An agent-based model of the urgent care system in the
UK National Health Service
In the UK National Health Service clinicians and managers
have until recently approached modelling from quite different perspectives.
Managers, encouraged (or driven) by performance targets for indicators such
as waiting times, have focused their attention on aggregate measures of the
actual system, and used spreadsheet models or discrete event simulations to
explore the implications of proposed organisational changes. Clinicians,
focused on the experience of individual patients, distrust aggregate models
because these fail to engage with the highly contingent decisions they make
every day. Both groups distrust aggregate models because they are driven
by historical data that they know to be unreliable.
We have used a variety of workshop and paper-based approaches
that bring together people who provide and manage health and social care services
with people who use them, to engage in dialogue about how things are and
how they might be improved. This approach has proved to be extremely
powerful, Pratt et al, 1999, but suffers from two important limitations. The
first is that the ‘attention to the whole’ that is present in workshop situations
degrades after the workshop. The second is that the system is so complex
that people cannot understand the consequences, for one part of the system,
of actions taken in another part.
We demonstrate RESCUE, an agent-based model of the urgent
care system, which we have used to engage a wide range of people using and
working in the health and social care system in dialogue about ‘how things
really work around here’, Boyle & Jessopp, 2001. The focus on the
decision rules of the agents proves to be an exceptionally powerful way of
engaging clinicians, while the generation of aggregate data enriches the understanding
of managers.
The formalisation necessary for modelling reveals the
implicit knowledge and decision rules that are currently organising the delivery
of health care, and makes these available for dialogue. The construction
of the model produces an icon that we believe is more durable and powerful
than any we have used before, and one that can be shared with people not directly
involved in its construction.
It provides a robust way of enabling people to share
the decisions made in their part of the system in the knowledge that this
contributes to an understanding of the whole, thereby facilitating the shifts
of attention between the parts and the whole. And it provides the opportunity
for them to explore together ‘what-ifs’ in a way that gives them more
nuanced conversations about the consequences for other parts, and for the
whole.
Seán Boyle
London School of Economics
LSE Health & Social Care
rescue_uk@yahoo.com
Julian Pratt
London School of Economics
Urban Partnerships Group
Department of Operational Research
www.wholesystems.co.uk
julian@wholesystems.co.uk