NHS England Data Science PhD Internships

Optimisation and Learning algorithms on patient pathways

Keywords: Synthetic, QLearning, TabularData

Need: The development of robust pathway analysis would enable a whole host of insights into patient experience and impact of interventions across the NHS. SyntheaTM is an open-source, synthetic patient generator that models the medical history of synthetic US patients. Using the Synthea pathways as a starting point, this project would seek to demonstrate the potential for learning algorithms to be applied to a pathway to show how the impact of a pathway change could be modelled across multiple criteria.

Current Knowledge/Examples & Possible Techniques/Approaches: Details of Synthea can be found on their Github and io site: Synthea by the Standard Health Record Collaborative. The MITRE team have started to address the issue of localization of Synthea for other areas and in particular have created a worked example for Shropshire as a demonstration of potential. One of our previous projects investigated different learning algorithms for pathway analysis including stochastic gradient descent, reinforcement learning (q-learning), monte carlo tree search and A* search.

Related Previous Internship Projects: SynPath_Diabetes

Enables Future Work: Pathway analysis, use of synthea, generation of agent-based simulations with learning for healthcare

Outcome/Learning Objectives:

Datasets: n/a

Desired skill set: When applying please highlight any experience around work with pathways or longitudinal data, software development, coding experience (including any coding in the open), and any other data science experience you feel relevant.

Return to list of all available projects.