A research team based at Kings College London is preparing for the launch of a first-of-its-kind study for managing and preventing diabetes amongst mental health patients.
People with conditions such as schizophrenia or bipolar disorder generally tend to develop cardiovascular disease at a much earlier stage than normal. This is also true for other physical health issues such as diabetes and stroke, which has been shown to result in higher incidences of mortality and morbidity amongst mentally ill patients.
In a study at South London and Maudsley (SLaM) NHS Foundation Trust, 20% of people with established psychosis had diabetes and a further 30% had pre-diabetes. In other words, the risk of type 2 diabetes in people with severe mental illness is approximately two-four times higher than in the general population.
As a result, the life expectancy of someone with a chronic mental health disorder is estimated to be ten years below average.
Researchers at Kings College London in collaboration with SLaM NHS Foundation Trust are seeking to address this problem in a new randomised cluster trial which is due to commence towards the end of this year.
The study is the first of its kind in a mental health setting and is focused on the application of digital tools to improve the physical wellbeing of mental health patients at the SLaM Trust.
As part of the study, two wards will have an electronic clinical decision support system (ECDSS) installed. Through a natural language processing tool called CogStack AI, the system helps clinicians to quickly identify mental health patients with symptoms of diabetes and direct them along the appropriate care pathway. CogStackAI is an open-source information retrieval and extraction platform which runs on Elastic.
An ECDSS is a computerised automated system designed to assist clinicians and other healthcare professionals in making clinical decisions.
These systems can be designed to automate electronic reminders and can refer to clinical guidelines to support the clinical decision-making process.
They are also able to condense large amounts of data. For example, if a patient has an electronic health record dating back several years, an ECDSS can automatically focus that data in a condensed form and pick out any relative data that could prove useful to the clinician.
In the context of this trial, CogStack AI can scour a patient’s medical history across the SLaM Trust’s medical records in a matter of minutes to identify incoming mental health patients who have abnormal blood sugars.
According to Clinical Research Fellow at King’s College London, Dr Dipen Patel, the system can perform manual coding and data collection tasks in less than a tenth of the time it takes a human analyst.
Through the use of a digital algorithm based on data provided by clinicians, the system automatically alerts doctors at key steps in the clinical process.
Patel says: “When we are looking at developing an electronic automated clinical decision support system, the first thing to do is work out what a human doctor should be doing.
“With regards to diabetes, we laid out a clinical pathway based on the existing guidelines. Once we had this clinical pathway in place, we then worked with CogStack to create a digital algorithm.
“This is where it is important to work out the points in the clinical pathway at which a computer should come in with an alert.
“For example, it could notify the doctor that a patient under their care hasn’t had a diabetes screening. Once a screening has been done, another alert could notify the clinician that the test has come back from the lab.”
The reasons for the mortality gap amongst mental health patients are varied and multifaceted.
Medication is a primary factor. Patel explains that, although they can be very effective for treating mental illness, many anti-psychotic medications can cause patients to develop metabolic side effects. These include a rise in sugar levels, cholesterol levels, as well as significant weight gain. Lifestyle choices, such as exercise, diet and smoking habits also come into play.
Over the past ten years, there has been a rapid rise in research interest in this field and Patel says improving the current healthcare system is key to addressing the issue.
He says: “We’re trying to work out how we can improve healthcare provisions for this group of patients. For example, trying to alter clinician behaviours, trying to improve our hospital systems and trying to improve the fragmentation that exists between primary care and secondary care.
“Often there is a blurred line between who is looking after the physical health of the patient. As patients tend to have more contact with their psychiatrist under secondary care, a lot of the physical health care is coordinated by mental health teams.
“However there’s a limit as to what can be done in mental health care compared to what can be done in a specialist setting for physical care.
“I think a system like CogStack, which has the ability to send automated alerts out to various end-users, can hopefully try and bridge that gap. The alerts that are sent out to secondary care can also be sent to the general practitioner to create a more rounded service for patients.”
Although electronic clinical decision support systems have been around since the early 2000s, they have not been widely used, especially within mental health services.
Early studies showed the potential for these systems to improve clinicians’ behaviours and proved that healthcare professionals were better at adhering to guidelines than they were without the system.
Patel explains that beyond the context of trials and studies, insufficient attention has been given to implementing these systems on a wider scale.
He says: “There wasn’t an appreciation for the barriers to scaling this technology up. So, particularly in mental health, there hasn’t been much use of this technology at all and the evidence base doesn’t really exist.”
Patel says the ultimate aim of the study is to show positive outcomes to scale up systems like this and then apply it to other conditions aside from diabetes, such as hypertension (blood pressure), high cholesterol, hepatitis screening and HIV screening.
If initial outcomes from the preliminary feasibility study are positive, the research team will move on to a larger trial in the future, testing the system on a greater number of wards.