Are there “Pillars of Therapeutic Wisdom” for Psychoanalytic Therapy?


The author-using the objective method of the social sciences-covers research dating back to 1930 on the therapeutic effectiveness of psychoanalysis, to which he adds the results of his own research conducted in Germany in the 1980’s and ’90’s. He examines the level of psychological health vs psychiatric severity, the therapeutic alliance, core relationship patterns, and the process from working through to mastery. He claims that these core concepts are true building blocks for clinical work in the varieties of psychoanalytic work.

Nothing is more difficult to determine than the future. However, we are all aware that nothing is more predictive than past accomplishments. For this reason we will start by taking a look at the past. The kind of wisdom(1) that I shall try to outline will be based not on clinical beliefs but on the findings of empirical research.

Research on therapeutic outcomes had a good start when the Berlin Institute commissioned Fenichel and his colleagues to prepare an evaluation of therapeutic effectiveness. In 1930 they proudly published their report.

The criteria of evaluation were most stringent, however, and were not applied by independent evaluators, but supplied by the analysts themselves-a measure of success we today are not very satisfied with.

Other more serious problems surfaced when Eysenck published his negative evaluation(2). As Bergin(3) pointed out, Eysenck included the dropout-today this is called “intent-to-treat analysis”-whereas Bergin made a much more favorable estimation of therapeutic outcomes, not including dropouts in his calculations and so he arrived at an amazing 91% improvement rate unmatched until today.


dropouts included.
improved = much improved

dropouts excluded.
improved = moderately improved or better
Alternative A
-dropouts included.
Improved = moderately improved or better
Alternative B
(Knight 1941)
dropouts excluded.
improved = much improved

The summary of Bergin is thus as follows:

The four divergent but equally reasonable tabulations of the Berlin data clearly establish my point that there is no valid way to assess the effect of psychoanalysis from the information available. I can see no clear justification for choosing one interpretation over another, even though I do have personal biases in certain directions The ambiguity in these data cannot be resolved(4).

Bergin adds some other observations to his careful and quite sophisticated evaluation, namely that the divergences in estimation of outcome are more pronounced where psychoanalytic treatments are concerned, and less so with eclectic therapies. Even though Bergin admits to not being an advocate of psychoanalysis, he concludes that the results until 1952 must be considered encouraging, though certainly not dramatic. “It is of particular interest, however, that the longer and more intensive the treatment, the better the results”(5).

Thirdy years have passed since Bergin’s careful review appeared, but why do we still have problems when we have to debate the outcome of psychoanalytic therapies? The development of an organized field of psychotherapeutic research has been fostered by research-minded psychoanalysts such as Robert Wallerstein, Otto Kernberg and Lester Luborsky-especially by their enduring performance in the show-case of psychoanalytic research, the Menninger Psychotherapy Research Project-as well as by researchers trained in psychodynamic such as H. Strupp, K. Howard and above all D. Orlinsky, who is remembered today as the founding father of the Society for Research in Psychotherapy. This society generated a research culture where the relentless demand: “show me your data!” prevailed over the dominance of clinicians’ opinions. The emphasis was on evidence and not on beliefs.

This research ethos has set new standards for the evaluation of psychoanalysis and therefore our position today can no longer be based on Bergin’s supportive statements. Psychoanalysis cannot escape the methodology of social science, be it qualitative-narrative or quantitative-psychometric in its approach. The idea of psychoanalysis as a special and unique science that creates its own methodological standards may be a comforting idea to some of us, but when it comes to the evaluation of outcome, then society and its agencies, the consumers of our treatments and our patients’ relatives and loved ones cannot be satisfied with an idiosyncratic believer’s view of psychoanalysis as an enterprise in which “the way is the goal”. One might maintain that for the therapeutic philosophy of the practitioner this point of view remains vital; however the demands of public accountability require more than this. If psychoanalytic therapies were seen as being a cultural event sought by customers willing to pay for a special treatment, no scientific achievements would be necessary but just agencies for organizing these cultural events. In fact, an increasing number of psychoanalysts do not share this view. They believe that Freud´s legacy entailed considering psychoanalysis not as an instrument for discovering how the mind works but using psychoanalytic therapy as an instrument for initiating beneficial change.

