Education and training and increasing capabilities and capacity in big data and applications in health and care were focus areas for GM CHC.
Four applied doctoral research projects were initiated in 2017/18 supporting the clinical care pathways as well as the infrastructure required to increase access to health care data.
Two examples of applied research projects
1. Commercial organisations accessing health data for research purposes
Given public concern about commercial organisations accessing health data for research purposes, there is a need to convey what patients, public and citizens would expect as part of rules for data access.
This research builds on the safeguards devised by citizens’ juries and uses the existing research literature to identify what are acceptable and unacceptable options for citizens. Focus groups were held to find out which options for commercial access to health data in the UK are preferable.
Future outputs will include the development of a social licence for commercial organisations accessing health data for research. Social licences outline the expectations of society regarding the conduct and activities of companies and organisations that go beyond the requirements of formal regulation. The recommendations can then be used in future projects where public and private organisations share health data for research purposes.
2. Dynamic prediction modelling in a learning health system
Clinical prediction models (CPMs) are used to predict future outcomes for individuals, and thus have the potential to be used for effective targeting of resource.
This project is developing methods within the dynamic modelling framework to help better predict patient outcomes and risks within healthcare.
Starting with systematic review of dynamic prediction models will obtain an understanding of the current methods, different approaches will then be used to develop a prediction model for 30-day mortality following coronary heart disease intervention and compared to a current CPM used in clinical practice.
Future outputs will include new methods to address some of the challenges with dynamic prediction models with potential impact illustrated through simulation studies.
