The place of psychoanalytic and psychodynamic treatments within psychiatry is controversial (Auchincloss 2002; Gabbard, Gunderson and Fonagy 2002; Kernberg 2002; Wallerstein 2002). Based on the now aged principles of Freudian psychology, as of today there are several theoretical orientations in understanding human emotion and development (Leichsenring, Rabung and Leibing 2004). However, the intervention generally refers to a specific type of treatment in which the patient is analysed by verbalising thoughts. A variety of techniques may be employed including free association, object relation and dream analysis. The therapists role is to understand the unconscious conflicts causing the patients symptoms by interpretation and allowing the patient to create a resolution (Freud 1903). In understanding their conflicts, transference will occur whereby the patient will substitute his or her relationship with the therapist for the troubling past relationship (often with a mother or father figure) that was at the root of the their problems. After transference has occurred, the patient will be much more inclined to remember the key moments that lead to their neurosis. In theory, this will eventually lead to a long-term recovery (Freud 1903). On paper this sounds perfectly plausible, but the psychodynamic approach has been overshadowed in recent years as behavioural and cognitive therapies have become the treatment of choice (Johnstone 2003). These alternatives are reported to be cheaper and more empirically sound, so does a psychoanalytic approach to mental health still deserve a place within modern psychiatry?
2. Outcome Studies in Psychodynamic Treatment
For long-term psychoanalytic psychotherapy and psychoanalysis, convincing outcome research is still required, but in the field of short-term psychodynamic psychotherapy (STPP) more evidence of its effectivity is available. Various meta-analysis’s have addressed the effectiveness of STPP (Chambless & Hollon 1998; Crits-Christoph 1992). Svartberg and Stiles (1991) found STPP to be superior to a no treatment control condition, but inferior to alternative psychotherapies such as cognitive behavioural therapy (CBT). Additionally, Crits-Christoph (1992) computed larger effect sizes compared with untreated waiting list control patients and STPP was found to be equally effective as other forms of treatment including CBT or psychopharmacological treatment.
While these observations are promising, the selection and quality of studies chosen to be included in a meta-analysis is likely to have a significant impact on the outcome of any comparison. Unlike Svartberg and Stiles, Crits-Christoph included studies of interpersonal therapy (IPT) as representative of STPP, which is surprising given that the relationship between IPT and STPP is a controversial one (Leichsenring et al 2004). Considering the empirical results, IPT is very close in outcome to CBT (Ablon & Jones 2002).
The methodological quality of the studies included in a meta-analysis also plays an important role. For example, many studies included in Svartberg and Stiles paper showed common methodological flaws. Other analysis’s from the likes of Crits-Christoph only included studies that met rigorous inclusion criteria (for example, experienced therapists, same number of sessions). Randomised controlled studies (RCTs) are regarded by many as the gold standard for any demonstration that a treatment actually works. This criteria is the same as that applied to both psychotherapy and evidence-based medicine in general (Leichsenring et al 2004). The assumption is further reflected in the guidelines by the Task Force on Promotion and Dissemination of Psychological Procedures of Division 12 (Clinical Psychology) of the American Psychological Association for the definition of empirically supported treatments. Accordingly, empirical support for psychotherapeutic methods can only be provided by RCTs whereby a therapy group is compared to an appropriate control condition (for example a waiting list or placebo group) or with another well established therapy (Chambless & Hollon 1998). On the other hand, given that psychological and biological models of madness are essentially incompatible, critics have argued that applying such a stringent criteria to psychoanalytic therapy is neither helpful or productive (Johnstone 2000)
The most recent meta-analysis concerning the use of STPP comes from Leichsenring et al (2004), who applied rigorous inclusion criteria and measured the effect of STPP in specific psychiatric disorders. Unfortunately, there have not been enough RCTs conducted to consider disorders separately so their results can only be applied in a more general sense. Their data conforms with earlier meta-analyses from Crits-Christoph (1992), but in this case, studies of IPT were not included. According to the results, STPP yielded significant and large effect sizes, but no difference was found between STPP and CBT concerning changes in target problems, general psychiatric problems or social functioning. In terms of the actual effects on patients, patients treated with STPP are better off with regard to their target problems than 92% of patients before therapy.
