July 27, 2017
Predicting clinical outcomes with PRONIA
By Stephen Wood
One important aim of research in the early phase of mental disorders is to better predict who will become unwell, and who will have severe illness.
Recently I attended a workshop in Pisa, Italy, dedicated to this issue. The workshop brought together investigators funded by the European Union through a specific call to develop effective imaging tools for diagnosis, monitoring and management of mental disorders.
One study presented in Pisa of particular interest to me, because I have been involved with it from the beginning, is PRONIA. This study, which will eventually include more than 1700 participants, has just started analysing the initial dataset. The ultimate aim is to develop a tool that predicts clinical outcomes on the basis of self-learning algorithms – that is, that will learn the best patterns that predict outcomes.
So far, we have been looking only at predicting a person’s functioning and, in Pisa, Dr Nikos Koutsouleris presented some data that suggested this might be possible with accuracy of more than 80%. This would be of great use clinically, as it would allow treatments to be delivered in a more focused way; specifically, it would mean people who needed more interventions (of whatever sort) could be targeted to receive them, while preventing over treatment of those who need less.
It remains to be seen if the result holds up in our validation sample, but at the moment things look promising.
For more, follow IEPA Network and Stephen Wood on Twitter.