The broadening scope of psychoanalysis has led to a remarkable development. Not only do we have a variety of psychoanalytically oriented treatments, we now also have to accept that the very notion of psychoanalysis as a self-sufficient form of treatment can no longer resist empirical scrutiny. Therefore, Wallerstein raises the question of which form of psychoanalysis one should employ in specific cases, asking whether one person’s proper psychoanalysis is someone else’s ‘mere’ psychotherapy? These considerations led him to pluralize the word “psychoanalyses” in the sub-title of his latest book(6). The state of outcome research in psychoanalysis requires the acknowledgment of that kind of diversity.

The recently published overview on outcome studies regarding psychoanalytic oriented treatments, the IPA OPEN DOOR REVIEW(7) was compiled with a view to furthering our understanding of the problems and chances provided by the philosophy behind psychoanalytic treatment.

For this reason the said review has been very liberal, in that it includes studies of many varied psychoanalytically informed treatments, as they all do contribute towards an assessment of the state of the art in psychoanalysis.

This is my first pillar of wisdom: it seems wise not to separate psychoanalysis proper from other psychoanalytic oriented therapies, but to favor the term psychoanalytic therapy to cover all treatments that use the basic notions of the theory of psychoanalytic treatment.

For many years research in psychodynamic treatment has accumulated an impressive array of therapeutic factors that have been shown to systematically influence the treatment outcome. It is likely that these findings apply to all forms of psychoanalytic therapy.

What are the other empirically supported pillars of therapeutic wisdom?

I. Level of Psychological Health vs Psychiatric Severity

Luborsky et al.(8) have summarized many studies using this dimension have shown that the average correspondence of observer-rated prediction of psychological health-sickness to the actual outcome was 0.27.. This finding is a reminder that psychoanalytic therapy like any other form of psychotherapy deals with the fact that our treatment tools are limited by the patient´s capacity to use them. However a number of findings point to the fact that more intensive psychoanalytic treatments benefit patients with more severe disorders. For the survival of the longer treatment domain these findings are encouraging; further comparative prospective studies of this issue are essential.

II The Therapeutic Alliance

# The therapeutic alliance has become the most researched clinical concept based on Freud’s initial formulation(9). It is now evident that the construct of this alliance has both a direct and an indirect impact on outcome and it is increasingly evident that this is true not only for short term treatment but also for long term analytic therapies(10).
The very extensive research has now passed the level of mere demonstration of its relevance to outcome. There is agreement that the therapeutic alliance is a multi-dimensional construct composed of four relatively independent dimensions(11):

a) the patient´s capacity to purposefully work in therapy
b) the patient´s affective bond to the therapist
c) the therapist´s empathic understanding and involvement
d) the patient and the therapist´s agreement on treatment goals and tasks

III The Core Relationship Patterns

Measures of the core construct of psychoanalysis, transference, have grown rapidly in recent years. After many years of simple research based on opinion retrieval, the advent of transference-related measures based on tape-recorded sessions has made clear that there is a fundamentum in re, i.e. that there are empirically demonstrable findings. Beginning with Luborsky´s discovery of the CCRT in 1977, a plethora of measures have been developed(12). The CCRT was followed by the plan diagnosis by Weiss & Sampson(13), the Dahl & Teller frame analysis(14), Gill & Hoffmann´s PERT(15), Strupp & Binder´s dynamic focus(16) and many more. The measurement of the core construct of transference thus became viable and psychodynamics moved towards becoming a basic science(17). A recent review by Henry et al. counts seventeen different methods that in varying degrees are in the process of being validated(18).

Our own recent study, carried out at three different universities, contributes toward ascertaining the validity of the CCRT-method. The connection between the “valence dimension” of the responses from others (RO), responses of the self (RS) and the severity of the psychic disorder has been analyzed within RAP-interviews of 266 female patients.
Both therapists and patients evaluated the severity of the impairment in a similar way. The main finding was that the more the patients are impaired, the more negatively they describe both their own reactions and those of their partners interaction, as shown in the relationship episodes(19).

