The heart of the matter: continuing the conversation

I would like to begin my contribution to the intercontinental conversation invited by Ivan Eisler in his recent editorial (Eisler, 2006) by thanking him for the considerable effort he obviously devoted to making a careful and thoughtful reading of my article in Family Process (Simon, 2006) and for his willingness to devote precious space in the Journal of Family Therapy to furthering the crucial dialogue, relevant to family therapy practice on whatever continent it occurs, regarding precisely what it is that makes family therapy work. As Eisler (2006) so ably summarized, it is currently a matter of some contention in North America whether it is common factors or model-specific factors that are primarily responsible for achieving positive outcome in family therapy. No participant in this debate of whom I am aware advocates a simplistic, either/or position on this question. Model-specific factors proponents recognize that common factors undoubtedly play a role in therapeutic outcome (e.g. Sexton et al., 2004), just as common factors proponents recognize that 'common factors are not islands – they work through models' (Sprenkle and Blow, 2004, p. 151). None the less, the parties in the debate remain divided by 'their different takes on the foreground/background gestalt in the picture of therapeutic efficacy. In the common-factors perspective, models are in the background, common-factors in the foreground. In the model-specific-factors perspective, the situation is reversed' (Simon, 2006, p. 336). This debate is not one of those abstract discussions of interest only to academics; rather, it entails considerable implications for the way in which family therapists are trained and supervised. For, if it turns out that common factors account for the bulk of the variance in therapy outcome, then the current pedagogical focus – at least in the United States – on having trainees learn treatment models will need to be replaced with a focus on helping trainees enact the generic behaviours and acquire the generic attitudes that common factors proponents maintain constitute the therapist's main contribution to achieving positive therapeutic outcome. If, on the other hand, it turns out that model-specific factors play a determinative role in therapeutic outcome, then the effort already devoted to having trainees learn, compare and contrast models of treatment will need to be maintained. Both the common factors position and the model-specific factors position in this debate have intuitive appeal, and both positions can be supported by some research findings. Therefore, I deemed it worthwhile to try to construct a hypothesis about the contributors to therapeutic outcome that would preserve and integrate the essential insight of both positions. Towards this end, I offered in my article the hypothesis that 'therapists achieve maximum effectiveness by committing themselves to a family therapy model of proven efficacy whose underlying worldview closely matches their own personal worldview' (Simon, 2006, p. 331; emphasis in original). The hypothesis preserves the key intuition of the common-factors perspective: that in the final analysis, what makes family therapy work is something that is common to all the major therapeutic models – namely, a certain quality of presence on the part of the therapist. However, the hypothesis also preserves the guiding intuition of the model-specific factors perspective, when it asserts that this therapist-related common factor is entirely dependent for its realization on the therapist committing to and adhering to a model of proven effectiveness. Eisler (2006) believes that he detects circularity in my hypothesis when I assert that, in order to achieve maximal personal effectiveness, a therapist needs to practise a model ofproven efficacy whose underlying worldview closely matches her own. 'This leads to a conundrum. How do we establish the efficacy of treatments in the first place if congruency is such a central ingredient' (Eisler, 2006, p. 331)? Any suspicion of circularity in the hypothesis disappears when one realizes – as Eisler (2006) himself suggests that he does – that therapist-worldview/model-worldview congruence plays a different kind of role in therapeutic outcome than do model-related factors and the usually identified therapist-related common factors (e.g. empathy, acceptance, respect, support). In a research study designed to test my hypothesis, model-related factors and therapist-related common factors would be the independent variables. Therapist-worldview/model-worldview congruence would, in distinction, be a moderating variable whose presence would be hypothesized to increase the impact of the independent variables, resulting in enhanced positive therapeutic outcome (the dependent variable). The hypothesis assumes that therapist-worldview/model-worldview congruence produces a synergistic effect between those elements in the model that contribute to therapeutic efficacy and the therapist factors that contribute to efficacy …. [It] assumes that when therapist-worldview/model-worldview congruence turns a model into an instrument for deep and authentic self-expression on the part of the therapist, that which is therapeutic in the model and that which is therapeutic in the therapist are mutually activated and enhanced, with the result that the therapist becomes maximally effective. Because model-related factors and worldview congruence are variables of a different kind, it is not at all self-contradictory for my hypothesis to assume that a therapeutic model makes a contribution to therapeutic efficacy that is independent of whatever worldview congruence might or might not exist between the model and its practitioner. It must be emphasized that the hypothesis does not assert that worldview congruence is sufficient to achieve positive therapeutic outcome. A therapist whose practice focuses on the reading of entrails might have a personal worldview that is highly congruent with the worldview underlying this 'model'. Despite this congruence, however, I would venture to guess that this therapist would achieve a rather meagre rate of therapeutic success, precisely because the 'model' he is practising lacks therapeutic efficacy. Model-worldview/therapist-worldview congruence cannot overcome the deficits entailed by a therapist who lacks generic therapeutic attitudes and skills, and/or by a model that is lacking in efficacy. Thus it entailed no contradiction for me to assert that a therapist maximizes his personal effectiveness when he practises a model of proven efficacy whose underlying worldview matches his own. Because my hypothesis conceives of models making an independent contribution to therapeutic outcome, there is nothing in the hypothesis that militates against the ongoing use of randomized clinical trials to determine which models of family therapy work. Such research isolates the contribution made by models to therapeutic outcome by controlling for or ignoring 'extraneous', non-model-related variables, such as therapist-related common factors and model-worldview/therapist-worldview congruence. However, I am entirely in agreement with Eisler (2006) when he asserts that family therapy research that is limited to randomized trials is intrinsically of limited usefulness. He is correct when he observes that while randomized trials tell us what works in family therapy, they provide little insight into how it works. I believe that this how might begin to emerge into view when we begin to conduct research which examines the role played by model-worldview/therapist-worldview congruence in therapeutic outcome. Well-constructed research focused on this question could begin to reveal the interactions between various factors that lead to positive therapeutic outcome. When I was writing my article, I was focused almost entirely on the interaction between two factors, namely the therapist and the treatment model she is employing. Eisler (2006) widened my focus when he noted in his editorial that therapist-worldview/model-worldview congruence might itself interact with and enhance client-related common factors, such as expectancy and hope. Research focused on worldview congruence, then, might reveal that this variable is the nexus where paths which originate in model, therapist and client system meet, merge and then proceed on to the destination of a successful therapeutic outcome. Of course, it is also possible that this research might reveal that worldview congruence plays no such role. However, because this research holds promise of moving the field of family therapy beyond its current focus on what works in therapy to a focus on how therapy works, I would venture to assert that conducting this research is worthwhile, regardless of the conclusion that is eventually reached about the salience, or lack thereof, of therapist-worldview/model-worldview congruence.