Lauren (Lara) Johnson

Position

Enrichment Student

Cohort year

2023

Partner Institution

Bio

Lara is a PhD student at the University of Edinburgh working on an interdisciplinary project in applied data science and geriatric medicine. She is using statistical and machine learning methods on patient health records and survey data to study people’s risk of adverse health outcomes in later life. She is also interested in the usability and ethics of big data in health for researchers, data visualisation and communicating about research to lay members of the public.

Before starting her PhD she worked as a Senior Data Analyst for Public Health Scotland, where she produced and automated statistical reports and dashboards on prescribing data and evaluated quality improvement projects. Lara also worked as a Senior Manager in consulting, where she had project director responsibility for multi-country mixed methods research projects.

Lara has an Msc in Psychology of Individual Differences from the University of Edinburgh and a BA in Classics and English from the University of Oxford.

Research interests

Lara’s research relates to the Turing’s research challenge of revolutionising healthcare through maximising the information in patient health records and increasing collaboration between data scientists and medical researchers.

She is exploring what different data sources and methods can tell us about frailty, a loss of fitness that can occur as a result of natural ageing combined with the effects of multiple long-term medical conditions. Lara is studying how the number and combination of health issues people have relate to their risk of worse health outcomes like falls, mortality or unanticipated increases in care needs. She is looking at how this manifests differently in different age groups and in men vs. women (who are more frail but live longer). She is drawing on both electronic health records and longitudinal survey data.

During her time at ATI, Lara will work on the methodological challenges in applying machine learning techniques – which are designed for continuous variables and try to minimise variance - to binary and sparse data. The data she works with is largely binary and can take on only two possible states (each health issue can only be present or absent). Moreover, it is generally sparse (most health issues are present in only a minority of people). Being able to turn binary data into useful information is important for making use of health records.