Abstract. Translating clinical guidelines into formal models is beneficial in many ways, but expensive. The progress in medical knowledge requires clinical guidelines to be updated at relatively short intervals, leading to the term living guideline. This causes potentially expensive, frequent updates of the corresponding formal models. When performing these updates, there are two goals: The modelling effort must be minimised and the links between the original document and the formal model must be maintained. In this paper, we describe our solution, using tools and techniques developed during the Protocure II project 1. 1

Summary Objective: Transforming narrative guidelines into a computer-interpretable formalism is still a bottleneck in the development of decision-support systems. Our goal was to support this step by providing computer-interpretable templates for representing guideline knowledge using clinical abstractions that are appropriate for particular guideline sub-domains. Methods and materials: We analyzed guidelines taken from the sub-domains of screening and immunization guidelines to find repeatable clinical abstractions and structured them as design templates to support encoding of these guidelines in a computer-interpretable format. To find guidelines for analysis and validation, we Results: We developed two visual templates that structure screening guidelines as algorithms of guideline steps used for screening and data collection and used them to represent the guidelines collected. We validated the computability of the screening templates by executing a screening guideline in a workflow engine. We validated the computability of immunization templates by writing code that, based on represented knowledge, computes immunization due dates and by creating an algorithm that translates the knowledge into computer-interpretable guidelines. Conclusion: We have demonstrated that our templates could be effectively applied to screening and immunization guidelines to produce computer-interpretable representations using domain-level abstractions.

A model is an abstraction of some "thing" (e.g., object, system, process, phenomenon) in our world that we create in order to understand it better. Models can be physical (e.g., an architectural model of a building, a prototype of a user interface), mathematical (e.g., a model for predicting the weather, for estimating population growth), or conceptual (e.g., a clinical algorithm, the logical relationships between data items in an EMR). Models can be used to describe an existing complex real-world object or phenomenon, or they can be used as a vehicle to design a manmade object or system. While conceptual models can also be specified in narrative, in this discussion we address conceptual models that have a symbolic representation with a diagrammatic notation. Thus, conceptual models specify objects or processes, their properties, and their relationships. A conceptual model of a proposed process or system has several important benefits, two of which are of great importance. First, the process of creating a conceptual model of a system helps its designer to study the problem domain better, to understand the system's components and their relationships, the system's desired functionality and behavior, and its interaction with users and other systems. Second, a conceptual model can facilitate the communication between different stakeholders of the process (or system), including for example, customers, end-users, (medical) domain experts, system analysts, and software developers; all stakeholders have different expectations from the system. Using a conceptual model is one of the ways by which we can narrow the design-reality gaps [1] between the conceptions of the system by its different stake holders. Conceptual modeling can play important roles in the development life-cycle of health information systems (HIS, e.g., electronic medical record (EMR) systems, computerized physician order-entry systems (CPOE), and clinical decision support