An early warning system for schizophrenia relapse, using Fitbit data and a smartphone app, could lead to dramatic improvements in treatment.
A group of London-based doctors and software engineers is developing the experimental system at London’s Bioinformatics Core of the NIHR Maudsley Biomedical Research Centre (BRC). Working from the centre, where clinicians, software engineers and statisticians seek out patterns in vast medical datasets, the team is studying the sleep patterns of schizophrenia patients to develop a possible alert system. “It’s well documented by clinicians that sleep disturbance is one of most highly self-reported factors [prior to a relapse in psychosis],” Amos Folarin, the BRC statistician behind the trial’s SleepSight app, told WIRED.co.uk. “But it’s hard to cover objectively.” “Most people use a sleep diary, but that is probably not reliable in populations of people with mental illness,” specialist registrar in psychiatry Dr Nicholas Meyer adds.
If there were a way to track sleep patterns in realtime, share that data with a physician, and alert them as soon as anomalies in those patterns occur, patients could potentially be treated prior to a full relapse.
According to project co-lead Dr Richard Dobson, patients are released fairly quickly into the community following their first “episode”. As a result, an inexpensive automatic alert system could be an excellent way to keep in touch with a patient and ensure they get the treatment they need before another episode strikes.
Enter the Fitbit Charge HR, which measures heart rate and movement. It will be used in an upcoming pilot trial of 20 patients (currently being recruited) to calculate how restful a wearer’s sleep is. The reason for using the commercial device is simple: medical-grade equivalents are neither built for long term use, nor cost-efficient enough to be handed out to patients — they cost around £300-500, versus the £120 Fitbit. “Broadly, the underlying tech is the same,” says Meyer. “But Fitbit also has Bluetooth.”
The use of Fitbit in medical trials is gaining credibility, he adds. But it’s also adding layers to trials that would otherwise have been impossible. The company has recently given researchers access to its real time data available via API partnerships, and patients on the upcoming trial will sync it up with a smartphone app called SleepSight.
Data from the phone will also be gathered, including accelerometer, light and battery data. The team is seeking approval to use information on messages and calls — times and activity, rather than content. They are taking a bit of a punt with the latter, not knowing how useful the outcome will be. But, Meyer says, “this is potentially a really rich source about various behaviours, showing what the life of someone with mental illness is like.” “When relapsing they might initiate fewer contacts, or use email or social networks less, for example. We don’t know, but it could be really important information. The more variables we collect, the better.” The team is also calling for the public to send in sleep data, as a means of developing a control set.
The trial will continue for two months, but anyone that takes part will be given the option of keeping the Fitbit for a year and sharing that data. Meyer points out that within a five year period, 80 percent of people with psychosis relapse. Whether that statistic is due to individuals failing to take medication, or the dose no longer being sufficient, the team wants as much data as possible to be able to calculate when sleep pattern anomalies can also be helpful indicators of relapse.
For the initial trial, the team will recruit “relatively stable” patients not undergoing a relapse. “Future studies will recruit more unwell patients who may have a history of not taking medication,” says Meyer. “The point is to show sleep objectively and uncover the relapse signature.” Meyer frequently hears from friends and families of patients that a sufferer’s sleep is highly disturbed prior to a relapse — he wants more than anecdotal evidence to prove it, and act on it in good time. In the meantime, they can also see if restless sleep corresponds with those days an individual experiences more paranoia, “even though people might not relapse fully”.
There is perhaps one significant issue surrounding the use of Fitbit that could feasibly be cause for concern. Many wearers find the data addictive, and behavioural changes — though the desired outcome of such a fitness product — can be significant. Apply that to an individual suffering from a mental illness, where paranoia and obsessive behaviours are common, and surely it would be a recipe for disaster?
Not so, says Meyer. He and the wider team strongly believe that giving patients a device, and the data that goes along with it, also gives them a sense of control over their own health and a sense of agency so frequently lacking when care is placed in the hands of senior physicians that have little time to get to know each individual. “The first thing people say is ‘that sounds like a crazy idea’, but I look at it another way,” says Meyer. “If you give them a tool that they understand, and work with patients, it’s something that will help you in building up a trusting alliance. Then there will be a chance people use it longterm. I hope we can show that people do use it and they have benefit, not just dispose of it when they become unwell.” “We want to promote the idea of self-monitoring, where they can see for themselves ‘I used alcohol there and then my mental state deteriorated over the next few days’. At the moment people get relatively little feedback — there’s a top down approach of ‘you must take medication because it’s good for you’. But if they see the consequences of not taking medication, or not sleeping well, it works to increase insights.”
No amount of technology, he agrees, could ever replace contact with a clinician the patient trusts, but in overstretched areas technology like this can be a way to ensure those in the community with mental health issues do not feel powerless, or alienated. Considering the NHS was one of the main battlefields of the recent General Election, with every party making greater promises for more GPs, nurses and budget than the last — but none pledging an increase in social care funding — it’s clear the services need additional support from somewhere, and technology can play a key role if used correctly. Meyer believes as trackers become ubiquitous in wider society, they will find a place in monitoring disease for preventative healthcare measures. This is in fact already happening on other small preventative health trials funded by NHS Trusts around the country, related to heart disease or obesity research.
Multidisciplinary places like the BRC are where some of the most interesting research in these areas is likely to come from. The unit is based within the King’s College London and the South London and Maudsley NHS Foundation Trust, Europe’s largest provider of mental health services, which has meant BRC researchers have access to an incredibly diverse range of rich datasets, including (anonymous) genomic data and brain scans from trials featuring some of the 250,000 patients the Trust has seen since 2007.
The ultimate goal? “Bringing personalised medicine to mental health issues,” says Dobson, just as healthcare practitioners are trying to do for so many other areas of the medical industry.