The Lucas, Gunter, et al. article describes a model of a comprehensive systems framework to improve clinical practice for a given disease across the continuum of care. The model is applicable in any clinical setting. Instead of using the old inspection-oriented, quality assurance or profiling approaches, the authors' model uses clinical practice improvement (CPI) methodology to determine in detail the content and timing of individual medical care process steps to be used for similarly ill patients. The goal of the model described is to produce best clinical outcomes for the least necessary cost. Similarity of illness is determined through disease-specific criteria that define a group of patients that clinicians agree should receive the same medical care process, except possibly for small, random differences among patients. Clinical outcomes are analyzed to identify statistically meaningful associations between one or more process steps and one or more medical outcomes.

  The final product of a CPI study is a research-based protocol, which is a decidable and executable diagnostic or therapeutic process of medical care, designed (a) to improve outcomes over those achieved by the clinician group before the CPI study and at the least necessary cost, or (b) to maintain substantially the same outcomes as those achieved by the clinician group before the CPI study, but at lower cost. There are inherent exceptions to research-based protocols as actually practiced by clinicians for patients. Clinicians explain their exceptions by citing non-stylistic, objective reasons for them, and the exceptions and reasons are discussed by the clinician group in order to confirm or selectively modify the protocol to cover particular exceptions.

  There is widespread interest in the potential of guidelines to improve the quality and lower the cost of medical care in all practice settings. Research-based protocols, developed through the CPI process, are even more likely than guidelines to succeed for these purposes. This article discusses the key features and advantages of the CPI approach, the challenges for implementation, and lessons learned. In particular, it points out the need to train clinicians in CPI methodology, to convert from a quality assurance or quality management department to a department of Clinical Practice Improvement, and the requirement to collect data about patients, process, and outcomes. Outcomes data alone can only identify outcome failures. If clinicians are expected to follow guidelines, they must be provided with outcomes data to show how well the guidelines work, and precise specifications of when to use the steps of the guidelines. All three components are needed: details about patient characteristics and the corresponding process steps associated with the best outcomes, so that the optimal processes can be determined and implemented consistently. The CPI methodology discussed in this paper is the most promising approach available today for determining how to practice high quality, cost-effective health care.