AI and VR – Cambridgeshire and Peterborough experts share vision of healthcare in podcasts

The podcasts are available to download from SoundCloud and the iTunes Store (search for ‘CPFT’)

Artificial intelligence, virtual reality and personalised medicine could help in the diagnosis and treatment of dementia and schizophrenia and other diseases in the future, according to the unique insights of clinicians at Cambridgeshire and Peterborough NHS Foundation Trust (CPFT) in the Trust’s first ever series of podcasts launched today.

Professor John O’Brien, Allison Bentley, Dr Rudolf Cardinal and Dr Emilio Fernandez are all experts in their respective fields. In the three podcasts available to download and listen to now, they discuss the causes and symptoms of, treatments for and research into some of the most common and debilitating physical and mental illnesses.

Future podcasts will delve into the fascinating work of the some of the other 140 research studies taking place at CPFT, including at the Trust’s Windsor Research Unit.

In the first podcast Professor O’Brien, CPFT honorary consultant psychiatrist and Professor of Old Age Psychiatry at the University of Cambridge, and Allison Bentley, dementia research nurse at CPFT, consider why some people are more at risk of dementia, and what people can do to stop the decline in memory and cognition in the brain. Some symptoms of dementia, such as Lewy body dementia, mirror those of schizophrenia, such as visual hallucinations, as discussed by Dr Fernandez in podcast 2.

John O’Brien says: “In general we know genetic factors are important in most dementias. They’re better known for things like Alzheimer’s disease rather than some of the other types of dementia. But in addition there are lifestyle factors as well.

“There are several different lines of research now to try to understand and identify people who are at risk of dementia. These range from biomarkers, biological tests, whether that’s brain imaging, looking in the blood or in the spinal fluid, to try and identify early markers that put people at increased risk. These markers are often based around the abnormal proteins in the dementias, amyloid and tau, but also other things as well, some inflammatory changes, which also seem to be early markers of risk for people.”

“Virtual reality tests to assess memory and cognitive function are being looked at to see if they can be very early markers for people who are at risk of getting dementia,” John adds.

Allison Bentley, who is studying for a PhD, says: “From recent research interviews I’ve completed it’s quite evident that people cope in a wide variety of different ways, from looking up information online, to going to carer support groups, accessing the Carers Trust and gaining support from friends, family and the local community. One of the main issues is recognising that family carers may not have time to look after their own health and may not feel able to say if they are not sleeping enough, having time or actually coping.”

In the second podcast consultant psychiatrist Dr Emilio Fernandez relays a patient’s vivid picture of what it is like to suffer from psychosis, hearing voices or experiencing hallucinations: “The patient finally mentioned to me that she had been hearing a voice continuously but also feeling a hand grabbing her heart and a voice telling her ‘If you disclose that you are hearing this voice I will just grab your heart, you will stop beating and you will die.’ She was scared to death but she did respond to medication and is now working full time and happy, but it was terrifying for her for five years.”

In around 30 per cent cases, however, the antipsychotic drugs do not work and a different drug, called clozapine, is used to treat these patients and is more effective in reducing the symptoms of psychosis. Dr Fernandez has been running clozapine clinics in Cambridgeshire for the last few years and he has developed a tool using anonymised patient records data to help him improve care for his patients.

Dr Fernandez adds: “My goal was to have the information available to make clinical decisions and that’s what made the database that I’ve been setting up is for.

“For example, because we know that sedation is a problem with clozapine I’ve been evaluating every time a patient comes to see me to see how many hours they sleep a night. After two or three years of gathering all the information I could define quickly all the patients that were having sleep problems. I could even define the factors that make them sleep less, which were finding a job, reducing the use of clozapine, but also adding a new medication, in this case it was aripiprazole. So I knew that that was an effective strategy to improve the sedation in the patients in my clinical practice.”

Schizophrenia won’t be seen as a single disease in the future, thinks Dr Fernandez: “When we are able to identify different sub-types of the illness, we could probably do much better targeting or deliver personalised medicine to provide patients with anti-inflammatories, antipsychotics, or antidepressants.”

In the third podcast Dr Rudolf Cardinal, a CPFT honorary consultant in general adult liaison psychiatry, has used a de-identified patient records system to help his colleagues who work with patients with schizophrenia and Lewy body dementia.

The Clinical Records Anonymisation and Text Extraction (CRATE) system was developed by Rudolf and his team and is based on work by London NHS trusts. CRATE, which is open source and free, can take clinical records, remove identifying information to create a research database, automatically collate information, such as questionnaire scores or blood tests from notes typed in by clinicians, and operate CPFT’s pioneering consent system through which patients can choose to be contacted about research.

Dr Cardinal, who is also a university lecturer at the Department of Psychiatry at the University of Cambridge, says: “We have looked at the relationship between different treatments for schizophrenia and how long people with schizophrenia spend in hospital. We examined the relationship between different medications and time spent in hospital, and provided some suggestions that might be useful for clinical care and to ensure patients’ stay in hospital is as short as possible.

“We’ve also looked at the lifespan comparisons of different types of dementia. It turns out that compared to other dementias, such as Alzheimer’s disease, Lewy body dementia is associated with shorter life expectancy, even when you can control for things like age and sex and other physical illnesses. So we’re exploring that further and trying to learn more about why.”

There are particular challenges in psychiatry, Rudolf adds, because lots of the information is recorded in note form – free text –  by clinicians: “We have applied techniques to try and extract information from that free text but also try to ‘scrub’ the identifying information from those letters or documents,” he says.

And what about the future?  Rudolf says artificial intelligence could play a role in the future of healthcare: “Artificial intelligence might be really useful. Sometimes humans are not great at spotting patterns or even of thinking of the patterns. And there are techniques to have machines do the learning and artificial intelligence techniques, that might be better at predicting some of these important things like risk factors for individual patients. So we would like to develop that as well.”

Rudolf concludes: “Research has got strong benefits in the long run for clinical care but we’d like it also to have benefits in the short term. So if you could improve care now for your patient using tools that we’ve developed then we think that’s a good thing and we try to contribute to that.”