Towards knowledge sharing and patient privacy in a clinical decision support system

Patient records and their disease and treatment history can be scattered among healthcare providers. Sharing the knowledge effectively and, at the same time, respecting patient privacy is crucial in providing safe and accurate clinical decision support systems (CDSSs). In this paper we reflect upon our experience in the HealthAgents project wherein a prototype system was developed and a novel approach employed that supports data transfer and decision making in human brain tumour diagnosis. Here we examine the capability of the lightweight coordination calculus (LCC), a process calculus-based language, in combining together distributed healthcare services and meeting security challenges in pervasive settings. The result is that various clinical specialisms, being captured in representational abstractions and making contribution to patient diagnosis and management, retain their autonomy. However, at the same time, the behaviour of specialists in sharing clinical knowledge about their patients and providing clinical support is constrained by policies and rules in respect of their own clinical duties and responsibilities. Being introduced into the programme of the HRB Centre for Primary Care Research, this novel approach has the potential to help the provision of optimal solutions in data linkage and sharing across the primary and cecondary care interface. As added value, its application also advances the process of integrating clinical prediction rules and implementing CDSSs in practice and, ultimately, the improvement of quality of car