In the context of this research, detailed micro-analytic studies were initiated. Bucci’s ‘Referential Activity'(20) has become a very successful tool for studying the basic mechanisms that are systematically connected to transferential processes.

IV The Interpretation of the Core Relationship Pattern

Many studies on the interpretation of transference-once thought to represent the top of psychoanalytic work-have revised the unanimous overestimation of this activity. Especially in the low frequency and/or short term psychodynamic treatments, several differential findings have enhanced our understanding of the problem(21):

= more is not better and may even be damaging
= transference interpretations do not necessarily repair poor alliances and may damage an already existing alliance
= transference interpretations do not elicit greater affective response or necessarily increase depth of experience, when compared to non-transference interpretations
= Interpretations are more likely to elicit defensive responses than other types of interventions
= the average level of therapist´accuracy may be much lower than assumed
= Strachey´s concept of the mutative interpretation has not yet been empirically demonstrated
= the therapist´s skill may make a difference to the outcome
= the quality of object relations mediates the response to both frequency and accuracy of interpretation

Another conclusion of this research is that little research has been done regarding high frequency long-term psychoanalytic therapies. Given the lack of public data in this field, we are still far from empirically knowing about the role of transference in analysis. The scientific community therefore awaits with great interest the findings of Kernberg et al. based on their ongoing systematic studies on Transference Focused Psychotherapy(22). For the severely affected patient population, the role of transference and the interpretation of transference will be at the center of research efforts. “However, in contrast to psychoanalysis (where a systematic focus on transference is a major treatment strategy), in the psychodynamic psychotherapy of borderline patients transference analysis is modified by attention to initial treatment goals and current external reality”(23).

V Working through to Mastery

The least understood “pillar of wisdom” is most probably the “working through” in order to achieve lasting effects. In traditional language Thomä & Kächele have described this phase of analytic treatment:

“The phase of working through begins after the patient has gained insight into the connections and processes marking the dynamics of previously unconscious conflicts. The goal is to use cognitive and affective insight to change behavior. Even if some patients achieve such behavioral changes without the analyst’s assistance, this is something that cannot generally be expected”24.

The goal of this process, called ‘mastery’, is defined as the acquisition of emotional self-control and intellectual self-understanding in the context of interpersonal relationships. The Australian psychologist Grenyer has developed a scale of measuring mastery that has been applied in a study with Luborsky, which demonstrates systematic progression to greater levels of mastery in interpersonal conflicts25.

There are other ways of expressing what needs to be changed and what needs to be better managed by the individual patient. In modern language Fonagy26 suggests three basic process dimensions that should be the focus of change. They are as follows:

(1) Intersubjective representational shifts;
(2) Changes of mental processes
(3) Changes in mental representations.

According to Fonagy, for the majority of neurotic patients the last of these three processes may be enough:

From a process point of view, representational changes are in the direction of a fuller and more elaborated representation of the mental states of internal objects and the self. Enhanced reflective capacity allows patients to integrate split-off parts of the self and create object representations with complex thoughts, mixed emotions, and differentiated desires. Symptomatic improvement should be associated with such changes. But from a technical point of view, we have argued that it is important for the analyst to be aware that the changes sought are not changes in the patient’s awareness of past events but rather changes in procedural and implicit memory. Thus recovery of past experience may be helpful but the understanding of current ways of being with the other is the key to change. For this, both self and other representation may need to alter and this can only be done effectively in the here and now.

For more disturbed patients, especially personality disordered individuals, the other two processes are essential. To undo representational shifts, the analyst must be able to permit externalizatons in order for the therapy to be tolerable for the patient. Secondly, the analyst has to help the patient to engage his mind in forms of mental activity which have felt dangerous in the past. The development or recovery of the reflective function stands out as the hypothesized core process.
As we have now dealt with what most psychoanalytic treatment researchers consider to be the building blocks of all and any of the psychoanalytic therapies, let me conclude with some statements on the state of the outcome research regarding the psychoanalytically oriented treatments.

