… and its potential to enhance clinical decision-making as a tool for safe value-based care. The authors discuss how …

This review article delineates how artificial intelligence approaches to medical diagnosis are vulnerable to adversarial attacks. The authors suggest that regulatory review of devices that employ machine learning include an assessment of resilience against such attacks.

Use of artificial intelligence (AI) and computer algorithms as tools to improve diagnosis have both risks and benefits. This commentary explores challenges to implementing AI systems at the front line of care in a safe manner and identifies weaknesses of advanced computing systems that influence their reliability.

Improving diagnosis remains a major focus within patient safety. For patients with rare diseases, diagnosis can often be delayed. Established in 2014 and funded by the National Institutes of Health, the Undiagnosed Diseases Network (UDN) applies a multidisciplinary approach to the most challenging diagnostic cases. Over a 20-month period, 601 out of 1519 patient cases were accepted by the UDN for evaluation. The authors report that of the 382 patients who underwent a complete evaluation, a diagnosis was identified in 132 patients.

Harm from opioids is a widely recognized patient safety concern. In this retrospective cohort study, investigators examined the effect of postoperative opioid prescribing in patients who had never received opioids before. As with prior studies, they found increased subsequent misuse of opioids among patients who received larger quantities of opioid medications following surgery compared to those who received fewer opioid medications. Longer duration of postoperative opioid prescription was also associated with higher odds of future diagnosis of opioid misuse. This study adds to evidence demonstrating the potential harms associated with even short-term opioid prescription. A recent PSNet interview discussed the opioid epidemic and strategies to address this growing patient safety concern.

In 2008, the Centers for Medicare and Medicaid Services (CMS) eliminated reimbursement for certain preventable errors and hospital-acquired infections. This landmark policy aimed to align financial disincentives with adverse events, an increasingly utilized strategy. However, this AHRQ-funded study found that the "no pay for errors" policy had no measurable effect on rates of catheter–associated bloodstream infections and catheter–associated urinary tract infections in hospitals in the United States. No subgroup of hospitals or patients identified in this national evaluation seemed to clearly benefit from this policy change. The benefits and limitations of the CMS policy are discussed in an AHRQ WebM&M interview with Dr. Robert Wachter.

… Hosp Epidemiol … The Institute for Healthcare Improvement's 100,000 Lives Campaign generated national attention for …

Antibiotic-resistant bacteria, such as methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus faecalis (VRE), are frequent sources of hospital-acquired infection (HAI) in the intensive care unit (ICU). Although the incidence of serious infections caused by MRSA has been decreasing, the optimal strategies to prevent spread of these bacteria remain unclear. In this cluster-randomized trial conducted in 18 ICUs, a protocol that involved universal surveillance and barrier precautions (gowns and gloves) for patients colonized with these bacteria was evaluated for effectiveness at preventing colonization and infection with MRSA or VRE. No reduction in colonization or infection was found, in part attributable to the fact that use of barrier precautions was suboptimal. Prior successful efforts to reduce HAI have emphasized the role of safety culture in addition to specific preventive interventions, an approach discussed in-depth in this analysis of the landmark Keystone ICU project.

This review highlights communication tactics to improve discharge in the emergency department.

Despite the past decade's intense focus on patient safety, one fundamental question remains unanswered: are hospitalized patients safer than they were 10 years ago? Unfortunately, this study indicates that the answer may be "no." The investigators analyzed medical records from 10 North Carolina hospitals over a 6-year period, using the Institute for Healthcare Improvement's Global Trigger Tool to identify possible adverse events, and found that the incidence of both preventable and non-preventable adverse events remained unchanged. Recent research has also confirmed a persistently high rate of adverse events in Medicare and privately insured patients. The recent 10th anniversary of the seminal Institute of Medicine report that launched the patient safety movement prompted several unsparing assessments of the state of the field, including a commentary by Dr. Robert Wachter and a plenary session at the National Patient Safety Congress.

This study found satisfactory reliability between internal and external reviewers using the IHI Global trigger tool to identify adverse hospital events.

… and safety. Investigators found no relationship between a hospital's quality status and its residents' curriculum or understanding of quality and safety. While a few programs had formal curricula, the majority of …

Medication errors are common in pediatric inpatients, but the best method of addressing them remains unclear. Studies of technological interventions, such as computerized provider order entry, have yielded inconsistent results. In this controlled trial, clinical pharmacists were deployed in the intensive care unit (ICU) and general medical and surgical wards in a pediatric hospital, and their effect on prevention of medication errors was assessed. Serious medication errors (including near misses) were significantly reduced in the ICU. No effect was seen on medication error rates for general ward patients, although the baseline rate of errors was much lower in those areas. A prior review documented the effectiveness of pharmacists at preventing medication errors in a variety of inpatient settings.

… prescriptions were at increased risk for suffering a preventable adverse drug event. Improved communication …

… The overall rate of preventable ADEs was similar to a prior outpatient study , but nearly three-quarters of …

… events (AEs), have been used to screen for errors in a variety of clinical settings . This AHRQ–funded study used methodology similar to a prior study in adult intensive care unit patients to develop a chart-based set of triggers for error identification in the …

… of medicine … N Engl J Med … The author argues that a systems approach  to hand hygiene is only useful if …

This study demonstrated that more than half of the patients in a single neonatal intensive care unit are at risk for misidentification errors due to similarities in patient names or medical record numbers.

This study found that medical errors often occur in patients with bronchiolitis, even more so in the critically ill.