Software Leveraging a Standards-Based Web Service Framework for Decision Support

DESCRIPTION (provided by applicant): Summary. Despite their demonstrated effectiveness, clinical decision support systems are not widely used to assist decision making in routine clinical practice. The long-range goal of this application is to leverage a standards-based Web service framework to create a tool that will make clinical decision support more available and more usable in health care. The innovative aspects of this project include a standards-based method for encoding and processing machine-executable medical knowledge that is easier to use than existing formalisms, as well as a method for delivering machine-executable knowledge through a Web service that is external to and independent of client systems. This Web service approach enables diverse applications from multiple institutions to provide clinicians with decision support regarding various conditions. This project responds directly to the interests of the National Library of Medicine (NLM) to create informatics tools that "provide decision support for clinicians." Based upon two proof-of-concept applications, a general prototype decision support Web service has been developed that includes (1) a patient information model based on the Health Level 7 Reference Information Model and concepts from the NLM Unified Medical Language System; (2) executable knowledge modules that capture medical knowledge in a simple, expandable, and computable format; and (3) a set of services for evaluating patient information and for administering the system. This application will focus on two specific aims for the purpose of showing the technical feasibility of building a clinical decision support system (CDSS) using this flexible Web service architecture. Aim 1 is to instantiate a prototype of this CDSS to deliver diabetes care reminders to the point of care in a primary care practice setting. Aim 2 is to evaluate the performance of the CDSS and to assess its acceptability to end users. Duke University and Religent, Inc. will collaborate to develop the software components required to implement this Web service-based CDSS in practice. Commercial opportunities for this system include supporting pay-for-performance initiatives for clinicians and improving care quality metrics for health plans. The successful development of a Web service-based decision support system should increase the availability and usability of decision support tools in diverse settings for various applications. It should also enable centralized management of medical knowledge resources and enable sharing of computable medical knowledge across a care delivery system or a geographic region. Increased use of decision support tools can be expected to improve care quality, reduce medical errors, lower healthcare costs, and augment disease management for patients and populations.