Feature Story | 5-Oct-2022

The rise of tech in health data research

A new and evolving way of treating mental illness in the UK’s population  

Health Data Research UK

Having lived with diabetes since the age of one, 34-year-old Londoner Manveer Sahota has always had a good understanding of his condition, its impact on his daily life and the long-term health implications.  

While holidaying in India aged 16, his parents accidentally injected him with too much insulin and he suffered severe hypoglycaemia when his blood sugar dropped dangerously low. Fortunately, they brought his hypo under control, before he lost consciousness, by injecting him with glucose. 

A few days later, Manveer experienced his first panic attack later leading to anxiety, which he has battled ever since.   




In an office in Glasgow University, Health Data Research UK - UK Research and Innovation Fellow and researcher Dr Rona Strawbridge is investigating the biological links that contribute to mental illness, diabetes, heart disease and obesity.   

Dr Strawbridge moved to Glasgow University in 2017 and started to use genetic data to explore the observed link between psychiatric disorders and cardio-metabolic diseases such as diabetes - and why these individuals have a life expectancy that is 10-20 years shorter than the general population.  

“For a long time, lifestyle factors have been thought to account for these conditions that occur together in people with mental health issues – like diet, levels of physical activity, and smoking; but actually, genetic studies show there’s more to it, and the same biological pathways might be contributing to both mental and physical health conditions,” she says.  

On learning about the HDR UK-facilitated research, Manveer says, “I had always suspected there was a link between mental illness and diabetes. I knew my diabetes was making me anxious and that the anxiety and depression was affecting how I managed my diabetes. But I didn’t think there could be a genetic link to this.”  

Dr Strawbridge’s hope is that by identifying links between mental and physical conditions, opportunities may present themselves “such as the repurposing of existing drugs for conditions like heart disease and diabetes for people with mental health conditions.” 

Rona goes on to explain that another advantage of linking data is that clinicians can intervene earlier to manage a patient’s mental and physical health.  

“Someone with a mental illness might be more vulnerable to poorer health outcomes, which could impact a clinician’s treatment decisions,” she says. “For example, if someone with a mental illness is struggling to control their diabetes, maybe they should be prioritised for more frequent monitoring of blood pressure as it is an indicator of heart disease, kidney function and eye health.”  

“Because my anxiety is linked to my fear of hypoing,” says Manveer, “I sometimes kept my sugars too high. Sadly, this has caused damage to my eyes, kidneys and nerves.” 

A family history of diabetes is often recorded when considering the likelihood of the condition presenting. “Perhaps family history of mental illness should also be considered in risk monitoring and in the prediction of physical conditions,” Dr Strawbridge adds with the caveat that previous generations may not have talked so freely about it.  

Beyond linking existing data, HDR UK works to improve the quality and diversity of the UK's data asset to improve the quality of the datasets, enable new discoveries and understanding about disease. In turn this provides fairer and more equal access to the latest treatments and medical technologies, benefiting as many people across the UK as possible.  

HDR UK’s mission to unite the UK’s diverse health data is praised by Dr Strawbridge, particularly as she considers the risk of making conclusions from existing data that is often gathered from individuals of European ancestry.  

“There is much less data on non-European ancestry individuals for a lot of conditions,” she says. “We need to be operating in a healthcare system with diverse data that is truly integrated. The more data we have access to, the more we can isolate trends and tailor care for certain groups.”  

Manveer says he has never had a problem with his data being used but now he understands how it is used to drive improvements in individual care as well as the population’s health, he wants to get involved.   

“I am very happy to share my story and my data, but it makes sense for it to be anonymised in research,” he says. “And if research can lead to early intervention that can help people with certain conditions, shouldn’t we all do our bit?” he says.   

Mental illness affects almost everyone at some point in their lives and, is the biggest cause of ill-health for people living in the UK. People living with a mental illness are less likely to take part in research studies and may also be actively excluded. Much of what is known about mental health is not based on the very people who are worst affected.   

To tackle this, HDR UK and the Medical Research Council (MRC) helped support the creation of DATAMIND, a collaborative hub for mental health data which aims to improve the use of big data for mental health research.

Co-director of DATAMIND, Ann John says, “The UK has a genuine opportunity to be world-leading in mental health data science, and the work of organisations like HDR UK and DATAMIND are making exciting progress towards this ambition. By working together with the public, patients, researchers, industry and the NHS we’re transforming both our understanding of mental health and the lives of people experiencing mental health problems. And we’re creating a hub where researchers and others can find and use mental health data to benefit patients and the public and improve care.”   

Health data, collected by GPs in routine health records rather than from specific clinical trials represent a wealth of information that can help to improve outcomes and tackle longstanding disadvantages faced by those with mental illness.  

However, the clinical notes produced by mental health professionals are typically written descriptively, meaning data on mental conditions is not captured in a way that can be easily analysed by computers. Until fairly recently, this prevented large-scale studies using this type of mental health record.  

Technicians and researchers at South London and Maudsley NHS Foundation Trust supported among others by HDR UK solved this problem by developing platforms to extract and analyse mental health records, alongside data security and governance frameworks to keep information anonymous and patients involved in its oversight. This Clinical Record Interactive Search (CRIS) infrastructure has been operational for 15 years and has supported over 250 research publications. 

The South London and Maudsley Trust is one of the largest providers of mental health care in Europe allowing researchers to analyse data from about 500,000 patients and 30 million clinical documents. 

Teams of bioinformaticians, biostatisticians and clinicians have designed a range of automated text mining tools to extract the information alongside manual analysis. For example, this approach has recently allowed the team to identify 21 physical health conditions in the records of the mentally ill patients. 

Richard Dobson, Professor of Medical Bioinformatics at King’s College London, who works in partnership with the Trust, says: “The key for us, has been to accept how data is collected and adapt our research style to this. Instead of building new, structured systems and trying to get clinicians to use them, we’re working to find ways to analyse existing health data.  

“For example, our tool CogStack is able to extract data from these different systems and sources - like PDFs, word docs, text fields and more - and is trained to understand and analyse the 'natural language’ in them. It does this by turning the data into something that is structured, like a spreadsheet, and coded into medical terms that machines can understand.”  

The result is data that can be discovered by researchers via a search-engine, and then linked and analysed at scale to improve public health.  

Since its inception, cutting edge AI research has required very specialised technical expertise, a lot of computing resources and a healthy budget. Now, systems like CogStack and CRIS are creating user-friendly interfaces and simple uploads allowing those in health with no computing background to work on datasets.  

This means healthcare analytics will continue to grow and evolve alongside the digitalisation of the NHS although the impact is only starting to be felt by Manveer and others of his generation.  

However, with the rise of cutting-edge software enabling clinicians to develop bespoke care plans alongside traditional ways of improving a patient’s health, today’s research will be tomorrow’s care for all.  



Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.