Almost 70 years ago, Eysenck1 stirred up the community of psychotherapists by postulating that psychotherapies—at that time predominantly psychoanalytic—are not effective in the treatment of psychological disorders. This led to a massive surge of empirically evaluated psychotherapy research and promoted particularly the rise of cognitive behavioural therapy. Today, we know that a range of psychotherapies work across a wide variety of mental disorders and numerous meta-analyses of randomised controlled trials prove that Eysenck’s conclusion is no longer relevant. However, despite ample evidence that psychotherapy is generally efficacious, only 30% of patients achieve remission while as many as 65% leave treatment without a measurable benefit or even with deterioration.2 Therefore, psychotherapy researchers face the challenge to improve the effectiveness of their interventions. In order to solve Gordon Paul’s3 fundamental question—‘What treatment, by whom, is most effective for this individual with that specific problem, and under which set of circumstances?’—we have to ask: What is hindering the development of the field of psychotherapy and how can it move forward?

Until today, categorical thinking still informs treatment selection and led to the development of intervention guilds and psychotherapy schools, which has retarded our progress in understanding and treating mental disorders. Mostly in absence of any empirical evidence, psychotherapy schools are usually based on plausible, yet unproven theories and on commercial and status interests of the representatives. Moreover, strong identification with one’s own school and its superiority over other schools reflects drastic allegiance effects and high risks of bias in research. As Marvin Goldfried,4 one of the pioneers of psychotherapy research, prominently calls out, the lack of consensus and disparate languages across theoretical orientations means that identifying the core factors that may underlie the effectiveness of psychotherapy is difficult if not impossible and holds back progress in the science and practice of psychotherapy.

In more recent times, a trend is emerging to move away from nosology and a strictly categorical diagnostic approach to dimensional, function-oriented, mechanistic constructs used as specific therapy targets. Abandoning the dichotomies, categorical approaches and guilds as well as overcoming mere ‘horse races’ in efficacy research may help us to understand mechanisms and to move towards a contextual model of psychotherapy. This coincides with an increasing interest in medicine and psychology to develop individualised precision therapy. By identifying the key elements that may be driving an intervention’s effect, transdiagnostic-modularised approaches can be developed addressing pathological mechanisms such as difficulties in emotion regulation or social threat hyperresponsiveness5 according to an evidence-based heurism across comorbidities and heterogeneous symptoms. A transdiagnostic-modularised approach could also help bridge the practice-research divide, facilitate digital mental health approaches6 and thus address the vast mental health treatment gap globally and locally7 8 by training psychotherapists and mental health trainees in the application of evidence-based core components rather than in theoretical orientations9 10. To speak with Gaines, Goldfried and Constantino11: ‘A consensually effective therapist would be one who, when faced with a specific clinical scenario, could astutely select and deploy the optimal evidence-based strategy at the appropriate time regardless of the main therapeutic orientation’.