GESDOR - a generic execution model for sharing of computer-interpretable clinical practice guidelines.

GESDOR – A Generic Execution Model for Sharing of Computer-Interpretable Clinical Practice Guidelines Dongwen Wang, PhDa; Mor Peleg, PhDb; Davis Bu, MDa; Michael Cantor, MDa; Giora Landesberg, MD, DScb; Eitan Lunenfeld, MDb; Samson W. Tu, MSb; Gail E. Kaiser, PhDc; George Hripcsak, MD, MSa; Vimla L. Patel, PhD, DSca; Edward H. Shortliffe, MD, PhDa a Department of Medical Informatics, Columbia University, New York, NY 10032 b Stanford Medical Informatics, Stanford University, Stanford, CA 94305 c Department of Computer Science, Columbia University, New York, NY 10027 We developed the Guideline Execution by Semantic Decomposition of Representation (GESDOR) model to share guidelines encoded in different formats at the execution level. For this purpose, we extracted a set of generalized guideline execution tasks from the existing guideline representation models. We then created the mappings between specific guideline representation models and the set of the common guideline execution tasks. Finally, we developed a generic task-scheduling model to harmonize the existing approaches to guideline task scheduling. The evaluation has shown that the GESDOR model can be used for the effective execution of guidelines encoded in different formats, and thus realizes guideline sharing at the execution level. INTRODUCTION Sharing of computer-interpretable clinical practice guidelines (CPGs) is a critical requirement for guideline development, dissemination and implementation1. In addition to conferring cost efficiency in guideline development, guideline sharing leads to improved acceptance of guideline implementation systems, and thus promotes the use of guidelines2. One approach to guideline sharing is to develop a universal standard for guideline representation to encode all the guidelines. Considering that no existing guideline representation model is dominant over the others, this approach is impractical at present. In this study, we propose an alternative approach, the Guideline Execution by Semantic Decomposition of Representation (GESDOR) model, to guideline sharing at the execution level. This approach is based on the observation that the different guideline representation models contain similar execution tasks, which are used to support the implementation of CPGs. According to the GESDOR model, guidelines can be encoded in different formats. A set of generalized guideline execution tasks are extracted from the existing guideline representation models. This set of generalized guideline execution tasks is then used to drive the execution of specific guidelines encoded in different formats. The relationship among the guideline instances, the guideline representation models in which the guideline instances are encoded, and the generalized guideline execution tasks is shown in Figure 1. Generalized Guideline Execution Tasks Guideline Models GLIF PROforma Immunization Guideline Hypertension Guideline Cough Guideline Breast Cancer Guideline Guideline Instances mapping between a model and the set of guideline tasks encoding of a guideline instance in a specific model execution of a guideline driven by the guideline tasks Figure 1. The relationship among the guideline instances, the guideline models, and the generalized guideline execution tasks in GESDOR. The guideline instances are encoded in specific representation models, while these models are mapped to the generalized guideline execution tasks. The guideline tasks are then used to drive the execution of the guideline instances encoded in different formats. METHODS The GESDOR guideline execution model comprises (1) a set of guideline representation models, which defines the domain to which the GESDOR guideline execution model can be applied, (2) a set of generalized guideline execution tasks that are extracted from the existing guideline representation models, (3) a set of mapping relationships, each of which corresponds to a specific guideline representation AMIA 2003 Symposium Proceedings − Page 694 model defined in (1) and provides the semantic links from the elements of that model to the guideline tasks defined in (2), and (4) a generic task-scheduling model, which harmonizes the existing approaches to task scheduling during guideline execution. To implement the GESDOR model, the generalized guideline execution tasks need to be extracted first. The mapping relationship between a specific guideline representation model and these guideline tasks needs then to be created. Finally, a generic taskscheduling model needs to be developed to harmonize the existing approaches to task scheduling. The Generalized Guideline Execution Tasks To extract the generalized guideline execution tasks, we performed a comprehensive literature review on the existing guideline representation models3. Guideline documentation models were used as complements to this review. Two specific guideline models, the 3rd version of the GLIF model (GLIF3)4 and a variant of the PROforma model (PROforma*)5, developed and structured as ontologies using the Protégé-2000 knowledge acquisition tool6, were used as the working templates during this process. Here the PROforma* model inherited most components of the original PROforma model, with the changes only in expression language, cyclic task execution, and patient data definition to simplify the implementation of the GESDOR execution engine. Based on these analyses, we have found a set of generalized guideline execution tasks and a guideline’s process structure that are common across different guideline representation models. These generalized guideline execution tasks include (1) the primary tasks, such as data collection, clinical intervention, medical decision making, patient state verification, branching, synchronization, and subguideline, which constitute the basic unit of a guideline’s process structure, and (2) the auxiliary tasks, such as criterion evaluation, event registration, and event invocation, which are used to support the execution of the primary tasks. To represent a generalized guideline execution task, we used (1) a set of input elements, which define the participants of the task, (2) a set of output elements, which define the execution effects of the task, (3) a set of subtasks, which define the other guideline execution tasks that are embedded within the task, and (4) a set of execution constraints, including preconditions, postconditions, and events, which define the restrictions on the invocation, completion, and triggering of a primary task. The generalized guideline execution tasks were then integrated and organized as an ontology, with each class representing a specific task, a structural element, or an execution constraint, and the slots of the class representing the attributes of that class or its relationships with other classes. We took an incremental approach to the development of this generalized guideline execution task ontology. During this process, we used Protégé-2000 (...truncated)