Retrospective studies have shown artificial intelligence (AI) to be at least as accurate as radiologists in detecting breast cancer in screening mammograms. This prospective randomized trial similarly demonstrated that AI readings were as accurate as double readings by radiologists, but with a significantly reduced workload.

Persons with dementia are at increased risk for adverse events compared to those without dementia, highlighting the importance of a timely diagnosis. In this study, researchers estimate approximately 20% of primary care patients aged 65 and older are expected to have a diagnosis of mild cognitive impairment or dementia; however, only 8% have received such a diagnosis. Missed diagnosis prevents patients from receiving appropriate care, including newly FDA-approved medications to slow cognitive decline.

Surgical safety checklists (SCC) are utilized around the world, but checklist completion at the operating room level remains inconsistent. This review summarizes facilitators and barriers to completion. Resistance or endorsement at the individual surgeon level remains a significant factor in SSC completion. Early inclusion of frontline staff in evaluation and implementation supported increased use.

Psychological safety is an important principle in identifying problems and improving patient outcomes. This narrative review highlights five best practices to foster psychological safety in radiology residencies – (1) establish clear goals and educational strategies, (2) build a formal mentoring program, (3) assess psychological safety, (4) advocate for radiologists as educators, and (5) support non-radiology staff. Although the review focuses on radiology residency programs, these strategies can be adapted to any residency program.

Non-English primary language (NEPL) patients may encounter barriers navigating the healthcare system and communicating with providers. In this retrospective study, researchers used the Safer Dx tool to explore differences in diagnostic errors among NEPL versus English-proficient (EP) patients. Among 172 patients who experienced a diagnostic error, the proportion was similar among EP and NEPL groups and NEPL did not predict higher odds of diagnostic error.

Artificial intelligence (AI) is increasingly used in healthcare, but concerns have been raised that it can exacerbate existing disparities because of underlying biases in AI tools. In this observational study, researchers evaluated biases in clinician versus AI chatbox responses to 19 clinical vignettes involving cardiology, emergency medicine, rheumatology, and dermatology. Findings indicate that both AI chatboxes and clinicians provide different clinical recommendations based on a patient’s gender, race/ethnicity, and socioeconomic status under certain clinical scenarios.

The pediatric population faces unique challenges in the prehospital setting. This prospective chart review study classified adverse safety events (ASE) of pediatric patients at 15 emergency medical services (EMS) agencies. More than 20% of encounters contained at least one ASE, although most were unlikely to cause harm (e.g., missed documentation).

To improve patient safety, payers such as the Centers for Medicare & Medicaid have implemented policies that limit reimbursement for certain healthcare-associated harms. This commentary introduces the “Payer Relationships for Improving Diagnoses (PRIDx)” framework describing how payers may implement similar policies to reduce diagnostic errors.

Diagnostic testing errors can contribute to delays in diagnosis and to serious patient harm. Researchers analyzed 327 voluntary incident reports from one medical center in the Netherlands and found that diagnostic testing errors most commonly occurred during the pre-analytic phase (77%), and were predominantly caused by human factors (59%). The researchers found that these diagnostic testing errors contributed to a potential diagnostic error in 60% of cases.

Clinician skills can decrease with age. This retrospective analysis of 1.9 million preliminary interpretations of radiology imaging findings examined the relationship between radiologist age and diagnostic errors. While the overall mean error rate for all radiologists was low (0.5%), increasing age was associated with increased relative risk of diagnostic errors.

Trigger tools are commonly used to detect adverse events and identify areas for safety improvement. This study found that trigger tools using electronic health record-based data can accurately identify delayed diagnosis of appendicitis in pediatric patients in community emergency department (ED) settings.

As high workload and interruptions are known contributors to diagnostic errors, significant research has been conducted to understand and ameliorate the impact of these factors. This study examined the association between hospitalist busyness (i.e., number of admissions and pages), resource utilization, number of differential diagnoses, and the hospitalist's diagnostic confidence and subjective awareness. Increasing levels of busyness were associated with hospitalists reporting it was "difficult to focus on what is happening in the present" but had no effect on diagnostic confidence.

Characteristics inherent to the emergency department (ED), such as overcrowding and unfamiliar patients, make it susceptible to errors. This article describes a novel process for identification of ED errors by adding the question, "Would you have done something differently?" to the chart review process. Adding this question and requiring a detailed explanation of what they would have done differently allowed for differentiation between a true medical error and a judgment call that coincides with an adverse event. Near misses, adverse events, and adverse events attributable to error were significantly higher when reviewers would have done something differently.

Emergency department triage nurses use their knowledge, experience, expertise, and critical thinking skills to prioritize patients by severity, ensuring the sickest patients are seen first. This study sought to identify cognitive biases that may negatively impact nurses' triage decision making. In a scenario describing a patient at Emergency Severity Index (ESI) level 1, the highest level, only 51% of nurses selected the cognitively unbiased triage response.