Unfortunately, there a lack of research concerning what treatment option might be best for a specific psychiatric complaint. Given that Esther may be suffering from post-traumatic stress disorder, would a psychoanalytic approach be appropriate? PTSD is often chronic, and persists for at least one year after the trauma in approximately 50% of cases (Davidson et al 1996). The most common precipitating events are combat trauma, physical and sexual assault and motor vehicle accidents (Norris 1992).
3. Treating Post Traumatic Stress Disorder: Esther’s Case (1)
Based on Esther’s case notes, she is likely to receive a diagnosis of post-traumatic stress disorder (PTSD). In the last ten years, there have been relatively few enquiries into what conditions would benefit the most from a psychodynamic approach, but a meta-analysis from Van Etten and Taylor (1998) considered the effectiveness of treatments for PTSD. In addition to psychodynamic psychotherapy, treatments considered included CBT, eye-movement desensitisation and relaxation training. In the case of a psychodynamic intervention, the aim is to uncover and resolving the unconscious conflicts arising form the traumatic events. In Ether’s case, this might be sexual abuse (Brom, Kleber & Defares,1989).
Unfortunately, this analysis did not have a particularly strict inclusion criteria, probably because at the time there was a lack of RCTs within many of the treatment groups. They found that behaviour therapy and EMDR (eye-movement desensitisation and reprocessing) were the most effective psychological therapies for PTSD. Effect sizes for these therapies were large relative to control conditions. Drop out rates were also very low in comparison to other interventions. While psychodynamic therapies were slightly more effective in comparison to psychopharmacological approaches, the effect size was considerably smaller in comparison to behaviour therapy.
When it comes to finding the best treatment, it would appear that psychodynamic approaches are often left as a last resort. Recently, Stiles and colleagues (2007) attempted to compare the outcomes of CBT, person centred therapy (PCT) and psychodynamic or psychoanalytic therapy (PDT) as delivered in routine primary-care. These are of course three very distinctive approaches in terms of their usual repertoires of interventions and their assumptions about then nature of psychiatric illness. The author’s approach followed a sensible logic of using clinically representative or effectiveness research. Accordingly, the risk of selection bias associated with the lack of assurance that the treatments were delivered in a standard way are balanced by a greater ecological or realistic validity (Seligman 1995). A previous study by Stiles et al (2006) found that there was little if any difference between CBT, PCT and PDT as delivered within the UK’s National Health Service, but in 2007 they repeated their study with a sample that was four times larger. All the therapies appeared to be effective in the second study, but the mean difference between the targeted approaches did not approach significance. The approximation was even closer than in the smaller 2006 study. These findings extend to those RCTs discussed perviously, but while the overall change scores overlapped considerably, there was a great deal of variation in the actual outcomes within each group.
If Esther’s treatment could be as equally effective by using CBT, PCT or psychoanalytic approaches, does it still make sense to use the old over the new? In the last ten years health economic evaluations have become increasing important in the assessment of treatment interventions (Parry & Richardson 1996; Sharfstein 1997). However, few studies have actually carried out an economic analysis. If Esther could have treatment which was cheaper and just as effective as psychoanalysis, then this would naturally become the preferred first choice option.
Psychotherapy is perceived as an expensive treatment for mental illness, but some evidence has suggested that it may result in cost savings as it can reduce the utilisation of other health care facilities (Guthrie et al 1999). It is possible that because psychodynamic approaches do not have to be focused on specific symptom complexes, this leads to a more secure patient base. Results from Guthrie et al (2009) demonstrated that brief PI patients who are high utilizers of psychiatric services experienced a significant improvement in their psychological status and a substantial reduction in health care utilization and health care costs in the 6 months following treatment. The additional costs of psychotherapy during the intervention period were offset by subsequent reduction after treatment. However, while there was significant improvement, the effect size was relatively small and most patients remained significantly ill. On the other hand, the intervention was extremely brief and a longer course of treatment may have resulted in further improvement. However, if this was to be true then these costs over the long term may not be recouped if treatment had to be continued over a longer timeframe.
