Therapeutic change processes link and clarify targets and outcomes

The dominance of the latent disease model of the DSM and ICD has led to an over‐emphasis on symptom reduction as the primary target and outcome of psychotherapeutic interventions, as Cuijpers1 points out.

Clients, employers, funders and the public at large did not demand the narrowing of focus that has accompanied psychiatric nosology. As Cuijpers1 correctly notes, there are other targets and outcomes that might be far more important, such as improvement in quality of life or life functioning, or economic outcomes. To those we might add prosocial and physical health variables, such as reductions in interpersonal violence or lifestyle related physical diseases.

Cuijpers1 concludes that the greatest weight should be given to patients when determining the priorities for the targets and outcomes of psychotherapies. We agree. But, if we are to consider a broader range of intervention outcomes, it will be all the more important to clarify how to move empirically from individual characteristics to individual goals by learning more about the “set of theory‐based, dynamic, progressive, and multilevel changes that occur in predictable empirically established sequences oriented toward the desirable outcomes”2. In other words, we will need to understand therapeutic change processes and link them to effective intervention kernels.

The core question in modern intervention science is “What core biopsychosocial processes should be targeted with this client given this goal in this situation, and how can they most efficiently and effectively be changed?”2. In that context, we are concerned with Cuijpers' dismissal of processes of change and other theory‐driven “intermediate outcomes” . Without a process focus, broadening our outcome perspective could result in even more technological proliferation and confusion than we have now.

Based on studies of mediation, Cuijpers concludes that “there is no evidence” that it is helpful to target processes of change. We disagree. Mediation is only one approach, and the traditional approach to studying mediation is flawed in many ways. Processes of change are idiographic by their nature3, and thus the statistical assumptions built into classical mediational analysis are universally violated.

Classical mediation focuses on a few processes, assumed to be related to outcomes linearly, unchanging across time, without any feedback loops or recursive processes. Such highly implausible assumptions form the basis of demands to prove that there have been no violations of temporality between mediators and outcomes, to show a dose‐response effect, or to prove that no third variable can be involved. In some areas (e.g., third variables) there is no agreed upon way to meet these requirements, and in others (e.g., temporality) little can be recommended beyond guesswork.

Nevertheless, it is supposedly scientifically conservative to prohibit publication of mediational results unless these methodological requirements are met. The result is a domain of ignorance at the core of psychotherapy research that has been to some degree artificially produced. Psychotherapy is rarely – if ever – a paucivariate, linear, continuous and unidirectional event. Instead, psychotherapy typically changes many interconnected variables that form a dynamic system in a non‐linear, bidirectional, dynamic and complex manner. This is best studied by adopting a dynamic systems and complex network approach4. Linear regression models of a few variables are simply inadequate.

Similarly, processes of change supposedly need to be treatment program specific. This idea emerges from a protocol focus – defending that a method engages unique processes of change – but it takes on a different hue when treatment is process‐based5, 6. If processes of change are central, why is it lethal if various technologies alter them? Treatment generality might in principle make change processes more important, not less.

Processes of change ultimately must be theory based and testable, but techniques under various banners and brand names may alter overlapping and broadly applicable processes of change. From the practitioners' point of view, so much the better. That fact empowers practitioners to broaden the range of methods they use in order to target an important change process.

Longitudinal evidence, basic research evidence, and component study evidence suggest that some processes of change are more important than others. For example, it would be strange if processes of change had no linkage to variation, selection, retention, and context sensitivity processes that are to be key to the evolution of complex systems in every other area of life3.

Indeed, it is worth noting that some of the patient‐supplied outcomes described by Cuijpers1 – such as interpersonal effectiveness, social support, the capacity for problem solving, accepting and valuing oneself, awareness, or self‐understanding – have been examined in other contexts under the rubric of processes of change. This suggests that patients themselves intuitively care about processes of change even when traditional intervention science has not focused effectively on them.

Departing from a nomothetic latent disease model and embracing the idiographic complexity of human suffering could free the field to pursue a more process‐based approach. Focusing on therapeutic change processes should not be a side note but should take center‐stage if we want clinical science to move forward.