Data Science. Sounds pretty intimidating, doesn’t it?
In fact, data science actually pops up in our lives in many ways. When we’re on Netflix, data science recommends programmes most relevant to our viewing habits, and those of millions of others. And when we search for information using Google, data science methods are being used to process huge amounts of data in order to work out what is relevant – providing a unique result each time.
But how can we use data science to transform how we understand, treat and prevent mental illness? In September we held our fifth annual Data Science Meeting, bringing together researchers and innovators from around the world to answer that very question. Here’s what we learned...
1. How could data science advance the treatment of mental illness?
One in four of us experiences mental illness each year, yet treatments are nowhere near effective enough - and a painful period of trial-and-error is often needed to move towards a helpful solution.
One of the sessions showcased 3 new ways data could be used to accurately predict and treat mental illness in future. One example is an algorithm that analyses patients’ medical records over time and alerts clinicians when a person’s mental health is deteriorating.
Another uses ‘deep learning’ – algorithms structured to create an “artificial neural network” that can learn and make intelligent decisions - to select the right treatment for each individual. The final project put forward is a system that monitors our activity through our smartphones and uses this information to assess any major changes to a person’s mental health - which can help to find the right moment for intervention.
2. Technology isn’t taking over…yet
In an age of Siri, Alexa and artificial intelligence, it’s safe to say technology is becoming more and more present in our daily lives. Algorithms that can process people’s health data and make recommendations could become more common – and may be used in the future by clinicians to inform their decisions regarding a patient’s treatment.
But is it right to hand over all the power to technology when it comes to mental health? The general consensus was that algorithms and technology can be used to support clinicians’ recommendations, but they should be as transparent as possible and well supported by research. The fact that people sometimes receive mental health treatments without their consent was an important point to bear in mind when increasing the reliance on technology to support the delivery of mental health care.
3. User-led research is the way forward
In this context, researchers could be led down a path directed by technological advances, rather than doing research that is considered useful by people with lived experience of mental illness.
When it comes to mental health research, it was agreed that it must be user-led – putting people who have used and continue to use mental health services front and centre. These people should all have the opportunity to decide on the issues and questions to be prioritised, shaping services and informing scientific advances that are relevant and necessary.
4. Collaboration is key in order to uncover breakthrough findings about mental health
Thousands of pieces of data are collected every day - whether in schools, GP clinics or hospitals - to provide researchers with crucial insight into people’s mental health and help them to tackle some of the major challenges. A huge resource in the UK is cohort data, which is data collected from the same group of people over an extended period of time.
However, there are no cohort studies dedicated to monitoring mental health – which means there’s often a lack of high-quality data for mental health researchers. A current priority in the field is developing mental health research platforms, where all available data on mental health can be linked and made easily accessible.
One example is Professor Ann John’s MQ-funded project, which is bringing data on young people’s mental health under one roof to help researchers and policy-makers worldwide. This will help reduce the costs and time involved in mental health research and assist with groundbreaking insights. The Medical Research Council are doing similar work – last year they announced their Mental Health Data Pathfinder Awards, through which they’ll invest £15 million in projects that can better mine this data and accelerate vital discoveries.
5. By collecting outcome data, we can make positive changes to diagnosing and treating mental illness
In many cases services do not collect data on how effective a treatment is – and if it’s really helping someone to get better. The NHS’ Improving Access to Psychological Therapies (IAPT) programme provides treatment for anxiety disorders and depression – and collects outcome data on all patients.
Over time, this data has informed striking improvements to the service. For example, it enabled them to see that the longer people waited for appointment, the less likely it was that their mental health would improve once they finally got treatment. IAPT also found that consistent improvements in patients were strongly associated with the number of sessions they had with a therapist. Using this information helped IAPT exceed its target of a 50% recovery rate – and proves the case for collecting data throughout the treatment process.
6. In an age of hacking and data breaches, trust is everything
A theme highlighted throughout the day was the importance of public trust. Mental health data science relies on collecting, analysing and protecting people’s data – and there was a unanimous feeling that the public have every right to understand the purpose that their data is being used for, alongside any potential risks.
As Natalie Banner from the Wellcome Trust said, “We mustn’t forget that participating in research is a profoundly generous act.”
Last updated: 15 May 2019