International collaboration set to facilitate the measurement of results in psychotherapy

This topic was discussed by specialists from several countries at the São Paulo School of Advanced Science on Mental Health and Evidence-Based Interventions, an event held with the support of FAPESP.

By Maria Fernanda Ziegler  |  Agência FAPESP – Mental health care is a key twenty-first century concern in light of the growing numbers of people who suffer from anxiety and depression. According to the World Health Organization (WHO), depression affects 322 million people worldwide. In Brazil, it affects 11.5 million (5.8% of the population), while anxiety-related disorders affect over 11.8 million (9.3%). The establishment of structures and public investment in psychotherapeutic care are, therefore, of paramount importance.

Research groups in clinical psychotherapy can achieve significant advances in such scenario by using assessment tools based on common results. The generation of data using similar methodologies would enable scientists in the field to deploy complex statistical techniques and advanced technology such as machine learning in search of repetitive patterns on which to base predictions, for example.

These and other issues were discussed by specialists in different psychotherapeutic approaches, including behavioral, cognitive behavioral and humanistic therapy, during the São Paulo School of Advanced Science on Mental Health and Evidence-Based Interventions held in fourth-quarter 2019 in Amparo, São Paulo State, Brazil.

The event was organized by the Center for Education and Human Sciences at the Federal University of São Carlos (CECH-UFSCar) in partnership with the Paradigma Center for Behavioral Science and Technology and was funded by FAPESP’s São Paulo School of Advanced Science program (SPSAS). It was attended by some 80 researchers, half from Brazil and the rest from Colombia, Ireland, Spain and the United States.

According to Maria de Jesus Dutra dos Reis, who coordinated the event, the scientific community that produces psychotherapeutic data should work together to advance international research collaboration.

However, in her view, the various approaches to psychotherapy have yet to reach a consensus on this matter. “Psychology is not paradigmatic. As a result, therapists who use different methodologies tend to measure the outcomes of their interventions in different ways. Each ‘school’ has its own way of systematizing intervention processes, so that it is often hard to quantify the evidence of effectiveness in a unified or systematic manner. In addition, the concepts and metrics used in psychotherapeutic treatment are often moot, especially in areas such as subjectivity, affects and emotions,” Reis said.

Collaborations

One of the aims of the organizers of the event was to foster scientific study based on tools for measuring results so as to ensure that research on the effectiveness of psychotherapy is reproducible. This requires collaboration among research centers in different countries or regions.

“The guidelines for empirically based therapy require different randomized controlled clinical trials conducted in different laboratories. After all, a single psychotherapist or laboratory is very unlikely to have a large enough number of patients to have sufficient data to constitute an empirical basis,” said Francisco Ruiz, a researcher in the Psychology Department of Konrad Lorenz University, Colombia (South America), and a speaker at the event.

Ruiz’s research in the field of clinical psychology includes collaborations with Spanish universities. “Collaboration tends to improve the quality of research and the impact of scientific articles. Moreover, when several different labs are involved in research in clinical psychology, we can take into account novel factors such as the culture of a specific country, which is still very hard to study scientifically,” he said.

Entropy

According to Reis, a parallel can be made between chaos theory and the variables of the therapeutic process relating to health workers, patients, social systems, the economy, the family and so on.

“It’s possible to identify similarities between the highly complex system involved in clinical psychology and disordered particles [high entropy] in chaos theory,” Reis said. “The process is sophisticated, but we shouldn’t take it for granted that the complexity of the investigation involved in the therapeutic process can’t have a common measure. The appropriate perspective on clinical psychology is that some order or rule emerges from the system and that a specific model to describe this apparent chaos can be found.”

Chaos theory is used in mathematics, physics and biology, among other fields. It posits that particles are not organized in an entropic environment. Nevertheless, chaotic systems have rules that can be well understood. “We’ll only have strong models in psychology if we have a large amount of data and can analyze it in the same way as entropy is analyzed in chaos theory,” Reis said.

“A chaotic environment isn’t lacking in order or an explanatory model. It’s an active and unstable system. Hence the need to create a large amount of data so that all subsystems can be identified. That’s what we’re doing here at this school. We’re talking about tools that can be used to analyze the data, particularly in discourse analysis, so we can understand how and why these ‘particles’ are chaotic and how they interact.”

Unified measuring tools and integrated cross-cultural analysis of therapeutic processes would also help to give research projects a more comprehensive scope and might even contribute over time to public policy formulation.

“Psychology is a science that needs to construct a theory for all possible variations in individuals’ behavior, cognition and emotion. It isn’t enough to describe the intervention procedures considered efficient only for most people [the normative average]. Like all human sciences, psychology should also describe and predict for those who are outside this normative majority,” Reis said.

While being as comprehensive as possible is important, it is equally crucial to establish which intervention procedures actually benefit most individuals. “At a time of scant funding and severe human suffering, it’s crucial to find evidence that can guide the different categories of health professionals and, as a resulting process, contribute to the construction of effective mental health policies,” Reis said.

Michel Lambert, a researcher at Brigham Young University in the US who has developed a method to track the outcomes of psychotherapeutic treatments that are used by clinics, health management organizations and governments in several countries, argued that treatment failures suggest the need to measure and monitor patient responses throughout the process.

“Psychotherapy unquestionably has a positive impact on psychological disorders,” he said. “However, clinical trials have shown that there are flaws in patient responses and that as many as 65% of people in routine care leave the treatment without measured benefits. In the case of research, studies should be replicated in order for progress to be achieved in this field.”

Computer science

The specialists who attended the event highlighted the benefit of using computer science and machine learning to analyze the many variables involved in psychotherapeutic treatment.

“Psychotherapy is highly complex. A large number of variables interact in the process, and these tools are productive when you study complex systems. In research on fear, for example, we take into account such factors as positive reinforcement, punishment and stimulus control. However, all these factors are analyzed separately. We can’t try to build a psychotherapy package like a pill. Each decision involves a different reaction. That’s not how it works in practice, even in the consulting room,” said Felipe Corchs, a professor at the Institute of Psychiatry at the general hospital (Hospital das Clínicas) run by the University of São Paulo’s Medical School (HCFM-USP).

According to Corchs, studies in psychiatry and neuroscience can be seen to have moved closer to computer science, although this trend is less decisive in clinical psychotherapy, probably because of the highly subjective nature of this activity.

“Psychotherapy involves many occurrences that aren’t easy to measure objectively because they happen in people’s private lives. Objective measurement of the level of oxygenation in a specific brain region or the effect of a drug is possible in neuroscience or psychiatry, but in psychotherapy, you can’t measure exactly what the patient is feeling. The patient may tell you, but that account can be biased for many reasons,” Corchs said.

The subjectivity of psychotherapy cannot be eliminated by the new technologies, of course, he added, but it may be possible to use them to access useful information that can move research forward.

“Scientific research will never be able to touch on an important part of psychotherapy that’s subjective,” he said. “However, I see no alternative to formulating computational, mathematical and predictive systems. This approach is the closest we can get to all the variables involved in the complexity of psychotherapy.”