Returning to Esther, it is important to remember that she is a child who is (a) not a high service user and (b) not likely to benefit from a short intervention. Cost is unlikely to be an issue when it comes to deciding if and when psychoanalytic therapy is appropriate in this case. Furthermore, all the studies considered so far have dealt with adult populations. There is actually very little evidence examining the effects of psychodynamic therapy on children, possibly because the treatment is often only seen as an option for adults, who suffered as children, rather than for children themselves (Bernsetin, Borchardt and Perwien 1996). At 8 years old, children are unlikely to understand their own feelings, let alone be patient enough to sit down and discuss at length their deep routed subconscious fears to a stranger. This however, would not be the only problem facing Esther is she were to undergo additional treatment.
4. The Problem with Dream Analysis: Esther’s Case (2)
Based on Esther’s case notes, a course of psychodynamic treatment is likely to consist of dream analysis in addition to understanding her anxious mood by using free association. The aim would be to break down her subsequent defence mechanisms leading to some form of transference.
Dream analysis has been discredited within modern day psychiatry (Auchincloss 2002). At the heart of this critical thought is the friction between current knowledge surrounding the purpose of dreams and the assertions regarding there importance made by psychodynamic theory. While Freud argued that the motivation of all dream content is wish-fulfilment, especially within the case of younger children, this is unlikely to be true (Freud 1913). Freud described the actual technique of dream analysis in great detail:
‘You entirely disregard the apparent connections between the elements in the manifest dream and collect the ideas that occur to you in connection with each separate element of the dream by free association according to the psychoanalytic rule of procedure. From this material you arrive at the latent-dream thoughts, just as you arrived at the patients hidden complexes from his associations to his symptoms and memories..The true meaning of the dream, which has now replaced the manifest content, is always clearly intelligible’ (Freud 1909; p435)
However, 100 years on, there are many opposing theories as to why we dream. Some researchers have proposed that dreams have no real purpose, while others believe that dreaming is still essential to our mental and emotional health (Hartmann 1995). Another more widely accepted theory has suggested that the function of a dream is simply to weave new material into memory in a way that can help us cope with future experiences (Hartman 2006).
Because the reasons underlying why we dream within normal behaviour are unclear, psychoanalytic dream analysis does not hold up against rigorous scientific enquiry. In Esther’s case her dreams may be related to her current feelings of anxiety, but many children experience nightmares even over the long term for no apparent reason. Telling a child what their dreams mean at a deeper level may be relatively useless in making them feel more secure about their current situation. Dreams may well still reflect unconscious thought, but this symptom alone is likely to suggest post-traumatic stress disorder, which always produces conscious symptoms. There is also an inherent danger of over analysing these dreams in Esther’s case. Many children have nightmares despite being happy the rest of the time and asking a child to consider these nightmares further may only increase their prevalence.
Based on the studies discussed so far and the initial treatment options available, it might be safe to have some reservations surrounding a modern psychodynamic approach. It is not clear that it would lead Esther to transference and a subsequent resolution.
5. The Problem with Evaluation: Introduction
The many approaches within psychoanalysis vary as much as the theories on which they are based. This causes an additional and as yet unresolved problem when evaluating subsequent treatments (Leichsenring et al 2004). In Eysneck’s view the road to proving the effect of any psychoanalytic therapy has been long and winding:
‘Just as eighteenth century physicists defended the concept of phlogiston to the death, long after Lavoisier had demonstrated that the concept had no scientific value, so the varied theories underlying psychotherapy are defended by practitioners and the media’ (Eyesneck 1994: p477)
While the debate could simply stop here, there are inherent weaknesses that require further investigation. These include additional problems surrounding transference, therapist effects and the lack of any concrete scientific theory. These issues must be explored in detail when considering psychodynamic therapy in a modern, empirically driven context.
6. The Problem with Evaluation (1): Definitions (The Transference Example)
While true Freudian psychoanalysts are becoming rarer with each passing year, one of the concepts he advocated for over a century has persisted: transference. Freud believed that transference was a fundamental part of the psychoanalytic process, sometimes only achieved after many years of careful therapy (Freud 1909). More recently, Susan Anderson and her colleagues have attempted to offer a new explanation for transference. Instead of being a difficult negotiation between patient and therapist, the application of transference is more like a stereotype (Pao 1979; Eyesenck 1994). In their view, transference is experienced all the time by everyone and for it to occur, all that is required is to encounter a person that reminds you of another, significant person in your life. Transference in this view is simply the misapplication of the traits of your significant other to this new unknown person (Kruglanski & Pierro 2008).
