Projects Delivered

Our understanding of life after brain tumour diagnosis is limited. Healthcare professionals often see brain tumour patients at the time of investigation or treatment, therefore only seeing snapshots of a patient’s life. Accurate description of life after brain tumour diagnosis is important to detect behavioural patterns that may be associated with deterioration and to inform research studies. The Brain Tumour Charity developed an online app accessible by computers and mobile phones called BRIAN that allows patients and their carers to share their information on events and quality of life over time. This provides a unique opportunity to study how aspects of life changes and how these changes may correlate with clinical events. This study will summarise the characteristics of people submitting data through the BRIAN app and use the data to describe the quality of life over time. We will also examine how the journey after brain tumour diagnosis may differ depending on the tumour type and treatment received. Findings from this study will help clinicians, patients and their carers to better understand and prepare for life after brain tumours diagnosis, and our results will generate more research questions to improve brain tumour experience.

People in the UK are living longer, but many older people have multiple health problems. The NHS and local authorities do not always provide joined-up or effective care to older people. Ensuring that care is targeted at those with the highest need is difficult. Ideally, we would provide preventive care to maximise people’s future independence, but that means we need to be able to identify the people most likely to lose their independence.

The aim of this project is to develop and test tools for predicting who is most likely to lose their independence or have other adverse outcomes. We will do this using information recorded electronically in hospital records. In future we would like to add in GP and local authority records to add more depth to our analysis. All the information we use will be pseudonymised (i.e. have names and similar information removed), kept secure and will not be used in individual patient care. The prediction tools we develop in this project will help us target care to those who most need it.

People in the UK are living longer, and more people now live with multiple conditions. Healthcare focuses on single diseases, and is often not good at dealing with many of the common problems of later life. For example, there are many reasons that people fall or get confused or have problems with their bowels or bladder. In electronic health records, these problems are almost always documented in ‘free-text’ (ie the typed notes that doctors or nurses make). Identifying who has this type of problem is therefore difficult because different doctors or nurses will use different words to describe them (‘confused’, ‘mixed-up’, ‘delirious’, ‘not themselves’, ‘cognitively impaired’). We will use data science methods that can automatically turn free-text into carefully-defined categories (eg all of the words in the last brackets are ‘delirium’). This will support research to understand how common these problems are and what causes them, and healthcare improvements.

Acute heart failure is a life-threatening condition where the heart is suddenly unable to pump blood around the body. It can be challenging to diagnose because the symptoms often mimic other conditions. Previous studies have showed that delays in making the correct diagnosis result in worse outcomes. We therefore developed a decision-support tool called CoDE-HF that uses a computer algorithm to combine levels of a blood test called NT-proBNP with patient factors to calculate the probability of acute heart failure for an individual.

In this project, we wish to evaluate the performance of CoDE-HF in approximately 2,000 patients attending the Emergency Department with suspected acute heart failure. We will store surplus material from their blood tests to measure NT-proBNP and link information from their electronic health records with other routinely collected medical information in regional and national databases in order to evaluate this algorithm.

Almost 10% of the global population has diabetes and rising; 15-17% of all UK hospital (inpatient) beds are occupied by people with diabetes at any time, costing ~£2.5 billion/ year. Significant numbers suffer a deterioration in diabetes care (such as low or high blood glucose events, and preventable foot ulcers) during hospital stays, as a result of poor management, in part due to general hospital staff being poorly trained in diabetes management.

This proposal aims to develop a computer-based (digital) tool to support clinicians in hospitals make better decisions. The tool will identify ‘at risk’ individuals; support triage of patients, and give advice around the correct decisions for foot care and medication changes/ titration. This tool makes use of historical information (data) contained in medical records to find patterns in the data that predict when escalation of treatment or expert input is needed, ahead of time, thus preventing serious health outcomes/ medical emergencies/ reducing hospital stay, preventing ill health and death. This project is a collaboration between MyWay Digital Health (MWDH), a University of Dundee spin-out, NHS staff and academia.

COVID-19 can be identified by looking for changes in the lungs on chest x-rays and computed tomography (CT) scans. Different imaging features may be associated with different patterns of disease. In addition, radiological imaging can identify features of other diseases that affect the heart and the lungs. These other diseases may affect the outcomes of patients with COVID-19. This study will review the radiological imaging of patients with COVID-19 and identify features of heart and lung disease which can be used to find patients who do better or worse after COVID-19 infection. This will help us identify the overlap between COVID-19 and other diseases affecting the heart and the lungs, and may help identify groups of patients who are at an increased risk.

In late April 2020, a change in COVID-19 testing strategy was implemented in NHS Lothian after a Scottish Government instruction to test all patients over 70 years old being admitted to hospital irrespective of displaying symptoms or not. This project sought to provide an initial analysis on what this change in testing criteria meant for the numbers of tests being undertaken for those in different age categories.

Chronic obstructive pulmonary disease (COPD) is a common lung condition, often caused by smoking. Patients living with COPD may suffer from shortness of breath and this can worsen unpredictably in flare-ups known as exacerbations. These often require hospital care. COPD is the commonest cause of emergency attendance to the hospital with breathlessness, and the third commonest cause of death worldwide. We plan to use health data from deidentified people with COPD to find risk factors for these exacerbations and other harmful outcomes including death. This will include using machine learning techniques, where advanced computers look for patterns in records that might otherwise be missed. Our aim is to create a new prediction tool that could be used to target care to patients identified at risk of deterioration. This is being delivered by Lenus Health in partnership with NHS Lothian to quickly move developments from this project into patient care.