Novel methods in data science to quantify viral and environmental triggers of chronic disease exacerbations

Lead Research Organisation: London Sch of Hygiene and Trop Medicine
Department Name: Epidemiology and Population Health

Abstract

People living with chronic diseases are at risk of early death, and suffer significant illness due to these diseases. Worsening periods of health, or "exacerbations" of asthma, chronic obstructive pulmonary disease, and coronary heart disease are costly and are bad for patients. Infection by a virus (like a cold, or 'flu) is known to trigger exacerbations, but there are other factors, like air pollution, and low or high temperatures that can also trigger them.

Despite knowing the importance of viruses, there is not good understanding of how circulation of viruses in the general population affects the timing of exacerbations at the level of a city or town. This also makes it hard to estimate the effect of other triggers, like air pollution, which is urgently needed.

This study will analyse the patterns of exacerbations, including respiratory virus circulation in the general population as well as non-infectious triggers, such as air pollution.

Technical Summary

This project focuses on exacerbations (periods of worsening) of three major diseases in the UK: asthma, chronic obstructive pulmonary disease, and coronary heart disease. People with these chronic diseases are at risk of early death, and suffer significant morbidity as a result of their conditions. Viral infections are known to trigger exacerbations from individual-level studies, but there is limited understanding of how circulation of triggering viruses impacts population-level patterns of exacerbations. A key tool for interrogating infectious disease dynamics - infectious disease modelling - has not been used to understand the role of infections in population-level patterns of chronic disease exacerbations. Current methods either inaccurately include infectious variables, or use statistical smoothing to remove their effect, leading to unreliable estimation of other triggers. Reliable estimation of the effect of risk factors is crucial for policymakers, and for patients. The accepted best tool for understanding infectious disease dynamics is transmission modelling. I aim to close this gap in our knowledge by using state of the art transmission modelling techniques to better quantify the burden of infectious disease on chronic conditions. By bringing together two disciplines and using novel methods, this project will strengthen the evidence base for possible interventions, such as vaccinations, preventive care, risk awareness, and treatment options.
 
Description Conference on Infectious Disease Dynamics
Geographic Reach National 
Policy Influence Type Influenced training of practitioners or researchers
Impact I started a conference (with 2 others) in 2017. After the pilot conference in 2017, the conference ran again in September 2018 during the award. It will run again in 2019 (also during the award). 105 scientists attended in 2018, to give presentations on research topics on infectious disease transmission dynamics and data scientific techniques in infectious disease research. The conference website is available at IDDconf.org, and the twitter account has nearly 500 followers: https://twitter.com/IDDconf through which increases engagement beyond attendees. For 2019 the conference is expanding to include 2 workshop training sessions for researchers and PhD students.
URL http://IDDconf.org
 
Description New short training course for postgraduate/professional
Geographic Reach National 
Policy Influence Type Influenced training of practitioners or researchers
Impact This is a new short training course called "Modern Techniques in Modelling" that will run in 2019. The course development was started in 2018 during the award. I am developing this course with 2 other members of staff at LSHTM. We will first run it internally in June 2019 for research degree students at our university, and then externally for research degree students and professional practitioners in November 2019. The course is 3 days on computational methods in infectious disease modelling and data science.
 
Description Outreach to Royal College of Nursing 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Professional Practitioners
Results and Impact A presentation/public engagement event was run at an outreach event at the Royal College of Nursing. The event was https://www.eventbrite.co.uk/e/rcn-late-beat-the-bugs-london-tickets-44405220209#

Unfortunately I could not attend the event (not available that day), but the presenters used a website that I made (with 2 others) and support the maintenance and development of: https://rozeggo.shinyapps.io/hpru_day/ as a key part of their outreach activities.

This aims to explain modelling methods, and also demonstrate impact of indirect effect of vaccination, using an example of influenza.

The event aimed to engage people in disease transmission processes, infectious diseases more broadly, and in the effect of vaccination.
Year(s) Of Engagement Activity 2018
URL https://rozeggo.shinyapps.io/hpru_day/
 
Description Speaking at TEDxThessaloniki 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact https://www.tedxthessaloniki.com/tedx_speaker/rosalind-eggo/
https://www.youtube.com/watch?v=GkLlqobr3PY

I was invited to give a public talk at a TEDx Event in Greece. It was a ticketed event with hundreds of participants. Also generated a YouTube video, which I will use for outreach in the future. This talk was about modelling in general, which is part of my project, but I did not talk about the work in the fellowship itself during the talk. This is because: i) at that time there were not results yet, ii) it was not in the requested title. But the methods from the fellowship are relevant to this talk, so it falls under engagement with modelling methods.
Year(s) Of Engagement Activity 2018