Edinburgh Futures Institute regularly supports data-driven research by providing small grants to successful applicants. In 2021, Dr Glenna Nightingale, Chancellors Fellow in Nursing Studies at the School of Health in Social Sciences, received a small grant for a study about the spread of Covid-19 across the University of Edinburgh’s Halls of Residence.
While media and political attention to Covid-19 has decreased, the need for continued research remains. Data analysis and statistical modeling, for instance, can help in understanding Covid-19 infection dynamics, enabling public health professionals and policymakers better prepare for future pandemics.
In a recently published journal article, Dr Glenna Nightingale and a team of clinical and academic researchers from NHS (National Health Service) Lothian, analysed the transmission of Covid-19 among University of Edinburgh students living in the University’s Halls of Residence. In the article, they proposed what they argued to be a relatively new approach to understanding how and why Covid-19 spread within this particular space, time, and group of people.
According to Dr Nightingale, epidemiological research at the intersection of public health and statistics often focuses on exploring temporal dynamics of the spread of diseases. For instance, many of us recall the daily reporting of the “R number” (basic reproduction) during the Covid-19 pandemic, which indicates how fast a disease spreads within a population. In contrast, Dr Nightingale argues, less attention is paid to how infectious diseases spread across space.
In their study, Dr Nightingale worked with her team to develop a combination of spatiotemporal epidemiological models to understand Covid-19 trends within University residences, over a particular period of time and space. To this end, the authors rely on a particular group of statistical models called “Log Gaussian Cox Process Models.” Log Gaussian Cox Process Models allow for the spatiotemporal analysis of individual data point patterns. The researchers used three new indices – Rspatial, Rspatiotemporal and Rscaling – which provided a more detailed picture of epidemiological dynamics within the study’s selected setting and period of study.
In their study, Dr Nightingale and her co-authors demonstrate the usefulness of this combined spatiotemporal approach. For instance, the statistical modeling reveals that between September and December 2020, relatively high infection rates were found among University of Edinburgh residence halls for which more fines were issued and which were non-ensuite. Moreover, the authors found that there was no significant relationship between levels of reported infections within the residence halls and community levels of Covid-19.
Dr Nightingale and her co-authors are hopeful that these insights can help the University administration and public health officials more broadly prepare for future public health emergencies. They also argue that the application of Log Gaussian Cox Process Models to epidemiological research presents a promising new avenue for research. These models can help provide a more fine-grained understanding of how infectious diseases spread both across time and space by taking into account the circumstances of individual patients, such as their living situation. The article’s first author, Megan Laxton, stresses the broader impact of their research:
“There are so many potential epidemiological applications for this type of modelling – it is really exciting to demonstrate the use of these models in understanding the spatiotemporal spread of disease.”
Dr Nightingale said that their study on Covid-19 is only the beginning of continued research on the subject:
“We plan to extend this study to include other universities to gain a better understanding and to allow us to compare the proposed indexes in different contexts. Additionally, we are in the initial stages of embarking on a project which proposes to use Log Gaussian Cox Process models in modeling Dengue and Zika.”