terovesalainen - stock.adobe.com
Universities use data analytics to tackle student mental health
UK universities turn to data analytics software to tackle student mental health. Are there duty of care lessons for other sectors?
Four years ago, after a student at the University of Manchester’s School of Physics and Astronomy committed suicide, lecturer Andrew Markwick wondered if the institution could try to identify students that might be struggling.
The school holds a variety of data on its students, including attendance at meetings and lectures and submission of work, but it was in a disparate set of systems.
So he built StudentCRT, which provides the personal tutors who guide individual students and welfare officers with a single view of this data and generates a score to spot those at risk. “It’s about identifying those students and having a conversation with them as soon as possible,” Markwick says.
Those conversations can reveal an innocent reason such as a holiday, but “it can unlock a whole world of struggle”, he adds, and some serious problems have been uncovered as a result of the system’s operation.
Since its introduction, the school has referred fewer students to the university’s counselling service while referrals from other schools have risen, indicating that the system is helping, although Markwick stresses that this has not been proved by research. It has also made it easier for staff to support students.
StudentCRT now covers data on some 2,000 students in the university’s physics and mechanical engineering schools and other schools may follow, although Marwick adds that it might be easier to implement in science-based subjects with more structured teaching methods where participation can easily be measured. He has founded a company, Third Floor Systems, to market the system to other universities.
Universities are certainly interested in providing better mental health support to their students, with numbers reporting mental health conditions having risen by nearly six times between 2007-2008, and 2016-2017, according to the Higher Education Statistics Agency. Universities UK, which represents the sector, says that at least 95 students killed themselves in 2016-2017.
Most students appear open to having universities monitor their mental health, with 66% happy for their parents or guardians to be contacted in extreme circumstances if their university is concerned about their mental health, and a further 15% content for this to happen in any circumstances, according to the Higher Education Policy Institute’s annual survey of 14,000 students.
Generating alerts from data
Jo Midgley, pro vice-chancellor for student experience at University of the West of England, says that today’s students face tough conditions, with most undergraduates borrowing tens of thousands of pounds to pay fees and living expenses, and working while studying to support themselves. The costs lead to high expectations from students, compounded by the pressures of social media. “Some of the anxiety we see is understandable,” she says.
But the result is that the university has seen demand for its counselling services rising unsustainably. So it is in its first year of using a system provided by Solutionpath to improve student well-being. “We were really keen to get upstream,” adds Midgley.
Each week, the system analyses a range of data on all foundation and first-year students (future year groups will be added as they matriculate). It looks for students whose engagement appears to be lessening, then emails or telephones them to discuss, typically contacting around 10% to 15% of students as a result.
Midgley says that the system has revealed increased anxiety from late January to early March due to particular academic demands at this time. In the future, the university may provide more advice before Christmas to pre-empt such problems, and expects to learn more about how it can change its processes to lessen stress.
But it also hopes to help many anxious students realise they are not mentally ill and do need counselling. “If we can deal with problems much earlier and smaller problems far more effectively, then we can ensure that our resources are used in the best way and that we can far more effectively support those students who genuinely need that higher level of support,” she says.
Nottingham Trent University is adapting an existing system that it uses to monitor engagement. It was Solutionpath’s first customer in 2012, but has recently started using an alert feature triggered if a student has had no engagement with the university for 14 days.
Jo Midgley, University of the West of England
Using this rather than a formula means staff and students understand it, although the university is looking at varying the period so newer students are contacted earlier.
As well as monitoring attendance and work submission, the system includes use of the university’s virtual learning environment (VLE), through which students access course materials.
“What we are finding is that students with mental health conditions are far more likely to generate an alert than their peers,” says Ed Foster, the university’s student engagement manager.
The alerts, which in its first year have been triggered for 3% to 4% of students, are sent to personal tutors, who can then contact the student. They also have access to a detailed dashboard on students’ engagement. “It helps them to prepare for that first discussion with a student, giving them some pointers,” says Foster.
Nottingham Trent also provides students with access to their own dashboard, allowing them to see how they compare to averages for their peer group. Foster says that some students have found themselves motivated by the realisation they were falling behind.
But some students at University of Leeds oppose such dashboards, according to Professor Neil Morris, dean of digital education. He is working on a code of practice for learning analytics developed with students, in advance of the university introducing a system of this kind.
