Guidelines provide a systematic means to improve patient management. Their use can enhance the quality, appropriateness, and effectiveness of health care, and may also contain costs. Optimal development of guidelines, however, would include input from experts in clinical medicine, meta-analyses, decision analyses, cost-effectiveness analyses, and evidence synthesis; such experts often are not available, or are prohibitively costly, on a local level. Therefore, national organizations often develop guidelines for a population that has """"""""average"""""""" characteristics. These guidelines may need to be adapted to local settings in which the characteristics of the patient population or practice are different. In addition, traditional guidelines are normally static and thus may become out of date as new information emerges. To address these problems, we propose that developers create guidelines from decision models. These guidelines contain the knowledge necessary to allow users to customize the guidelines to a local patient or population, and to modify the guidelines over time as the underlying decision model or evidence evolves. To achieve this goal, our proposed research has three specific aims: (1) to develop methods for the automated creation of guidelines from evidence-based decision models, using a web-based system called ALCHEMIST, (2) to create and evaluate guidelines based on the findings of the Cardiac Arrhythmia PORT, a 6-year multi-institutional study of methods to prevent sudden cardiac death, and (3) to evaluate, in a randomized controlled trial, clinicians' use of and their satisfaction with our developed web- based guideline system compared with traditional guidelines for making treatment recommendations on a set of hypothetical clinical scenarios. If we successfully complete our specific aims, we will have: (1) demonstrated that developers can use decision models to create guidelines automatically, (2) used the ALCHEMIST system to create an evidence-based customizable guideline for prevention of sudden cardiac death based on the findings of the Cardiac Arrhythmia PORT, (3) evaluated the benefits and quality of the ALCHEMIST- generated guidelines, and (4) laid the groundwork for a randomized trial of clinical implementation of the ALCHEMIST guideline system. These achievements will help further the missions of the AHRQ and the NLM by leading toward a formalized method for combining evidence to create guidelines, and facilitating dissemination and translation of these guidelines into clinical practice.
