“As a clinician I try to follow outcome research in psychotherapy, and it seems like the whole biomedical field is moving away from randomized trials to research on who benefits from treatment. This whole move towards personalization does not seem to be happening in mental health research. Why is that?”
An old question we still cannot answer
It is not new to stress the importance of examining for whom and how therapies work, in addition to examining if treatments work. In a seminal paper from 1967, Gordon Paul already said that outcome research should focus on the question:
“What treatment, by whom, is most effective for this individual with that specific problem, and under which set of circumstances?”
Since then, many scientists and clinicians have repeated the importance of examining who benefits from which treatment. But unfortunately, little progress has been made since this question was posed about 60 years ago.
There are several thousand randomized trials showing that many types of psychotherapy are effective for many different mental health problems. But we have made hardly any progress on answering who benefits from which treatment.
For example, in depression we have many different effective therapies, but it is not known which therapy is best for which patient. If you must choose a therapy for a specific patient, there is still no evidence-based recommendation that you can use. The whole biomedical field is moving towards “personalized treatments”, but little progress has been made in the field of depression and other mental disorders.
The search for subtypes of mental disorders
Most research has been done in the field of depression, because depression is such a heterogeneous condition. There have been many attempts to find different ‘types’ of depression that would respond better to specific types of therapy, including for example melancholic depression, atypical depression, anxious depression, early onset depression, early trauma depression, etc.
Unfortunately, none of these subtypes has been found to respond better to one type of treatment compared to another. Exceptions are bipolar and psychotic depression that do need specific treatments, and, for example for seasonal depression, specific treatments have been developed. But other than that, the attempts to differentiate different types of depression have not been successful. Treatments work on average, but knowledge on who benefits from which treatment is almost completely absent.
Three approaches to the problem
One major problem of examining who benefits from which treatment is that large studies with many participants are needed. Most effect studies are designed to show that a treatment works, not for whom, resulting in far too few participants to examine that.
One innovative approach to find more knowledge on this question is through individual patient data meta-analyses. In these studies, the data of multiple primary studies examining the same intervention are combined into a very large new dataset, which solves the problem of small numbers of participants.
A growing number of such meta-analyses are being conducted, especially in the field of depression, but only a few have found characteristics that predict who will benefit more from one therapy compared to another. None of the findings are ready to be used in clinical practice.
Another major new method to identify who benefits more from one treatment compared to the other is the use of machine learning methods in large samples of patient data in routine care. A growing number of such studies has been conducted in recent years, but again no results have been found yet that are ready to be used in clinical practice. This relates to the problem that the machine learning methods are usually too unreliable, impractical, or both, to inform practice.
A third method that can be used to predict outcome in patients is the use of very large randomized controlled trials with enough participants, and the inclusion of many characteristics that may predict outcome. A growing number of such studies is now being done, but unfortunately no characteristics have been found until now that are ready to be used in clinical care.
In summary
In summary, yes, it is not just important to examine if a therapy works, but also for whom. There has also been much research examining this question, but unfortunately this research has not yet come up with characteristics that can predict which treatment is better for which patient.
Clinicians often have certain ideas about who benefits more from which treatment, but unfortunately such ideas are rarely based on good scientific knowledge. And this means that it is basically a matter of trial and error which treatment will work for which patient.
- For why no single therapy is best for everyone, see Is cognitive behavior therapy the best treatment for everyone?
- For why differences between therapies are so hard to establish, see Are all psychotherapies equally effective?
- For what to do when the first treatment does not help, see What can I do if therapy does not seem to work?
Pim Cuijpers is professor emeritus of clinical psychology and scientific director of Metapsy. He has been involved in more than 1,100 scientific studies, mostly on psychological treatments of mental health problems. This is one of a series of evidence summaries in which Prof. Cuijpers tries to answer questions from patients and clinicians, based on what is known in science about treatments. The knowledge is mostly drawn from collective work of the Metapsy collaboration of at least 15 years. Do you have other questions you would like Prof. Cuijpers to discuss? Feel free to contact us.
Literature
- Paul GL. (1967). Strategy of outcome research in psychotherapy. Journal of Consulting Psychology, 31(2), 109–118.
- Cuijpers P, Harrer M, Furukawa TA. (2025). Innovations to improve outcomes and uptake of psychotherapies for mental disorders: a state-of-the-art review. World Psychiatry, 25, 4–33. https://doi.org/10.1002/wps.70002
- Cuijpers P, Ciharova M, Quero S, Miguel C, Driessen E, Harrer M, Purgato M, Ebert D, Karyotaki E. (2022). The contribution of “individual participant data” meta-analyses of psychotherapies for depression to the development of personalized treatments: A systematic review. Journal of Personalized Medicine, 12, 93. https://doi.org/10.3390/jpm12010093
