SAIL Databank is excited to announce this year’s cohort of student projects as part of their 3rd year Capstone viva project. Their projects will benefit from the vast repository of health, social and administrative data held within SAIL and help nurture the data science knowledge and expertise amongst these budding medical and data research professionals.
As well as working with anonymised, real-world data, the students will gain experience in navigating the robust access procedures of a privacy-protecting Trusted Research Environment (TRE) to which all SAIL users are subject. In addition, the students will receive guidance and supervisory support from experts within Population Data Science at Swansea University Medical School, who recognise the need to grow the workforce capacity in this field.
Utilising linked health, social and administrative data is a complex undertaking that requires specific knowledge to maximise utility and efficiency. Population Data Science is committed to providing the environment and mentoring expertise to support the Economic and Social Research Council’s (ESRC) capacity-building strategy in partnership with Administrative Data Research (ADR) UK and Health Data Research (HDR) UK.
We asked this year’s student researchers to tell us about their projects and how SAIL can help answer their research questions:
The diagnosis and progression of the comorbidity of Severe Mental Illness (SMI) and Atherosclerotic Cardiovascular Disease (ASCVD) can be greatly affected by several risk factors, such as diabetes, obesity, and unhealthy lifestyle, which ultimately limit people’s ability to participate and engage in everyday activities. As there are many research gaps, the relationship of this comorbidity is not yet clear. Therefore, this data-linkage study aims to examine whether Welsh adult individuals with a diagnosed SMI will go on to develop an ASCVD event and to what extent they manage the comorbidity. Access to SAIL provided us with blood pressure and lipid level values and allowed us to statistically compare those values of individuals with and without comorbidity. Finally, this study can help identify the effects of comorbidity in clinical practice.
How likely an elderly person is to get diabetes in Wales if they are close to fast-food locations? I have chosen to answer this question because the population of Wales and the UK is ageing, and with a generation of people growing older who have been brought up on less healthy food, it’s important to know if the multitude of fast-food locations in Wales are having an impact on their health. The SAIL Databank will help answer this question by extracting the data of everyone in Wales over 65 who has type 2 diabetes and using that data, I will be able to make a map of their locations using a geographical information system and a Lower-layer Super Output Area (LSOA). Depending on the results of my project, it will be able to outline many issues with elderly people who have diabetes in our cities and rural areas, hopefully raising awareness and giving them further support from our healthcare system.
How have prescribing patterns for epilepsy medication changed in Wales? There are currently very limited studies on prescribing patterns for medications used to treat epilepsy [Anti-seizure medications (ASMs)] in Wales. It is important to know whether prescribing patterns in Wales have changed to reflect the latest guidance and evidence. I will use prescribing and diagnosis data from electronic primary care records within SAIL between 2010-2020. I will analyse data for the whole Welsh population, including the first ASM prescribed, sex, age, year of diagnosis and whether any other drugs were later co-prescribed. I will then use statistical techniques to see if any changes are significant. This will allow me to find out if the prescribing patterns for treating epilepsy have changed and, if so, why. If prescribing trends have not followed recent guidance or evidence, then this can be reported to healthcare professionals to ensure that people with epilepsy in Wales have the best possible treatment.
Predicting premature mortality in people with epilepsy and the associated risk of different antiepileptic drugs and polytherapy will be investigated using the SAIL Databank. People with epilepsy have a 2-3 times higher risk of dying prematurely than the general population, and males with epilepsy have been shown to have a higher risk of premature death. The project will use SAIL’s healthcare data to identify the most frequent risk factors for early mortality in people with epilepsy and build machine learning models to predict their premature death. Since SAIL’s healthcare data is routinely collected, this will allow us to incorporate lots of information, such as GP visits and hospitalisations, and the model developed will be portable and widely available for healthcare systems.
Diabetes and epilepsy are both common, long-term conditions that can greatly impact the quality of life. Diabetes is a metabolic disorder effecting the endocrine system, whilst epilepsy is a neurological condition characterised by recurrent epileptic seizures. It has been shown that there is a link between these two conditions, however, this relationship is not yet well understood. I plan to use data from the SAIL Databank to explore the relationship between diabetes and epilepsy in the Welsh population whilst also looking at factors such as age and deprivation levels. If certain subgroups of the population seem to be more at risk of receiving a dual diagnosis of epilepsy and diabetes, this could mean that prevention measures could be explored.
Commenting on the student projects programme, Associate Professor for Population Data Science Research, Ashley Akbari, said, “Enabling access to these types of data, software, systems and experiences will equip students with the skills and experiences they need to understand population-scale data, which is vital to support the future generations of research professionals who can advance society’s health and wellbeing. In addition to tackling health crises, more and more data-intensive organisations are realising the power of their data to improve lives, which when linked together, can provide a depth of insight into potential improvements to services and people’s lives. This initiative, supported by the Population Data Science group, provides students with the foundation and confidence to interrogate large-scale data to answer real-world problems.”
Head of Health Data Science in Swansea University’s Faculty of Medicine, Health and Life Science, Professor Mike Gravenor, added, “It is wonderful to see our students from a wide range of degree courses taking on data science projects. Using routine health data requires skills from many disciplines, and a SAIL project gives them the opportunity to see a world-class trusted research environment in action.”