Some unbiased observers realize that there are considerably more studies on long term psychoanalytic treatments than most of us are aware of. This is especially true if the observer happens to be a native English speaking person with no mastery of a foreign language (which is often the case). In fact, a substantial amount of these studies have been carried out at European University Research centers. In general they share a preference for effectiveness-methodology. Very few studies of the longer-term psychoanalytic therapies seek to fulfill the gold standard of experimental efficacy research. “When reality doesn’t fit the blueprint”, researchers can tell sad stories of how a planned experimental design failed but an intelligent data analysis is able to retrieve relevant findings(27). We should not reject randomized studies in principle; sometimes it can be successfully done as demonstrated by the ongoing Munich study comparing once a week psychodynamic with three times a week psychoanalytic therapy on patients suffering from major depression(28).
Most studies have major limitations which might lead critics of the discipline to discount their results. Others have such serious limitations that even a sympathetic reviewer might be inclined to dismiss their findings.

For example, is the analyst in a position to judge the outcome of a treatment? Not only is there the issue of a self-serving bias, but is the context of free association not also totally incompatible with the systematic gathering of data concerning adjustment and the like? At present we feel the answer is straightforward: the analyst´s view in evaluating the outcome is one among many others, it has to be compared to the patient´s view, and to the views of his relatives or friends, and it has to be compared to independent assessment with well designed instrumentations.

The most common problems that we encounter in many studies are still: lack of standardized diagnoses,
inadequate specification of treatment procedures,
lack of control for selection biases in sampling,
absence of intent to carry out controls and the failure to follow up drop outs,
use of inexperienced therapists,
lack of homogeneity of the patient groups considered,
heterogeneous methods of intervention with the related lack of a generally accepted manualized method of intervention,
lack of statistical power,
lack of random assignments to treatment groups,
lack of independent assessment of outcome,
lack of standardization of measures of outcome,
questionable validity of some outcome measures,
poorly matched comparison groups,
absence of control for the law of initial and of regression to the mean,
failure to take adequate baseline measures, and related to this reliance on retrospectively collected data,
inadequate detail on statistical analysis and inappropriate statistics reported,
inadequate control for intercurrent treatments,
and so on.

Up to now there are no valid methods available that might definitively indicate the existence of a psychoanalytic process although process research has made substantial progress in the analysis of what constitutes the essential features of analytic processes in contrast to other therapeutic processes29.

Notwithstanding their many limitations, the sheer number of studies available is encouraging, particularly the range of ongoing studies. The OPEN DOOR REVIEW was by no means an exhaustive review. In fact this review was labeled “open” in order to underline the intention of including further studies in future as these are brought to the attention of the review board. Many of its conclusions should therefore be heavily qualified in the light of the questionable internal validity of the observations reported.

1. Outcome of Psychoanalysis

1.1 Estimates of the percentage of patients benefiting varies widely across studies, even for similar conditions and similar measures, probably as a function of methodological factors
1.2 Completed analytic treatments are invariably associated with greater benefits
1.3 Longer treatment has better outcome
1.4 Psychoanalysis can bring the functioning of a clinical group to the level of the normal population
1.5 Intensive psychoanalytic treatment is generally more effective than psychoanalytic psychotherapy
1.6 Superiority of psychoanalysis over psychotherapy sometimes only becomes apparent some years after treatment has ended
1.7 Superiority of psychoanalysis over psychotherapy is sometimes not maintained at long term follow up
1.8 Psychoanalysis can lead to a reduction in expenditure for health-care services and this is the case for a number of years after therapy ends

