The Power of Connections

Mapping the Behaviour of Health Care Networks

Dr Jonathan Clarke
Supervised by Professor Ara Darzi (IGHI, Surgery and Cancer), Professor Mauricio Barahona (CMPH), Dr Joachim Marti (Center for Health Policy)

This work is done in collaboration with the Big Data and Analytics Unit (BDAU) at Imperial. Please see here for more information. 

No hospital functions in isolation. Patients flow from one care provider to another creating a complex network of interconnected entities that together form the National Health Service. The performance of one component of this network influences the demand placed on adjacent care providers, and in turn may significantly change the quality of care provided to communities as a whole.

The national performance of Accident and Emergency departments over recent winters demonstrates that hospitals have a finite capacity to provide efficient care. If demand exceeds this limitation, delays and a deterioration in the quality and safety of care may result. This relationship is found across healthcare, in cancer care, surgical waiting lists and inpatient care. Predicting the circumstances in which such systems fail at a local or national level relies upon an understanding of how providers are connected to one another, the demands patients place upon providers for care, and the capacity of providers to meet the demands placed upon them.

This work seeks to examine how patients move from their local communities through primary and secondary care providers, and thereby develop a rich understanding of the dynamic structure of health care at a national level.

Three primary studies will be conducted:

• Patient choice will be examined through the decisions individuals make when presented with a choice of proximate care providers, and the influence such decisions have on system demand and quality of care.
• Novel community detection methodologies will be applied to optimally partition the health system into scalable organisational subgroups and specialty networks.
• Simulation of patient flows through the health network will identify predictable vulnerabilities in the health system and indicate areas for efficient focussed resource allocation to increase the resilience of the network.