Miscommunication is a major contributor to adverse events. This article describes the development of a framework to classify communication errors that contributed to a patient safety incident. Nine types of communication errors were identified. Falls and delays in diagnosis, treatment, or surgery were the most common adverse events related to communication errors.

Mnemonics are a commonly used memory aid to ensure standardized, structured handoffs between providers. This review summarizes mnemonics used in perioperative handoffs. Situation, Background, Assessment, Recommendation (SBAR) and SBAR variations were the most common, followed by I-PASS (Illness severity, Patient summary, Actions list, Situation awareness, Synthesis), and I-PASS variants. Most studies reported only on process outcomes; only four measured patient outcomes.

Artificial intelligence (AI) and machine learning (ML) are being used and tested in numerous ways. This review highlights how they are being used to detect and mitigate human error in safety-critical industries, the limitations and challenges of AI/ML, and insights from the recent literature. Examples from health care include using AI to detect diagnostic errors and combining AI with clinician expertise, with the ultimate decision to follow AI’s suggestion resting with the clinician.

Learning from errors is a fundamental component of health professions education. This randomized trial examined the difference between two error-framing strategies—Error Management (EM) and Error Avoidance (EA)— on cardiopulmonary resuscitation (CPR) performance among first-year health professions students. The EA perspective aims to avoid errors as much as possible, even during the learning process. In contrast, the EM perspective fosters a positive approach to errors and considers them part of the learning process. Results found that EM led to equivalent or slightly better performance than EA and control arms. The authors conclude that EM is a promising approach for medical education given its potential long-term benefits for patient safety and non-detrimental effect on short-term performance.

Artificial intelligence (AI) and machine learning (ML) are being used and tested in numerous ways. This review highlights how they are being used to detect and mitigate human error in safety-critical industries, the limitations and challenges of AI/ML, and insights from the recent literature. Examples from health care include using AI to detect diagnostic errors and combining AI with clinician expertise, with the ultimate decision to follow AI’s suggestion resting with the clinician.

Learning from errors is a fundamental component of health professions education. This randomized trial examined the difference between two error-framing strategies—Error Management (EM) and Error Avoidance (EA)— on cardiopulmonary resuscitation (CPR) performance among first-year health professions students. The EA perspective aims to avoid errors as much as possible, even during the learning process. In contrast, the EM perspective fosters a positive approach to errors and considers them part of the learning process. Results found that EM led to equivalent or slightly better performance than EA and control arms. The authors conclude that EM is a promising approach for medical education given its potential long-term benefits for patient safety and non-detrimental effect on short-term performance.