TY - JOUR AU - Bucur, Anca AU - van Leeuwen, Jasper AU - Vdovjak, Richard PY - 2016/09/12 TI - Efficient pathology case distribution leveraging workflow and domain modeling, and flexible definition of policies JF - Diagnostic Pathology; Vol 2 No 1 (2016): 2016DO - 10.17629/www.diagnosticpathology.eu-2016-2:115 KW - Case distribution, digital pathology, workflow modeling and optimization, automatic dispatching, workflow engine, rules engine N2 - Background The value of digital pathology and its potential to improve the current pathology practice are increasingly recognized, with a growing number of examples of successful implementation in a variety of use cases. While current widely-spread uses relate to research, consultation, second opinion and education, success stories are emerging demonstrating outcome and cost benefits of leveraging digital pathology for standard diagnosis as well. In all these cases optimization of pathology workflows, standardization and seamless integration in the environment are emphasized as prerequisites to reaching the desired improvements. Aims We develop workflow-driven applications to enable clinical users to efficiently and effectively leverage deployed digital pathology solutions for faster diagnosis and better patient outcomes. The work addresses information integration requirements, and aims to identify and propose solutions for performance bottlenecks in existing processes. A process with potential for improvement is the case distribution to pathologists for diagnosis. Materials and Methods We propose an approach to workflow modeling and implementation that is open and flexible, and leverages existing standards (BPMN2.0) developed in the context of business process modeling and widely adopted in other domains. This enables efficient implementation and give us access to an existing platform (KIE) that delivers a workflow-engine (jBPM), a rule-engine (Drools), and a constraint satisfaction solver (OptaPlanner). By externalizing the definitions and implementations of tasks within the workflows, we build workflows that are adaptive to the environment and that can incorporate decision models and applications of the desired complexity. We selected case distribution for diagnosis as the first process to optimize and built an application that can be efficiently customized to the needs of each lab with respect to dispatching rules, operational domain model, optimization goals and visual elements. Results We implemented a case distribution application that can be deployed within an integrated workflow implementation to optimize the distribution of cases to pathologists for diagnosis. To automatically assign cases to pathologists according to defined policies, the optimization component applies the defined policy model (scoring rules within the domain model). The schedules are generated according to the desired optimization goals, e.g. to improve throughput or turnaround. This application seamlessly connects with relevant systems in the environment, the LIMS and the IMS, in a loosely-coupled standard way through the integration engine to retrieve all needed data about the cases to be visualized and used in the dispatching decisions, and to provide the case assignation information to the IMS. Conclusions The optimization of the case distribution to pathologists for diagnosis has been identified in the literature as an important area when aiming at workflow improvements and automation of the processes in the pathology labs. A common theme emphasized by previous research identifying user needs and requirements for efficient digital pathology implementation is the use of standards to support adoption and the seamless integration with all relevant systems in the environment. Within a standards-based workflow-driven approach, we implement a case distribution application that optimizes the case assignation for diagnosis by modeling the local workflows and the requirements of each deployment site. The optimization component applies the local policy models (i.e. business rules, domain models with roles and characteristics, and optimization goals) to propose the most suitable solution for distributing the incoming case load among the available pathologists for diagnosis. UR - https://www.diagnosticpathology.eu/content/index.php/dpath/article/view/115