In a typical study, Anderson’s team had a participant describe an important person in their life (Katrina & Andersen 1996). Then after a delay, they read a description of a new person that has some similarities with their friend or loved one. In a subsequent memory test, subjects often misapply traits of their loved one to the new person. These traits were not present in the description they read. However, Anderson’s transference may not be considered the same as what Freud describes. One difference is that Andersen’s transference appears to arise during moments of cognitive confusion, whereas Freud said transference was the product of intense cognitive effort (Freud). Recently, Arie and Kruglanski and Antonio Perro (2008) have devised an additional experiment to consider which explanation has more support.
They asked 42 participants to indicate whether they were morning people or evening people – whether they functioned better in the morning or night. Additionally, they had to describe a significant other as in pervious studies. Two weeks later, they invited the subjects back for a supposedly unrelated treatment. Half the morning people were asked to either show up at 7am or 8pm and the evening people were divided in the same way. Participants had to read descriptions of ‘strangers’ that either incorporated eight of the items from their own lists of traits from their loved ones, or instead included random items from another participant’s list.
After a short break they were tested with 15 items, some of which has been used to describe the stranger, some of which were traits they themselves had used to describe their loved ones (but were not included in the description of the stranger), and some of which they had said were irrelevant to their loved ones. When people were at their cognitive best (for example, morning people tested in the morning) there was no difference in the number of false alarms however, when morning people were tested in the evening, or vice versa, a different pattern was observed. When the stranger was similar to their loved one, these people had significant false alarms, misattributing more of their loved one’s traits to the stranger.
In summary, transference was only found when people were cognitively weak and the effect of transference disappeared when people were at their cognitive peak. The authours’ argue that this effect casts doubt on the entire Freudian notion of transference. Transference may not occur as a result of cognitive effort, but is alternatively a product of cognitive confusion. Having already shown that dream analysis has been discredited within modern day psychology, it would appear that the theory surrounding the most crucial aspect within psychodynamic theory is also highly questionable.
7. The Problem with Evaluation (2): Therapist Effects
As far back as 1936, Rosenzweig proposed that common factors were responsible for the efficacy of psychotherapy and used the conclusion of the Dodo bird from Alice in Wonderland, where the Dodo bird proclaims ‘everybody has won and all must have prizes’. To examine if this is still present in modern day psychodynamic treatments and interventions, Wampold and colleagues (1997) conducted a meta-analysis including behavioural and psychoanalytic treatments. Their meta-analysis differed in that the the corpus of studies was limited to only those that compared two or more treatments, which were not classified into general types. The results demonstrated that the distribution of effect sizes produced by comparing two bonafide psychotherapeutic treatments was consistent with the hypothesis that the true difference is zero. Surprisingly, the effect sizes produced by such comparisons were not related to the similarity of the treatments compared, nor did they increase as a function of time. Lastly, this study examined direct comparisons between therapies so that the results are not confounded between differences in outcome measures and all the findings are consistent with the Dodo bird conjecture.
However, what is more surprising is that when breaking down the reasons that define a patients recovery, the actual treatment used only accounts for 1% of the variance. 74% comes from unexplained stressors such as social support, the patients health condition or other supports. The clinician accounted for 6% of the outcome, so regardless of whether a cognitive, behavioural or psychodynamic method is used, a good therapist will make all the difference. Furthermore, a bad clinician can dramatically reduce the effect of a good ‘treatment’ whereas a good clinician can being a poor ‘treatment’ up to an effectiveness level to equal that of the best treatment (Eyesneck 1994; Wampold, Mondin, Moody, Stich, Benson and Ahn 1997).
