Semi-supervised classification of patient safety event reports.

OBJECTIVES: The Veterans Health Administration patient safety reporting system receives more than 100,000 reports annually. The information contained in these reports is primarily in the form of natural language text. Improving the ability to efficiently mine these patient safety reports for information is the objective of a proposed semi-supervised method. METHODS: A semi-supervised classification method leverages information from both labeled and unlabeled reports to predict categories for the unlabeled reports. RESULTS: Two different scenarios involving a semi-supervised learning process are examined, and both demonstrate good predictive results. CONCLUSIONS: The semi-supervised method shows much promise in assisting researchers and analysts toward accurately and more quickly separating reports of varying and often overlapping topics. The method is able to use the "stories" provided in patient safety reports to extend existing patient safety taxonomies beyond their static design.

Improving RCA performance: the Cornerstone Award and the power of positive reinforcement.

BACKGROUND: The Veterans Health Administration has had a comprehensive patient safety program since 1999 that includes conducting root cause analysis (RCA) of adverse medical events. Improving the quality and timeliness of the RCAs at the local level has been a continual challenge. METHODS: We initiated a non-monetary program called the Cornerstone Award into our patient safety reporting system to recognise facilities conducting high-quality and timely RCAs containing deterministic corrective actions that are implemented and evaluated for effectiveness. RESULTS: Since the Cornerstone Program began in 2008, the per cent of RCAs completed in a time-critical manner (≤45 days) has increased from an average of 52% pre-Cornerstone to an average of 94% post-Cornerstone. The per cent of action plans with stronger deterministic actions and outcomes has increased from an average of 34% pre-Cornerstone to an average of 70% post-Cornerstone. DISCUSSION: Implementing a non-monetary recognition award that was tied to specific improvement goals greatly improved the timeliness and quality of the RCA reports in the Veterans Health Administration System.

A cognitive aid for cardiac arrest: you can't use it if you don't know about it.

BACKGROUND: A cognitive aid developed by the Department of Veterans Affairs (VA) and distributed to all VA facilities provides caregivers with information to minimize omission of critical steps when diagnosing and treating cardiac arrest. In 2002, caregivers were surveyed about the usefulness of the cognitive aid and the success of its dissemination throughout the VA. METHODS: Fifty randomly selected VA hospitals were sent a letter to alert them of the upcoming survey. Twenty surveys were sent to each of the selected hospitals with instructions to distribute the survey to specific caregiver types. RESULTS: Nine (18%) of the VA hospitals had not used the cognitive aid tool because of dissemination problems. Of the 565 caregivers responding to the survey, 59% (332) were aware of the cognitive aid. Of these 332, 96% agreed that putting the cognitive aid on code carts is a good idea. There were 234 respondents who were both aware of the cognitive aid and had been involved in at least one code within the past 30 days. Of these 234, some 29 (12%) used the aid during a code, 28 of whom agreed that the cognitive aid was helpful during the code. DISCUSSION: Both new and experienced caregivers find the cognitive aid helpful when responding to "code" situations. However, cognitive aids cannot be helpful if theintended users are unaware of their availability. Dissemination and awareness of the aids can be problematic in large health care systems.

The Veterans Affairs root cause analysis system in action.

BACKGROUND: The patient safety program in the Department of Veterans Affairs (VA) began in 1998, when the National Center for Patient Safety (NCPS) was established to lead the effort on a day-to-day basis. NCPS provides the structure, training, and tools, and VA facilities provide front-line expertise, feedback about the process, and root cause analysis (RCA) of adverse events and close calls. MONITORING THE PROCESS: Facility patient safety managers determine the disposition of adverse events and close calls occurring at their facilities. They use a safety assessment code (SAC) to prioritize the actual and potential severity and frequency of an event. BEFORE-AND-AFTER STUDY: Before the new RCA system was implemented in 2000, the VA used another adverse event reporting system, focused review (FR). A comparison of the two processes indicates that the RCA process has shifted analyses of adverse events toward a human factors engineering approach-entailing a search for system vulnerabilities rather than human errors and other less actionable root causes. CASE EXAMPLES: Two case examples--on hazards in the magnetic resonance imaging (MRI) room and on a cardiac pacemaker malfunction--illustrate how the RCA system works in actual operation. The cases illustrate that broadly applicable, high-impact actions can result from a thorough RCA process. DISCUSSION: NCPS monitors the quality and completeness of RCAs through the immediate review and feedback process. Still to be investigated is the effectiveness of RCA actions addressing the hypothesized root causes and contributing factors of the close calls and adverse events.