Design of clinical pathway management system for preventing adverse event

A clinical pathway (CP) - also known as care map or integrated care pathway - is a collection of evidence-based recommendation on the sequence and timing of care interventions for a diagnosis. CP analysis has experienced increased attention over the years because it is useful both for the health-care management in general and for the administration, computerization and scheduling the best practice for individual patients. Indeed, significant improvement in the quality of almost all available clinical processes was observed after the CP development and implementation. In addition, reductions in health care macro-variation phenomena (length of stay, patient trajectories, etc.) and in performance micro-variation (variations in diagnostic and therapeutic prescriptions, protocol implementation, etc.) were shown in sites where CPs were implemented successfully. Moreover, CPs are helpulf for clinical risk management. For example, it is possible to discover a CP model from past clinical pathway instances (i.e., patient paths), detect the anomalies in clinical pathways and identify care points where patient trajectory deviate from expected and/or normative medical behaviors. Ideally a patient trajectory (PT), namely actual care events provided to a patient during diagnosis and treatment, should adhere to the corresponding CP. Deviation occurs when the care received by a patient differs from what was defined in the pathway. Differences are common in practice and they can have positive or negative implications on care delivery. In certain cases, deviations from CPs could be due to patient specific conditions or operational efficiency, but on the other hand they could indicate sub-optimal care, errors and potential adverse events. For this reason, it is important to measure variance from prescribed CPs because both these conditions could provide an opportunity for improvement in care delivery if early recognized. However, the measurement of deviations of trajectories representing the actual care received by patients is a non-trivial task because of the potentially complicated structure of clinical pathways themselves. CP management computerization can contribute to improve adherence to CPs and to measure deviations in a timely and efficient manner. Previous works demonstrated that an information technologhy (IT) application that support CPs -embedded into clinical routine work - can help to increase pathway compliance. The aim of this thesis is the design of a Clinical Pathway Management System (CPMS) able to track patient trajectory (PT) and identify deviations on actual PT from its corresponding CP and cases of missing records. Main expected benefits are an increased availability of information for healthcare staff, activities and responsabilities tracing along all phases of care path, greater clinical decision support for health workers and a better control over the process due to integration between services and resources. The ultimate goals are both the prevention of adverse events in clinical processes and a posteriori identificazion of errors to adopt corrective actions. Within the project, a methodology able to model and compare CPs and PTs has been developed, validated and integrated in the CPMS. The methodology consists of a set of models for the complete description of CPs and PTs and of an index for measuring deviations in PTs, called Clinical Pathway Deviation Index (CPDI). It can be applied for real-time evaluations, even when patient trajectory is still ongoing, in order to assess actual services - provided to a certain patient - and to generate alerts, if deviations have been found. In addition, it can be used by the Quality Manager to identify occurred adverse events, recognize an excessive variability among PTs, analyze trend and staff behavior and to identify possible improvements to be made to CP. CPMS was designed as multi-agent system (MAS) - due to the need to implement a scalable and flexible system, able to support different scenarios. The Prometheus methodology and the Prometheus Design Tool (PDT) were employed for MAS design. In addition, functional requirements were accurately described using the Unified Model Language (UML). A web application architecture was chosen in order to guarantee ease-of-installation, ease-of-deployment, ease-of-update, simplified access and scalability. A prototype was developed and tested.