Coronavirus (Covid-19): Latest updates and information
Skip to main content Skip to navigation

A Conception of Computing with Promise for Medical Education

Meurig Beynon and Will Beynon

Computer Science, University of Warwick, Coventry, UK
University Hospitals Leicester, Leicester, UK


Abstract

Introduction: Advances in computing promise to revolutionise medical reasoning. For instance, the virtual physiological human (VPH) initiative aims at ultimately developing virtual clones for personalised and predictive healthcare, and the OpenClinical project at providing AI decision-support tools to enable healthcare professionals to deliver evidenced-based care accurately reflecting the ever-changing wealth of current medical knowledge. Both the VPH and OpenClinical draw inspiration from approaches to developing complex software systems within the established conceptual framework for computing. We argue that this 'computational thinking' framework is not best-matched to the issues they raise, and that an alternative perspective is essential to fully reflect the qualities of medicine as both art and science. Furthermore, non-traditional approaches to computing potentially provide adaptive and expressive tools for creating distinctive applications well-suited to medical educational objectives.

Objective: Computing advances have such potential impact on medical decision making that understanding how computing and medicine can best be integrated both in theory and practice is critical. We aim to raise awareness of ways of applying computing in medical education that respect the exceptionally broad nature of medicine as a discipline. By adopting too narrow a view of computing, we are in danger of construing the VPH and OpenClinical as respectively reducing the study of medicine to a grand challenge in computational science or to a sophisticated form of protocol-driven interaction. We instead propose a conceptual framework for computing - "Empirical Modelling (EM)" - compatible with the aspirations of the VPH and OpenClinical projects, but giving greater consideration to the essential role for human intelligence and judgement. In this way, we illuminate what separates the medical student both from the robot and from the experienced doctor.

Method: We introduce and illustrate EM as an approach to computing based on ‘making construals’ rather than ‘developing computer programs’. Construals are physical artefacts - typically built most effectively with computers - that 'metaphorically' express expectations about relationships in the world. The central idea is that we make sense of phenomena by thinking about putative causes - what is acting in the situation to make changes (agents); what these actors are deemed to 'sense' and respond to (observables); and how agents' state-changing actions immediately affect several observables simultaneously in predictable ways (dependencies). We have constructed a simple online construal relating to physiological and clinical aspects of vivax malaria to demonstrate the key notions of agency, observation and dependency in a tool applicable to medical education.

Result: The development and interpretation of our interactive online construal is a potentially open-ended activity in which medical students, clinicians and researchers can all participate. Each participant interacts to change the state of the construal in essentially the same way - by introducing new observables or redefining existing observables - according to their perspective and experience. The construal serves – inter alia - to model the evolving states in the lifecycle of a malaria parasite, to show how symptoms may be correlated with these states, and to highlight gaps in our explanations for state-transitions. To the non-expert, it serves as a 'provocative model' that, rather than presenting definitive textbook answers, promotes questions, providing stimuli and bases for independent research or self-directed learning.

Conclusion: Unlike traditional conceptions of computing, EM emphasises activities where there is limited support from theory and the mechanisms operating must be explored speculatively. This makes it an ideal vehicle for developing medical experience where learning does not depend on knowledge alone or where uncertainties are central to the problems encountered. Similarly, EM provides an alternative platform for the vision for systems biology that underpins the VPH and the pragmatic decision-support afforded in applications such as OpenClinical.


Keywords: Medical science, medical education, clinical decision support, computational science, Empirical Modelling, construals