Thanks humans, we’ll take it from here…

Professor Hamish McLeod is a co-investigator on the EMPOWER study, playing a role in encouraging reliability in research measurement and supporting people using the EMPOWER App. Here he describes the digital explosion as a fourth industrial revolution and considers implications for mental health care and support. 

If you look at the headlines, you might think that we are on the cusp of yet another revolution in technological capacity that will transform society and industry. There may be some truth to this. Seismic changes have happened before with the various industrial revolutions that have dramatically altered the way we work. The first industrial revolution came with the mechanisation of the textile industry in late 18th century Britain. People like Henry Ford gave us the second industrial revolution with the development of production line methods that changed the way that complex products such as cars could be speedily and efficiently mass produced. And now we are deep in the third industrial revolution marked by technologies such as 3-D printing, robotic manufacturing methods and increasing reliance on computers to do tasks faster and with more accuracy than humans can.

Some people have even begun to predict the dawn of the fourth industrial revolution where Artificial Intelligence (AI), cloud computing, big data, and analytics will further disrupt and transform society. All industries are expected to be affected, from manufacturing to healthcare and education.  But, does this extend to issues of mental health and wellbeing? Self-driving cars might be the taxis of the future, but will technology replace real humans who do mental health work? This blog takes a peek under the hood of digital healthcare and tries to separate the hype from believable trends.

Self driving taxi

Self-driving cars might be the taxis of the future, but will technology replace real humans who do mental health work?

The Excitement of New Technology

I should start by acknowledging that I like the excitement of being involved in developing new technology such as the work we are doing in the EMPOWER trial. Through this study we get to do important clinical work with a modern technological twist and even though we are at an early stage, we are already learning a lot of things that we didn’t predict before the project started. Our main aim is to test and refine a smart phone app that could help people to understand and manage the ebb and flow of their wellbeing (including psychotic symptoms) over time. This information about fluctuations in wellbeing could be enormously valuable in the early detection and prevention of the worsening of distress from psychosis.

A positive outcome would be very exciting – maybe digital algorithms and smartphone apps really will bring about a revolution in mental health care. That would be very cool. But, we have to remember that novelty and innovation can be extremely alluring merely because it is new. We can all invest a lot of hope in the prospect that a new gadget or tool will defeat a problem that we have been struggling with for a long time. Distressing and life disrupting psychotic relapse is one of those challenging problems that we would love to be able to manage in a much more effective way. But, my scientific training and multiple lessons from the past reminds me that hoping that a new strategy will work is not the same as it actually working. We need high quality data before we can trust new strategies enough to warrant recommending them to the people who want new ways of managing their wellbeing.

my scientific training and multiple lessons from the past reminds me that hoping that a new strategy will work is not the same as it actually working

A New Californian Gold Rush?

In 1848 James W Marshall found gold at Sutters Mill in Coloma, California, marking the start of a seven year gold rush that brought 300,000 people to California seeking their fortune. Coloma is about 140 miles from San Francisco, the city most closely associated with the modern-day gold rush of the third and fourth industrial revolutions.


A second Californian gold rush?

Plenty of people in Silicon Valley are keen to tell us that they can use technology to close the gap between mental health care needs and the capacity of traditional mental health services. For instance, a former Uber executive has recently hit the headlines for raising $3.75m to launch a mental health start-up born out of his own experience of mental health difficulties experienced in the pressure cooker world of app development.

There are also “digital therapist” chatbots such as Woebot and Wysa that have begun to attract attention as a sign that AI might soon replace human therapists. To be honest, I think this reflects the kind of over-reach that happens when people get excited about a new idea. There are also commercial incentives to encourage tech entrepreneurs to over-claim that their new wellbeing app can transform your life. Lots of money can be made from closing the gap between mental health treatment needs and access to suitable care. As a result, there are currently more than 325,000 health and wellbeing apps available across app stores with 78,000 of them added in 2017 alone. So, how can you choose an app that is right for you?

Hey Siri – What’s the Evidence Base for Wellbeing Apps?

Siri’s AI algorithms will struggle to help you access much evidence about the effectiveness of mental health and wellbeing apps. It is possible to design and launch an app without any proof that it makes any difference to the problems being addressed. Instead, the market is allowed to decide – the presumption is that good apps will become more popular and the best ones will come to dominate. Sadly, this argument does not hold water as we have learned to our detriment with the antibiotic resistance crisis that can partly be attributable to the natural human desire to try any treatment when you are suffering – even if it is known that the offered treatment isn’t going to be effective.

the market is allowed to decide – the presumption is that good apps will become more popular and the best ones will come to dominate

One solution to this problem is to sift the products with an evidence base from those that are untested or proven to have problems or side effects. In the UK, NHS Digital provide a curated library of apps that have sufficient evidence of their suitability for public use. So, how many of the more than three hundred thousand wellbeing apps in market make it on to the list? Seventy-five are currently listed and of these, only 18 have a mental health focus. That’s 0.000055% of the total pool of apps! No wonder Siri has trouble helping you find the evidence base.

Hey Siri

Siri may struggle to find evidence based wellbeing apps.

Using Data for Good, Not Evil

I have already discussed in a previous blog how data protection and management of privacy are fundamental issues that need to be addressed in the development of technologies like EMPOWER. The data we collect from EMPOWER is managed to the highest standard of privacy and confidentiality protections so that people involved in the trial can be reassured that their personal information will not be accessed without their consent. We also use data to test whether the EMPOWER app is showing enough signs of effectiveness to warrant spending money on a large scale definitive randomised controlled trial. This is a long and challenging process, but one that will mean that any future offering of EMPOWER to the public will be based on data, not hope and hype.

EMPOWER is also the first even mental health app to be subjected to monitoring and approval by the MHRA – the government agency that regulates the approvals of drugs and medical devices. This provides the highest level of protection for the public by ensuring that any claims about the effectiveness and safety of EMPOWER will be backed up by independently scrutinised data. Unlike the “tech gold rush” approach of rapid app development without any robust testing of efficacy, our approach means that it will take a lot longer for the EMPOWER app to be widely accessible to the people who might benefit from it. But we would much rather use data and the scientific process to ensure that EMPOWER is rigorously tested and checked before it hits and app store near you.

So, thanks for the offer AI but for now we’ll use our own non-artificial intelligence to solve the problem of how technology can help us to provide better quality care and support to people who are learning to more effectively manage the ebb and flow of psychosis.