Principal Investigator: Dr Safa Salim
Institution: Imperial College LondonTags: 61217, chronic disease, epidemiology, Machine Learning, risk factors, risk prediction, venous ulceration
Veins are blood vessels in the legs that return blood back up to the heart. When these veins don’t work well they can become enlarged, cause swellings and ulcers in the legs (Venous Ulcers). Venous ulcers are estimated to affect 1 in 100 people, and can cause significant distress to the people who have them. It is also estimated to cost the health service approximately £2.7 billion pounds every year. We will identify people in the UK Biobank with venous ulceration and describe what this population look like e.g. the average age, which regions they are in and whether they have other healthcare problems. This can help us understand what the healthcare needs of this group of people are.
It is unclear about why some people get venous ulcers and why some people don’t. However, it is thought to be due to several factors relating to each individual. This includes factors like age, gender and lifestyle choices. Before people develop venous ulcers, they get other signs in their legs such as swelling and skin changes. People with these changes can be treated to help prevent a venous ulcer developing. Using large datasets, like UK Biobank, can help us understand what these factors might be which will ultimately help to identify who is at higher risk and should be treated urgently.
There are several ways that we can use the data to try and discover what these factors might be, one of these ways is using artificial intelligence (AI) techniques such as Machine Learning. Machine learning can process large amounts of data, like from UK Biobank, to try and discover the factors that lead to some people getting venous ulcers. One of the benefits of using machine learning to do this is that it may help identify risk factors that are not already known about.
Therefore the aims of our study are to:
1- Describe characteristics of patients with venous ulcers in the UK biobank database
2- Identify factors that are associated with venous ulcers
3- Use Machine Learning to analyse large amounts of data to find factors that we might not already know about.
We estimate that this will take us approximately a year to do and we will share all of our findings with patients affected with venous ulcers, the general public and the scientific community to try and help more research in the area.