Mark Brown takes a look at new research which is aiming to personalise mental health care with sophisticated algorithms – so more people can get treatment that works for them.
Sometimes when we experience mental health difficulty and seek treatment or support it can feel we are a round peg being hammered into a square hole. We are individuals. We are all made up of idiosyncrasies; each of us a living history of what we have been, how we have lived, where we have travelled and what the world has done to us and with us. We are not only our age; not only our gender; or sexuality; not only our mental health difficulties or our distress. We are all of those things and more. Despite this, in mental health, treatments often feel like a very approximate fit, where we have to ‘suck them and see’ before we know whether they work or not.
I recently spoke to Zach Cohen, a researcher seeking to solve this problem with a project using statistical modelling to predict with greater accuracy which mental health treatments will work for people. Zach and his colleagues have been working on something they are referring to as a Personalised Advantage Index, a way of working out which treatments might be best for an individual based on who they are. To create this personalised treatment suggestion engine they have been mashing up data from existing studies with knowledge of the way that treatments interact with each other and what we know about the ways different people respond to them.
Mental health treatment doesn’t always make people feel like they are regarded as individuals. Medical professionals are often unlikely to be fully up-to-date with all existing research and are likely to fall back on past experience and recommendations from colleagues, something as true for talking therapies as it is for medications. We know what symptoms a particular treatment or intervention is supposed to work for and have a hazy-at-best knowledge of which kinds of people might respond best to a particular treatment or intervention. What we often don't know is whether this particular treatment will be the most effective for that particular person. What might work for one person with a particular diagnosis might prove actively harmful to another. In mental health, diagnosis alone often overshadows consideration of other factors.
Over the last decade or so the idea of precision medicine has been gathering pace. The US National Research Council defines precision medicine as an approach to medicine that gives practitioners “the ability to classify individuals into subpopulations that differ in their susceptibility to a particular disease, in the biology or prognosis of those diseases they may develop, or in their response to a specific treatment.”
Many factors can lead to being prescribed treatments for a particular diagnosis that might not be helpful, especially in a mental health system where choices might be limited. Too often people feel they must like it or lump it. The intention of Zach’s work is to make it possible to suggest to someone experiencing a mental health difficulty a treatment or intervention will work best for them based on age, gender, life situation, level of prior experience of treatments or other factors.
At present, Zach’s project has a model made from previous research that they are testing against existing large-scale studies to see whether the treatments their Personalised Advantage Index predicts matches with what actually had the best outcomes for people.
The promise of mental health treatments that start with who we are; in all of our weird, lumpy, contradictory glory, is a tantalising one. When experiencing mental health difficulty people complain of either being told that they must endure a period of trial-and-error to find the best fit treatment for them or that they should learn to live with the treatment they are offered even if it doesn’t feel right or doesn’t work as anticipated. Zach’s work could make mental health prescribing more intelligent and responsive to individual situations, histories and circumstances. Zach and his colleagues are working on something that might help to end the curse of one-size-fits-all treatment decisions and bring a new precision to the treatment people are offered.
Last updated: 7 September 2017