process, needs to be taken into account. On the other hand, in cases stagnating in alliance over the 4 sessions of BPI, therapist competence is also important but in the opposite sense: low to moderate competence yields the best outcome and the more competent the therapist, the less positive the outcome. It could be said that, in the latter cases, the therapist does “more of the same” by delivering competent interventions, which finally have only a limited impact on symptom change. Two reasons might be at stake: (1) in these patients, the impact of the intervention is confined to an internal psychodynamic level, with no direct impact on our outcome measure (2) The patient–due to his dysfunctional relationship patterns- establishes a rigid level of alliance and is thus highly resistant to the therapeutic relationship and the therapist’s interpretations, even more so if they are competently delivered. These results complete Barber et al’s (2006) study on the moderator effect of adherence. For competence, as we defined it, a linear moderator model might be most accurate, compared with adherence, where a curvilinear yields similar effects of alliance. Using these approaches, we conclude that the highest competencescore is the optimal within the differential context of growing alliance, whereas for adherence, the median adherence is in any case the optimal (Barber et al., 2006). Because context-embeddedness of the technique (e.g., “skillfulness” and “providing a therapeutic milieu”) is the main difference between Barber et al’s (2006) definition of adherence and ours of competence, we hypothesize that more outcome variance is explained with the wide concept of competence, compared with adherence, when taking into account the context of the applied psychodynamic technique. If this assumption holds true, it might also account for the absence of effect of the competence measure by Barber et al. (2006) who defined competence less broadly than it was done in our study (see Introduction section). On the other hand, it might be argued that our definition of competence is so broad that risks of confounds with other context variables, such as the therapeutic alliance, are not excluded. Empirically, such a critic does not stand further examination, as the differential effect of stable alliance demonstrates: even if the therapist’s competence varies in these dyads, alliance remains the same, indicating at least some independence between these variables. Several limitations of this study should be underlined. This is a naturalistic study; although the distribution of the patients between the therapists was controlled for, the patients were not randomly assigned to the therapists. Consequently, there was no controlled distribution of patients to therapists according to their number of years of training and experience. Such a control would have enabled us to partial out the influence of therapists’ training and level of experience. In this study, it confounds with competence due to high correlations. G$#.1/.\!
Jean-Nicolas Despland, Prof., Yves de Roten, PhD, Martin Drapeau, Prof., Thierry Currat, MD,* Veronique Beretta, MPs, and Ueli Kramer, PhD
CLP
.01230/1.40/5&&'67894/0/571.8/5&&/6648./1.40&
Made with FlippingBook HTML5