Analytics in Action
Dr Lisa McFetridge
My research develops statistical techniques for the modelling, analysis, forecasting and visualisation of real-world data. It ranges from empathic artificial intelligence, location planning for brands and city councils to a wide variety of medical applications, including nephrology, paediatric medicine, vascular disease and, more recently, analysing instances of Covid-19 in children.
Due to the synergy between patients' repeated measurements and the time until an event of interest, my work on longitudinal and survival analysis using joint modelling approaches naturally lends itself to the analysis of clinical data. Such methods allow a deeper understanding of the individual-specific reaction that patients can have to new treatments and thus is a key tool in the development of precision medicine, providing a personalised patient prognosis. This has led to collaborative projects with clinicians in Belfast City Hospital and the Royal Victoria Hospital, alongside various medical centres and hospitals throughout the United Kingdom.
In particular, during the Covid-19 pandemic, I am part of a UK collaboration to investigate the changing seroprevalence and symptomology of Covid-19 in children. This UK multicentre longitudinal study has the potential to inform public health decisions regarding school closures and other services vital to the well-being of children, with early results already indicating the need to widen the range of symptoms listed for children to better identify positive cases.
Another aspect of my research concentrates on the development of an innovative system to qualify the retail and foodservice potential of town centres and enable effective location planning through the use of discrete choice experiments and analysis. This research is part of an ongoing Knowledge Transfer Partnership with Client Analysis and Relationship Development (CARD) Group, a Belfast based market research company. Together, we are innovating in the area of brand analysis.
Alongside statistical methods, my research also focuses on machine learning techniques. Joint with the company Sensum, we recently completed a Knowledge Transfer Partnership to develop artificial intelligence models of felt emotion. This work has led to the development of a personalised mobility experience with empathic artificial intelligence and has been awarded an ‘Outstanding’ classification by Innovate UK.
The nature of statistical analysis allows the impact of my research to be felt in a wide variety of areas, both within academia and industry. This impact spans from the advancement of diagnostic tools and guidelines in medicine, enhanced driver safety through the understanding of driver emotions and possible need for intervention, to advising city councils in brand management for the development of sustainable and successful town centres.
The research I undertake as part of Knowledge Transfer Partnerships (KTPs) frequently leads to societal impacts through the ongoing work of the company utilising that research. The KTP joint with Sensum has led to the development of automated monitoring, enhancement and insights within cars. By understanding the stress, fatigue, focus etc. of drivers, the automated system provides enhanced safety and comfort to drivers through emotionally intelligent transportation. My KTP joint with CARD Group, despite being in the early stages of the project, has already gained interest from UK city councils who are keen to utilise the discrete choice models we are building to better understand the relationship people have with certain brands.
Examples of the impact my research has within medical diagnostics includes my collaboration with the Randox Laboratories Centre of Excellence for Biomedical Applications, for whom I provide statistical expertise in the development of advanced diagnostic tools in medicine. Through my collaboration with Wellcome Wolfson Institute of Experimental Medicine, my work influences current practices in paediatric medicine. My investigation into the performance of existing clinical practice guidelines (CPG) for diagnosis of invasive meningococcal disease in children has highlighted that current NICE guidance performs poorly when compared to alternative CPGs. As a result, we are now in the process of developing a new CPG to reduce the number of unnecessary invasive procedures on children. This will increase safety and consequently reduce financial cost for the health service.
Due to the limited knowledge of prevalence and symptomology of Covid-19 in children, a UK collaboration was established to investigate Covid-19 in younger people. Initial results are already being fed into the discussion around the symptom list for children and, of particular note, is the high level of asymptomatic cases within a younger population.
- EPSRC grant entitled “Time-dependent Robust Joint Modelling: Analysing a wealth of longitudinal outliers” Total value: £100,419 - (https://gtr.ukri.org/projects?ref=EP%2FP026028%2F1)
- Randox Laboratories Centre of Excellence for Biomedical Applications grant. Total value: £479,768 - (https://www.randox.com/50-million-centres-excellence/)
- KTP joint with CARD Group. Total value: £112,236 - (https://info.ktponline.org.uk/action/details/partnership.aspx?id=11886)
- KTP joint with Sensum. Total value: £96,821 - (https://info.ktponline.org.uk/action/details/partnership.aspx?id=10774)
- Dr Thomas Waterfield, Wellcome Wolfson Institute of Experimental Medicine - https://pure.qub.ac.uk/en/persons/tom-waterfield
- Client Analysis and Relationship Development (CARD) Group - https://www.card-group.com/
- Sensum - https://sensum.co/uk
- Randox Laboratories Centre of Excellence for Biomedical Applications - https://www.randox.com/50-million-centres-excellence/
- Statistical Advisor for HSC Statistical and Methodological Support Unit - https://research.hscni.net/hsc-statistical-and-methodological-support
- QUB/Belfast Trust Paediatric Sepsis Research Group
- Özgür Asar (Acibadem University, Turkey) - https://www.acibadem.edu.tr/en/academician/ozgur-asar
- Jonas Wallin (Lund University, Sweden) - http://www.lunduniversity.lu.se/lucat/user/b7c33e09e7abf3f4bc36682616d87765