Data relevant to health are increasing - in volume, variety and velocity. These ‘big’ or ‘diverse’ data are also become more available and more usable, including individual participant data (IPD) from trials, data obtained from electronic medical records (EMRs), administrative data and associated data linkage systems; as well as ‘~omics’ data (e.g. genomics, proteomics) and data from social networks, wearable and mobile devices.
Many argue that the increasing availability of these data will enable decision-making to draw increasingly on information that is closer to the circumstances of the individual. They predict that less value will be placed on summary measures or population estimates and greater emphasis on systems and data that describe smaller and smaller sub-groups approximating to each individual. Others are concerned about the unclear provenance of these data, risks of false positive findings, ability to assert causation and challenges in the transparency of data source.
This session will explore these issues with a series of presentations and a panel discussion. The presentation is relevant to all working in evidence synthesis (i.e. beyond health) and will highlight the importance to all in this field of engaging critically with debates around the use of diverse sources of data to inform decisions.
The five talks will be:
- Reflections on data, inference and causality – Ida Sim
- Using big data to learn about effects of interventions: holy grail or fool’s gold? – Jonathan Sterne
- Understanding prognosis in a world of diverse data – Karel Moons
- Big Data and NICE – opportunities, challenges and threats – Gillian Leng
The facilitated panel discussion will give participants the opportunity to pose questions to the panel, as well as allow further discussion on wider challenges and practical ways forward.