2. Patient variables
2.1 More severe disorders benefit from psychoanalysis rather than psychotherapy
2.2 Psychoanalytic therapy at sub-clinical doses may have negative outcomes
2.3 Behavioral disorders respond less well to psychoanalysis than emotional disorders
2.4 Younger children benefit more from psychoanalysis than older ones
2.5 Anaclitic problems are better dealt with in psychotherapy, introjective problems in psychoanalysis
3. Process-outcome variables
3.1 Successful psychoanalytic treatment of severe personality disorder may require a combination of supportive and expressive techniques
3.2 Therapeutic alliance at the beginning of treatment predicts outcome
3.3 Supportive therapy may be better for psychotic patients as regards improved capacity for adaptation
3.4 Anxiety, guilt and idealization in transference may be associated with successful treatment whereas shame, humiliation and existential anxiety are associated with failed treatments
3.5 Successful therapists compensate for the patient’s affect
3.6 Systematic changes in dreams illustrate structural changes
4. Therapist variables
4.1 More experienced analysts not inevitably the most effective
4.2 Perhaps classical psychoanalytic attitudes are unhelpful in psychoanalytic psychotherapy
4.3 Match between analyst and patient is a key predictor of outcome
5. Methodological considerations
5.1 The psychoanalyst’s estimate of a patient’s disturbance increases over the course of the treatment
5.2 Psychoanalysts tend to overestimate outcome but not necessarily
5.3 The definition of outcomes in terms of individualized treatment goals emphasizes the effectiveness of psychoanalysis
5.4 Simple measures of user satisfaction highlight the ambivalence of patients about analysis
5.5 Measurements show psychoanalysis to produce greater psychiatric symptom changes than personality or relationship changes
5.6 The psychoanalyst’s estimate of therapeutic alliance is more relevant to outcome
5.7 Psychoanalytic technique varies considerably between analysts, even those trained in the same institution
6. Some questions soon to be answered
6.1 How effective is psychoanalysis for major depression and anxiety compared to less intensive treatments?
6.2 Is psychoanalysis more effective than CBT for severely anxious children from a developmental standpoint?
6.3 What are the qualitative differences between nature of change and psychoanalysis and psychotherapy?
6.4 Is psychoanalysis cost-beneficial?
6.5 How effective is modified psychoanalysis for borderline personality disorder?
6.6 What is the value of the contract in the psychoanalytic psychotherapy of borderline patients?
6.7 How important is mentalizing in the psychoanalytic treatment process?

In summarizing these results-as done in these tables-we will adopt a cautiously optimistic attitude towards the evidence presented. In this way we will not be disregarding the weakness of the evidence, but rather we wish to highlight what could be shown by these studies and what the present evidence indicates. Many of the ongoing studies are methodologically “state of the art” and this is of course encouraging as regards persuading skeptics in the field. In general, the findings underscore the effectiveness of psychoanalytic work and should encourage us to undertake further, even more rigorous, explorations of treatment outcome.

In the light of available evidence, the future of psychoanalytic therapies does not look too bad. Nevertheless, psychoanalytic therapies deserve to be studied in much more extensive fashion as they have been in the past.


1) The term is borrowed from the book by T.H. Lawrence The seven pillars of wisdom.

2) Hans Jürgen Eysenck, “The effects of psychotherapy: an evaluation”, J. Consulting Psychology, 1952,16, pp. 319-324.

3) Allan E. Bergin, “The evaluation of therapeutic outcomes” in Allan E. Bergin, Sol L. Garfield (eds), Handbook of Psychotherapy and Behavior Change (New York: Wiley, 1971), pp. 217-270.

4) Bergin, cit., p. 225.

5) Ibid., p. 227.

6) Robert Wallerstein, The Talking Cures. The Psychoanalyses and the Psychotherapies (New Haven, CT: Yale University Press, 1995), p. XV.

7) Peter Fonagy, Horst Kächele, Rainer Krause, Enrico Jones, David Lopez, eds., An open door review of the outcome of psychoanalysis. Research Committee of the International Psychoanalytic Association (London:, 1999).

8) Lester Luborsky, Louis Diguer, Ellen Luborsky, Arthur T. McLellan, George Woody, Louis Alexander, “Psychological health as predictor of the outcomes of psychotherapy”, J con clin psychol, 61, 1993, pp. 542-548.

9) Sigmund Freud (1912), Zur Dynamik der Übertragung, GW, VIII, pp. 363-374.

10) See Louise Gaston, William E. Piper, Edwin G. Debbane, Jean-Paul Bienvenu , John Garant, “Alliance and technique interaction in predicting outcome of short and long term dynamic psychotherapy”, Psychotherapy Research, in press. Timothy Eaton, Nick Abeles, Marvin J. Gutfreund, “Therapeutic alliance and outcome: Impact of treatment length and pretreatment symptomatology”, Psychotherapy, 1988, 25, pp. 536-542. Arlene F. Frank, John G. Gunderson, “The role of the therapeutic alliance in the treatment of schizophrenia”, Archives of Gen Psychiatry, 1990, 47, pp. 228-236.