This brings into question current methods of training given to those who wish to practice psychoanalytic or psychodynamic therapy. In a traditional Freudian sense, to become a psychoanalysts involved being psychoanalysed first, taking many years. However, as recent research from Ward and Bucher (2003) has shown, fast track training may not improve delivery or the outcome for patients. Mental health nurses are now becoming increasingly involved in the delivery of psychological based interventions. Ward and Bucher gave training to those who volunteered over several weeks. The hypothesis for the study was that receiving training would result in significant changes to the nurses’ behaviours during therapeutic interaction. Three month follow ups showed only one significant behaviour change out of the 12 expected. Experienced counsellors however, who also completed training did adhere to the model. At the time of writing it is unclear whether these psychodynamic training schemes are having any impact on patients themselves. On the other hand, evidence from mental health care service user groups suggests that people with mental health problems strongly favour a more psychodynamic approach because they feel more informed and in control (Paley, Myers, Patrick, Reid & Shapiro 2003).
8. The Problem with Evaluation (3): Theory within Psychodynamic Treatment
Unfortunately, the empirical validation strategy weakens support for psychotherapy as a mental health treatment rather than making it stronger. Klein (1996), an advocate of psychopharmacological treatments, summed up the issue:
‘The bottom line is that if the Food and Drug Administration (FDA) was responsible for the evaluation of psychotherapy, then no current therapy would be approvable’
(Klein 1996; p84)
Why is it then, that researchers persist in attempts to find treatment differences when they know that these effects are small in comparison to therapist effects or effects of treatment versus no-treatment comparisons (Wampold et al 1997) ?
The view of many academics is that psychotherapy is essentially a technology without any scientific basis. Critics continue to argue that our main concern should be with the creation and working of a scientifically valid theory of why these therapies actually help people (Eysenck 1994). Freud’s own theories have consistently been shown to as scientifically invalid and without any valid theory, the alleged facts are impossible to interpret (Grawe 1993). Does psychodynamic therapy work cannot really be answered in any form without a sound theory. The Freudian theory rejects concern with symptom removal or alleviation of symptoms; symptoms, in psychoanalytical theory, are like fever, indicative of an underlying illness and it is this illness that has to be cured to lead to a symptom-reduction, not a permanent recovery. On the other hand, behaviour therapy is based on the theory that symptoms are directly produced by conditioning processes, and that any intervention is aimed at removing the conditioned response will permanently remove the symptom. What is to be criterion of success in psychodynamic therapy, the removal of all symptoms or a general personality change as demanded by Freud? (Freud 1903).
Another important variable is the duration of treatment. Freud predicted that the acquisition and fixed resolution of transference effects is essential to a cure and can take a long time. As discussed perviously, the term transference is theory laden and and should never be used until the theory on which it is based can be made clear (Eysenck 1994). However, the widespread use of short-term psychotherapy would suggest that Freudians have acknowledged that Freudian theory is completely wrong on this point.
Finally, when it comes to assessing the outcome of a therapy, there are a bewildering number of different measures to choose from. There are projective measures, introspective measures from the patient and ratings by depended judges or members of close family. A sound theory would govern the choice of measure and some instruments would be more acceptable than others. Meta-analyses adding together results obtained through the use of all these different measures simply confound incommensurate data. While the initial outcome studies discussed appear to show a positive effect, it is still, after decades of research, unclear what constitutes a positive change in behaviour.
9. Conclusions – What Next for Esther?
It is difficult to come to any solid conclusions surrounding the use and effectiveness of psychoanalytic treatment. It is probably not suitable for those who are of low intelligence or children because it requires the patient to be introspective, something that Esther is likely to lack at 8 years old. Ethically, it would be unwise to prescribe a course of treatment for a child so young when there is little evidence to suggest it would ever work and if it did, this would have more to do with the child than the choice of therapy. There is always the danger of over analysing her problems and causing additional damage that may effect her well into adulthood. On the other hand, a brief psychodynamic-interpersonal approach may well prevent later episodes of psychosis. It is guesswork and the correct therapist with the right experience is going to be more important that the chosen intervention. As an alternative, Esther might benefit from a more child friendly approach such as Art Therapy (Malchiodi 2003). Unfortunately, this too suffers from the same problems that underlie most other psychodynamic approaches; theory and sound clinical evidence.