“We have students who have said, ‘I will not look at this dashboard, I will not log in, because I don’t want to know how I’m doing against to everyone else’,” he says.
Some express worries that such systems will not consider students as individuals.
“Students ask, ‘What are you going to do in terms of making decisions about me as a person, when I’m an individual not a data point?’. That’s been quite challenging,” Morris says.
These concerns have been reflected in the code. “We’ve been very clear that we only want the data to inform a conversation, so it’s the starting point for a conversation between student and tutor,” says Morris.
The university will focus on using data to help individual students to overcome barriers and to improve how it delivers its teaching, with Morris saying: “We’re looking at it from a student success perspective.”
He adds that students studying remotely could particularly benefit, as most or all of their engagement takes place digitally – making it easy to measure and assess accurately – and they often study while working, meaning they are more likely to welcome advice.
Predicting mental health concerns
There is also work to develop systems that can be used across higher education, with Northumbria University leading a £2m project largely funded by the government’s Office for Students to use a range of data to predict which students are at risk of mental health crisis.
Working with other universities, students’ unions, charities and suppliers including Microsoft, the project aims to provide tools and techniques by 2021.
James Murray, whose son Ben committed suicide in 2018 when he was a student at Bristol University, is working with Universities UK on the project.
Murray has worked in data analytics and believes universities should consider a wide range of information in identifying at-risk students, including their behaviour.
“Serious concern is when someone is displaying a pattern of behaviour that is out of the ordinary for them, and that could potentially be leading towards a crisis state,” he says.
Murray says it is crucial that institutions do a better job of joining up what they already hold. His son’s inquest showed that information about his illnesses and anxiety was discussed in emails, but was not recorded on the spreadsheet used in the meeting which decided to dismiss him, partly due to data protection concerns. Ben killed himself soon after receiving a letter passing on this decision.
“The one time where that it could have been vital, potentially in saving a life, was at that dismissal meeting,” James Murray says. “There are critical moments during the processes that people are following where the injection of information could make a difference, could be life and death. It was life and death in Ben’s case.”
University of Bristol, which has since changed a range of processes including how it handles student dismissals, is one of those involved in the Northumbria project.
Mental health at work
Murray sees potential for similar systems in the workplace, such as processes that may lead to suspension or firing.
“Technology should enable this to happen more safely by bringing the right information to the right person at the right time, to make the right decision,” he says. As well as helping to protect employees, organisations may protect themselves from legal and reputational damage.
Some suppliers are looking at what they could offer employers to support staff mental health, although not yet in ways that could identify individuals at risk. People First, a service which asks employees questions on engagement with work, provides consolidated data for employers, but only for groups rather than individuals.
Mark Williams, chief product officer at People First, says that individuals would not provide honest feedback if they could be identified.
Wellbeing app provider Remente, which provides training and advice, similarly carries out regular surveys of employees for its corporate subscribers and again provides only consolidated results to employers.
Andrew Markwick, University of Manchester
Chief executive David Brudö says that the system works better when it is not mandatory. “There needs to be a pull rather than a push. This should be a resource that is offered from the employer,” he says.
There might be potential to go further, particularly with specific groups and types of employee. Brudö says that young employees particularly value work-life balance, while People First’s Williams reckons that remote workers could particularly benefit from such services given they are at risk of isolation.
University of Manchester’s Andrew Markwick says that he has discussed StudentCRT with police authorities, armed forces and those managing sportspeople.
“The thing that unites our students and these groups is that these people are pretty talented, but they are under pressure and in some cases they struggle in silence,” he says. Some may find it hard to admit they have problems, which makes identifying them all the more important.
Although systems designed to spot for mental health problems may start with data analysis, Markwick adds that this should be followed with human contact, something echoed by other universities.
“A lot of cases are dealt with simply by having a conversation,” he says. “The fact that a student knows someone is looking out for them – not to tell them off, but because they care about them – I think that matters.”
Read more about data analytics and business applications in Higher Education and health
- Qlik Sense and predictive analytics in education flag troubled students.
- Predictive analytics in healthcare helps manage high-risk patients.
- Customer relationship management software may be targeted at businesses, but the healthcare and higher education industries have redefined it to gain better insight into their users.
In the UK, Samaritans can be contacted on 116 123 or email [email protected]. Helplines for other countries can be found at www.befrienders.org.