Skip to main content

Unfortunately we don't fully support your browser. If you have the option to, please upgrade to a newer version or use Mozilla Firefox, Microsoft Edge, Google Chrome, or Safari 14 or newer. If you are unable to, and need support, please send us your feedback.

Publish with us

Mental health researcher uses AI to achieve what’s important to patients

September 24, 2020

By Libby Plummer

A fascinating journey from football data scientist to mental health researcher and team leader

Caption: Dr. Sagar Jilke, a Post Doctoral Research Associate in the Institute of Psychology, Psychiatry and Neuroscience at Kings College London, talks about how his unusual career path allowed him to carve a niche in the world of mental health.

Conducting mental health research during a pandemic has its challenges. For Dr. Sagar Jilkaopens in new tab/window, a Post Doctoral Research Associate at the Institute of Psychology, Psychiatry & Neuroscience at King's College London (KCL)opens in new tab/window, the solution has involved drawing on his unusual background in industry and technology.

Having completed his PhD jointly at Imperial College London and Goldsmiths College, University of London, he is now three years into a five-year fellowship at KCL, where his work encompasses the important task of supporting a number of junior researchers. While a postdoc traditionally focuses primarily on research plus a small amount of teaching, Sagar's role is split into two distinct arms, he explained:

The first is research, including quantitative aspects, like coding, big data, and number crunching. And then there's the qualitative side, like interviews and focus groups with mental health service users. The aim of all this is to ensure that mental health patients' voices are being heard in research and that we're working on stuff that's important to them. This also includes facilitating patient and public involvement meetings.

The second arm is all management. I'm responsible for supporting the career progression of my six pre-doctoral researchers. We just had the last round of recruits leaving: one's gone off to do a PhD, one's gone to become a clinical psychology, one's doing a master's and one's enrolled in medical school.

Research evolves in response to what’s important to patients

Sagar and his team work on a range of research projects simultaneously, with the current lineup including a timely paper on the impact of social media on mental health. "The research projects that we work on evolve in response to what patients tell us is important," Sagar said:

Sometimes mental health patients don't like scrolling through social media because while they might be there looking for support, searching certain terms such as 'psychosis' or 'schizophrenia' can dig up hate-filled search results. Obviously, this is not ideal for somebody experiencing these symptoms.

This particular project is based on developing a machine learning algorithm to automatically identify stigma on social media.

Another project is aimed at understanding what patients think about Ketamine as a form of depression treatment. Widely known as a horse tranquillizer or an illegal party drug, Ketamine is being repurposed as a potential antidepressant, so it's essential to understand what patients think about this and the best way to go about any future rollout.

While Sagar's team has adapted quickly to remote working as a result of COVID-19, fine-tuning the logistics of their research has been slightly more challenging. Not only did the team have to modify their methodology to accommodate the remote nature of the work, they also had to give a great deal of consideration to people's access to technology:

We've already run one study online, but it's been a little bit cumbersome because we had to apply for ethics amendments because we were changing the procedure of the study. We had four or five studies on the go, so we had to pause some of them and focus on amendments, or ones that were a priority.

Building rapport with participants online is very different to speaking in person, but it's been OK. The technical aspects are more challenging: we have to call the participants beforehand to make sure they understand what's going on, and we have a 15-minute window at the start of each focus group to ensure that everyone can access the link OK. We have to be aware that some of our participants might not have easy access to a laptop or the internet, so we need to make sure that they have the equipment required.

Scopus award helps shape his career path

Tackling the problems presented by the pandemic is no easy task, but Sagar has a slightly wider range of professional experience than many researchers. Following his time at Imperial College London, where he won the Scopus Young Researcher UK Award in the Arts & Humanities category in 2015, Sagar took a slightly unusual route to his current role, with a stint at a London tech startup as a football data scientist. Working for a company that traded on football matches, Sagar's job was to use machine learning to predict the outcome of football matches as accurately as possible, enabling clients to place bets on the results. As an avid Arsenal fan, Sagar was well suited to the subject matter but found the fast-paced culture and seven-day working week to be a challenge. "But I have no regrets," he said. "I'm really glad I tried it and I learned a lot."

Early Career Researcher UK Awards

The Early Career Researcher Awards recognize and reward outstanding early career researchers making a significant contribution to their field and thus having a positive effect on society. Nominate your most talented peers for this year’s Researchers’ Choice Communication Award:

One of the many things Sagar learned in his football-flavored role was learning to code in Python. A comprehensive understanding of this ubiquitous programming language is one of many elements of Sagar's data science experience that he's been able to bring with him to his current academic role. "I now teach PhD students at King's — I've set up a module and a couple of courses in data science and Python," he explained. "And I guess my USP is that I'm teaching it from an industry perspective."

Sagar believes that the commercial understanding gained from this role has changed the way he works as a researcher and allowed him to be more efficient:

When I came to this post at KCL, I was able to automate a lot of things. So I write a piece of code and automate a certain task, and then it would run every week or every month. I think that's a result of that commercial mindset, where time is money. I know that if I just spend a bit of time writing a piece of code once that will run itself every month, then it will serve me for the next three years.

While Sagar has been able to port many of the skills that he honed at his previous job to academia, what's the biggest difference between the two roles?

I think it's the rigor. In the papers I write now, the depth in which I have to describe the methods is practically a paper in itself. The depth in which I need to explain the models and the technicalities can take the same amount of time as the actual data collection.

Whereas in the industry, you never really needed to explain the subtle differences in the models. All people wanted to know was what's going to give the best outcome. And then it's up to the data science team to find it. So it's the trade-off between depth and speed.

Sagar credits the Scopus award with giving him the reassurance he was on the right path and the confidence to apply for the data scientist job as well as his current KCL role: "It was a really big boost. I think having that on my CV and having something like the YouTube videoopens in new tab/window that I can link to really helped me. The fact that it's a peer-reviewed award, and that I was nominated and won it based on my achievements, was very reassuring. And I think it gave me a lot of confidence because prior to that, I felt a little bit of imposter syndrome."

Like many of his peers, Sagar still uses Scopus for searching articles. Without it, he says, finding literature would be incredibly time-consuming and result in significant mental fatigue.

Using the right tools is a given for a busy postdoc, but does Sagar have any other advice for early career researchers?

Just go for it, full throttle. If this is what you enjoy — collecting data, understanding people, academic literature, discovering novel findings — go for it, without any reservation