Skip to Main Content
Clinical practice guidelines (CPGs) are systematically developed healthcare recommendations designed to improve quality and control costs by reducing errors, minimizing practice variability, and promoting best practices. Although considerable effort and resources have been applied to CPG design, development, and deployment, the impact of CPGs on clinician behavior is inconsistent at best. The biomedical literature suggests that CPG efficacy could be improved using information systems that make CPGs and related patient-specific decision support functions available and easier to use at the point of care. This approach also depends upon the ability to integrate CPG information systems into the clinician workflow. Based on an analysis of the barriers to efficient and effective guideline use, researchers from City of Hope National Medical Center's Division of Information Sciences were motivated to develop a Web-based expert system to facilitate CPG utilization in a variety of clinical environments. The resulting Graphical Decision Support Interface (GDSI) employs a relational database-driven state machine architecture adapted from an instrument control system developed by one of the authors. The user interface was designed to facilitate clinician interaction with large, complex decision hierarchies. As users enter information about an individual patient, the system computes additional derived values and provides context-specific recommendations. It also documents when and why clinicians intentionally deviate from guideline recommendations to assist with organizational benchmarking and CPG modification efforts. This report describes the GDSI prototype and our experience with the translation of breast cancer treatment guidelines into algorithms with explicit states and decision points. It also describes a pilot study of the system at two Southern California cancer treatment facilities. Preliminary results suggest that the GDSI system is well received among clinicians and is capable of making a positive contribution toward CPG development and use. Lessons learned and directions for future research are also discussed.