A considerable void exists in the information available regarding anatomic pathology diagnostic errors and their impact on clinical outcomes. To fill this void and improve patient safety, four institutional pathology departments (University of Pittsburgh, Western Pennsylvania Hospital, University of Iowa Hospitals and Clinics, and Henry Ford Hospital System) have proposed the development of a voluntary, Web-based, multi-institutional database for the collection and analysis of diagnostic errors. These institutions intend to use these data proactively to implement internal changes in pathology practice and to measure the effect of such changes on errors and clinical outcomes. They believe that the successful implementation of this project will result in the study of other types of diagnostic pathology error and the expansion to national participation. The project will involve the collection of multi-institutional anatomic pathology diagnostic errors in a large database that will facilitate a more detailed analysis of these errors, including their effect on patient outcomes. Participating institutions will perform root cause analysis for diagnostic errors and plan and execute appropriate process changes aimed at error reduction. The success of these interventions will be tracked through analysis of postintervention error data collected in the database. Based on their preliminary studies, these institutions proposed the following specific aims: Specific aim #1: To use a Web-based database to collect diagnostic errors detected by cytologic histologic correlation and by second-pathologist review of conference cases. Specific aim #2: To analyze the collected error data quantitatively and generate quality performance reports that are useful for institutional quality improvement programs. Specific aim #3: To plan and implement interventions to reduce errors and improve clinical outcomes, based on information derived from root cause analysis of diagnostic errors. Specific aim #4: To assess the success of implemented interventions by quantitative measure of postinterventional errors and clinical outcomes and by qualitative assessment by project participants. Funding for this project was approved by the Agency for Health Care Research and Quality in September 2002, and data collection and analysis are ongoing. Over 5000 errors have been collected in the database, and the clinical outcomes of these errors have been tracked. At a national meeting in November 2003, root cause analysis was performed to determine causes of errors. The findings of these root cause analyses have been presented at national pathology meetings and are currently being published.