The publication of the report by experts from Inserm (France) about the efficacy of psychotherapies (2004) has raised many criticisms, mainly from psychoanalysts. The criticisms of Perron, Brusset, Baruch and Emmanuelli, professors of psychology, published on Internet (www.techniques-psychotherapiques.org) in 2004, attack the methodology that they say introduced numerous biases in favour of cognitive behaviour therapies (CBT) compared to psychodynamic psychotherapies (PP). In order to argue against these criticisms, I have selected ten points raised by this group that seemed to represent the essence of their criticisms. Perron et al. say that the Inserm report is neither objective nor scientific; although this report cannot be fully objective due to its narrative nature, it is, nevertheless, based on empirical studies that satisfy scientific criteria for quality and that are well documented. * Perron et al. suggest that the Inserm report contains numerous biases; it is not possible to avoid all statistical biases so this criticism is an attack on accepted scientific methodology. * Perron et al. do not provide any evidence in favour of their criticism that the meta-analyses included in the report have accumulated biases due to prejudices common among psychotherapy researchers; even if researchers were all subject to therapy allegiance, this would not lead, systematically, to biases in favour of CBT. * Perron et al. state that meta-analyses automatically reinforce biases from randomised controlled trials (RCTs), whereas I argue that biases in RCTs are not systematically accumulated, and that the sensitivity of meta-analyses to doubtful RCTs can be tested by comparing results from analyses with and without these studies. * They also say that, in RCTs, the comparability between groups cannot be complete, however, the aim is not identity, and the comparability can be empirically tested after randomisation. * They also declare that the use of DSM automatically favours CBT, however, despite the fact that DSM is unsatisfactory even for biological psychiatrists and CBT therapists; no unanimously accepted alternative system has yet been identified. * Perron et al. say that the use of quantitative assessment of the psychotherapy outcome favours CBT because of an epistemological split between descriptive-nosological and functional-psychodynamic approaches. However such approaches are not clearly exclusive from each other, and a reduction of psychiatry into only two approaches does not fit with the observed variety of etiological hypotheses. * The statement that the functional-psychodynamic approach is contradictory with quantitative assessment of outcome is challenged by the fact that numerous outcome instruments for psychodynamic assessment exist. Furthermore, I have examined the size of the differences between CBT and PP, by selecting studies and meta-analyses that report the direct comparison of these therapies, but this does not provide evidence of major differences. Therefore, the existence of major bias in favour of CBT, due to either the use of DSM, symptom rating scales or other causes, is not supported by these facts. * The accusation that studies included in the Inserm report did not assess patients from control groups at endpoint has being found to be untrue. * The argument that the various effect size statistics used in the meta-analyses give rise to major bias is very unlikely. However, the authors overlooked the more important problem of pooling different outcome measures. * The allegation that a bias in favour of CBT was induced because there were more studies conducted with CBT than with PP is false, because the number of studies does not influence the effect-sizes. The Inserm report does, however, because of the presentation of results by DSM disorder and emphasis on absolute efficacy studies, find more positive results for CBT than for PP compared with no treatment. In conclusion, Perron's et al. criticisms are, in fact, against the application of scientific methodology to the assessment of the efficacy of psychotherapies as a whole. Nevertheless, the Inserm report has some weaknesses; the low number of studies of PP, the lack of direct comparisons between CBT and PP, the fact that multiple efficacy criteria were not taken into consideration, and the problem of comorbidities which was insufficiently taken into consideration.
Read full abstract