11) See Louise Gaston, “The concept of the alliance and its role in psychotherapy”, Psychotherapy, 1990, 27, pp. 143-153. Lester Luborsky, “Helping alliance in psychotherapy: the groundwork for a study of their relationship to its outcome” in James L. Claghorn (ed.), Successful Psychotherapy (New York: Brunner, Mazel), pp. 92-116; “A pattern-setting therapeutic alliance study revisited”, Psychotherapy Research, 2000,10, pp.17-29.

12) Lester Luborsky, Paul Crits-Christoph, John Mellon, “The advent of objective measures of the transference concept”, J Consult Clin Psychol, 1986, 54, pp. 39-47.

13) Jim Weiss, Harold Sampson, Group at MZPR, The Psychoanalytic Process: Theory, Clinical Observation, and Empirical Research (New York: Guilford Press, 1986).

14) Hartvig Dahl, “Frames of mind” in Hartvig Dahl, Horst Kächele, Hans Thomä, eds., Psychoanalytic Process Research Strategies (Berlin: Springer, 1988), pp. 51-66.

15) Merton M. Gill, Irwin Z. Hoffman, “A method for studying the analysis of aspects of the patient’s experience in psychoanalysis and psychotherapy”, J Am Psychoanal Assoc, 1982, 30, pp. 137-167

16) Hans H Strupp, Jeff Binder, Psychotherapy in a New Key. A Guide to Time-Limited Dynamic Psychotherapy (New York: Basic Books, 1984).

17) Hartvig Dahl, “Introduction” in Dahl, Kächele, Thomä, eds., cit., pp. VII-XVI.

18) William Henry, Hans H. Strupp, Timothy E. Schacht, Louise Gaston, “Psychodynamic approaches” in Bergin, Garfield, eds., Handbook of Psychotherapy and Behavior Change, cit.

19) Cornelia Albani, Dieter Benninghofen, Gerd Blaser, et al., “On the connection between affective evaluation of recollected relationship experiences and the severity of psychic impairment”, Psychotherapy Research, 1999, 9(4), pp. 452-467.

20) Wilma Bucci, “Converging evidence for emotional structures: Theory and method” in Dahl, Kächele, Thomä, eds., op.cit.; W. Bucci, “Pattern of discourse in good and troubled hours”, J Am Psychoanal Ass, 1997, 45, pp. 155-188.

21) Henry, Strupp, Schacht, Gaston, op.cit.

22) John F. Clarkin, Frank E. Yeomans, Otto Kernberg, Psychotherapy for Borderline Patients (New York: Wiley, 1999).

23) Cf. Otto Kernberg, John F. Clarkin, “Developing a disorder-specific manual: the treatment of borderline character disorder” in N. Miller, L. Luborksy, J.E. Barber, J.P. Docherty, eds., Psychodynamic Treatment Research. A handbook (New York: Basic Books, 1993), pp. 227-244.

24) Thomä & Kächele, cit., p.317.

25) Brin F.S. Grenyer & Lester Luborsky, “Dynamic change in psychotherapy. Mastery of interpersonal conflicts”, J Con Clin Psychol, 1996, 64, pp. 411 -416.

26) Peter Fonagy, “The process of change and the change of processes: what can change in a `good analysis`”,, 1999.

27) Rolf Sandell, Jan Blomberg, Alan Lazar, “When reality doesn’t fit the blueprint: doing research on psychoanalysis and long-term psychotherapy in a public health service program”, Psychotherapy Research, 1997, 7, pp. 333-344.

28) Dorothea Huber, Günther Klug, Michael Rad, “Münchener Psychotherapie Studie” in M. Leuzinger-Bohleber, U Stuhr, eds., Psychoanalysen im Rückblick (Giessen: Psychosozial-Verlag, 1997), pp. 454-469.

29) See Fonagy, Kächele, Krause, Jones, Lopez, eds., op.cit., pp. 237-264.

Share This Article


European Journal of Psychoanalysis