The Turing-Roche Knowledge Share Series: Diversifying Health Research
online
Hosted by the Turing Institute
This event will be on the theme of diversifying health research. As AI in health research progresses, how do we ensure representativeness across the system and the inclusion of all voices?
We will be hearing from Spencer James, Principal Data Scientist at Roche and Lara Johnson, PhD student at the University of Edinburgh and Turing Enrichment Student.
Spencer's talk will explore inclusive research, which is increasingly a priority at Roche and Genentech as well as across the broader biomedical research industry. Inclusive research includes assessing whether study populations in clinical trials are representative of the underlying population with the disease of interest. Historically representativeness in this sense has been affected by population-level variation in access to care as well as diagnosis, treatment, and enrolment in clinical trials. Improving representativeness in this setting is predicated on our ability to measure it. Surprisingly a standardized metric for measuring this concept in the setting of drug development does not seem to exist. Currently different organizations, publications, and investigators tend to report different combinations of error metrics, statistical methods, and qualitative assessments as proxies for the concept of representativeness, particularly as it pertains to variables including race, ethnicity, gender, and others. In this talk, he will introduce a novel, standardized metric for measuring representativeness in demographic groups for the purpose of advancing inclusive research and designing representative clinical trials at Roche and elsewhere.
Lara will talk about her experience involving patients and clinicians in her PhD project "What do data tell us about frailty?", which uses ML methods to look for patterns in the health deficits older people have and assesses their ability to predict adverse outcomes (quality of life, care needs, falls, hospitalisation, deaths, etc.). This is a complex topic emotionally, medically and technically from a data science perspective. During her talk, Lara will highlight how she has embedded Patient and Public Involvement (PPI) through her project. Her collaborative efforts have led to the development of an accessible information resource on ageing and frailty, which is currently in use within NHS Lothian. Lara will also share how patients and clinicians have played a pivotal role in shaping the project by discussing research questions, methodology and exploring the implications of preliminary research findings.