Everybody has won and all must have prizes

For eighty years the research has said no. Put two forms of therapy side by side and you almost always find the same thing: they work about equally well. In the literature this is called the dodo bird verdict, after the dodo in Alice in Wonderland who declares after the race that everybody has won and all must have prizes. What does make a difference is the care provider. Some have systematically better outcomes than their colleagues, and that difference is not related to their school, their title or their years of service.

So what does work? Research into common factors arrives at the same list every time. A care provider who is really there and does not judge you. Someone who thinks with you and not about you. Trust. A shared idea of where you want to go. Not spectacular, but stubbornly reproducible.

Proven better, or proven easier to study?

That is uncomfortable for a field that sells itself on technique. And it clashes with the word on which all of mental health care turns: evidence-based. In the 1990s that term was hijacked by the logic of insurance. What was paid for had to be proven, and proof meant: a protocol, a manual, a group of people with the same DSM diagnosis, eight to twelve weeks, and at the end a questionnaire about symptoms. Only therapies with a handbook fit that format. That is why most trials are about CBT and its variants. And that is why many people now think that CBT is proven better. That is simply not true. It is proven easier to study, which is something quite different.

There is another problem with that questionnaire at the end. If you ask people why they go into therapy, they seldom mention fewer symptoms. They mention: understanding myself better, becoming milder towards myself, looking at my situation differently, better relationships. We measure something other than what people come for, and then draw conclusions about what works.

Your brain predicts, and you call that prediction reality

And this is the interesting part. An explanation has recently come on the table that makes the whole quarrel between the schools unnecessary, and it comes from an unexpected corner: the neuroscience of the predictive brain.

The idea is this. Your eyes and ears do not show you what is there. Your brain continuously builds a model of the world and uses it to predict what is about to happen. You do not experience that prediction as a prediction; you experience it as reality. Only when the information from the senses clashes with the prediction does your brain have to choose: do I believe myself, or do I believe what is coming in? Usually the incoming information wins and the model adapts. That is how you learn.

Except when the model has been built from experiences that had to do with survival. Then the prediction wins, also when it is no longer right. Suppose I walk through town and someone passes me without saying anything. And then I think: you see, nobody loves me. While he simply did not have his glasses on. That is not a strange thought from a broken brain; it is a prediction based on old material. Early experiences weigh heaviest here, because as a small child you depended on others and rejection was life-threatening. Those memories are stored in a place your conscious thinking cannot reach.

What this means for psychosis susceptibility

For everyone who lives with psychosis susceptibility this is not an abstract story. This is the core. Meaning does not come in from outside; meaning is made. If you are under great strain, sleep badly, or have lived for years in an environment where distrust was sensible, your brain starts to lean more heavily on its own predictions. The look of that man in the tram acquires meaning. The radio is talking about you. And as long as the correction does not come in, the model keeps confirming itself. That is the opposite of a defect. It is a normal mechanism that has gone too far, in the same way that high blood pressure takes place in a perfectly normal blood vessel.

Safety and precision

What follows from this for help? Two things, and neither of them is printed on a certificate.

The first is safety. A brain that feels threatened holds on harder to its own predictions. Fear makes learning impossible. That is the reason the relationship matters so much, that predictability and availability matter, that social holding matters. Someone who holds you is not a soft precondition for the real work. It is the work.

The second is precision. Safety alone changes nothing; then everything stays as it was, only more pleasant. Experiences have to be added that touch exactly the prediction that is stuck, so that your brain can notice the error and adjust its model. Someone who does not leave at the moment you expect everyone to leave. A group in which you name your voices and nothing bad happens. A job in which it turns out that you can do it. We call that recontextualising: the same experience, a different place in your story.

And this very often does not happen in a consulting room. It is often the passing moments that make your model falter: a neighbour who asks something ordinary at the moment your world starts to arrange itself strangely, a colleague who simply keeps talking, someone who phones for no reason. No protocol, no session, no billing code.

Where those places are

That is what it is really about. If safety and precision are the active ingredients, the question is not which brand of therapy someone gets, but whether someone has a place where this can happen. And then it is rather absurd that we have placed that place exclusively with a specialist with a long waiting list. Recovery colleges, peer support, resource groups, groups of voice hearers, social prescribing, a workplace, a neighbour: these are not alternatives to the real help. These are the places where people can find safety and correction at the same time.

There is work to do

On two fronts. For care providers: stop defending your school and invest in what you share with all the other schools. For the system: fund the places where people hold each other, not only the sessions in which symptoms are counted.

And for you, if you are looking now: do not ask first which method someone uses. Ask whether you have the feeling that this person can handle you, takes you seriously and does not walk away when it gets complicated. That predicts more than the label on the box.

Jim van Os

Prof. dr. Jim van Os, Chair Division Neuroscience, Utrecht University Medical Centre. Jim works at the interface of ‘hard’ brain science, health services research, art and subjective experiences of people with ‘lived experience’ in mental healthcare. 

Jim has been appearing on the Thomson-Reuter Web of Science list of ‘most influential scientific minds of our time’ since 2014. In 2014 he published his book ‘Beyond DSM-5‘, and in 2016 the book ‘Good Mental Health Care’. In 2026 he published 'Wat psychisch lijden is'.

Want to read more?

Blogs by Jim van Os can be read here

Did you know that PsychosisNet regularly posts new content?

Online tests 
Compare antidepressants tool
Compare antipsychotics tool
Podcasts
Books