One of a series of knowledge share series aiming to bring together members of Roche (a large biotech company, and leading provider of in-vitro diagnostics and innovative solutions across major disease areas) and The Alan Turing Institute’s networks (the UK’s national institute for data science and artificial intelligence) as well as the wider scientific community, to showcase partnership updates and research, knowledge share and hear different academic and industry perspectives on data science topics to gain insight and help build new connections and collaborations.
The event was on the theme of Digital Health, exploring how the increasing amount of collected ‘footprint’ data can be used to develop healthcare research and products and considerations around this. The event reached to around 150 people with 80-90 viewing live and 75 views on YouTube to date.
I gave a talk introducing the Turing Special Interest Group (SIG) Novel Data Linkages for Health and Wellbeing and how the groups work applies to my own research in using AI in population health surveillance through digital footprint data. Jian Dai and Yajing Zhu both experienced data scientists from Roche, gave an overview of the Data, Analytics, and Imaging Data Science Capabilities in Roche’s personalized healthcare department. After we took part in a panel discussion.
I addressed how three themes emerging from the work SIG is doing bringing multidisciplinary and multi sector communities together to discuss linking novel digital footprint data to health outcomes applied to my own work. These themes were:
• The value to policy makers & healthcare organisations
• Public acceptability
• Industry as data providers
I demonstrated how they applied to my studies Assessing the value of integrating national longitudinal shopping data into respiratory disease forecasting models, Public attitudes towards sharing loyalty card data for academic health research: a qualitative study and a third study which I am currently writing up Using shopping data to predict respiratory disease and COVID-19: CIDS (Covid Individual Data Study).
My presentation generated a lot of interest and a lot of questions, many of which had to be followed up in the Roche-Turing slack channel post-event due to time restraints. There was more discussion around the inclusion of wider audiences in medical research, data collection and wrangling, and public acceptance of using these ‘novel’ data types than there was around modelling techniques. May be this is because in ‘real-world’ AI health care applications, appropriate data collection and the representativeness of the data can prove to be more challenging then the computational analysis itself.
You can watch the webinar here