorders analysts are most likely to treat—the so-called Axis II personality disorders. There has been an unfortunate, politically motivated attempt to keep dynamic ideas out of this part of the system, on ideological grounds, rather than considering their specific potential for thinking about personality. In this regard, it's no accident that an interest in psychiatric nosology developed when we first developed specific, differential treatments for psychiatric disorders. It was of very little importance to differentiate a manic psychosis from a schizophrenic psychosis until we had the alternatives of lithium or phenothiazines. If we had differential treatment strategies for the Axis II disorders, diagnostic concepts and criteria would become immensely important and no doubt would be refined. But right now it isn't clear what we would do with a more elegant nosological system. I'm not sure who would buy the product once it was developed. I don't think clinicians are hungry for it, and I don't think our researchers are hungry for it yet.
Jacobson: Couldn't it be argued that we have a treatment that, though expensive, is demonstrably more effective than the alternatives for some patients, and since we haven't yet identified the specific subpopulations for whom it's the treatment of choice, we need the nosology to help us determine this?
Michels: Only if we're ready to do the kind of research that would develop those answers—that means controlled comparative efficacy studies of patients randomly assigned. I'm not sure the field is ready for that. Many of the friendlier critics of psychoanalysis have suggested not that it is an ineffective treatment, but that it hasn't demonstrated a differential superiority to treatments that are cheaper, quicker, or easier.
Jacobson: Bob, thanks for sharing your views.