Building on previous research on the use of text mining related to medication administration error incidents, researchers in this study found that artificial intelligence can be used to accurately classify the free text of medication incident reports causing serious or moderate harm, to identify target risk management areas.

Many strategies have been developed to improve recognition of, and response, to clinically deteriorating patients. This review found that simulation-based educational strategies was the most effective educational method for training staff to recognize unwell patients. However, the quality of evidence was low and additional research into simulation-based education is needed.

Misuse of prescription opioids represents a serious patient safety issue. Using commercial claims from 2014 - 2018, researchers examined the association between the 2016 CDC guidelines to reduce unsafe opioid prescribing and opioid dispensing for patients with four common chronic pain diagnoses. Findings indicate that the release of the 2016 guidelines was associated with reductions in the percentage of patients receiving opioids, average dose prescribed, percentage receiving high-dose prescriptions, number of days supplied, and the percentage of patients receiving concurrent opioid/benzodiazepine prescriptions. The authors observe that questions remain about how clinicians are tailoring opioid reductions using a patient-centered approach.

Prioritization of patient complaints allows inspectors to follow up more quickly on those that pose the most severe safety risk. Using text mining and sentiment analysis, more than 22,000 patient complaints were assigned a severity category. The ‘bag-of-words representation’ was most successful for severity predicting and could be used to triage patient complaints.

There are many challenges associated with detecting and measuring patient safety events. This meta-review provides an overview of approaches to measuring and monitoring safety in primary care. The authors suggest that instead of developing new methods for measuring and monitoring safety, researchers should focus on expanding the generalizability and comparability of existing methods, many of which are readily available, quick to administer, do not require external involvement, and are inexpensive.

The World Health Organization (WHO) released the International Classification of Diseases, 11th Revision (ICD-11) in 2018. In addition to the medical entities such as disease and injury, it contains a second component, the ICD-11 Mortality and Morbidity Statistics (MMS) linearization. The authors evaluated whether the ICD-11 MMS is appropriate for use in patient safety and quality or if a USA-specific clinical modification is necessary. 

Trauma resuscitation is a complex, specialized process with a high risk for errors. Researchers analyzed videotapes of in situ simulations to evaluate latent safety events occurring during trauma resuscitation. Themes influencing latent safety events related to physical workspace, mental model formation, equipment, unclear accountability, demands exceeding individuals’ capacity, and task-specific issues.

Despite public policies and guidelines to reduce opioid prescribing, providers continue to overprescribe these medications to children, adolescents, and young adults. In this analysis of US retail pharmacy data, 3.5% of US children and young adults were dispensed at least one opioid prescription; nearly half of those included at least one factor indicating they were high risk. Consistent with prior research, dentists and surgeons were the most frequent prescribers, writing 61% of all opiate prescriptions.

Artificial intelligence (AI) and big data have potential to impact health care improvement in a variety of ways. This commentary examines the intersection of AI, diagnosis, and inequality. The authors highlight several innovations that could reduce health disparities and diagnostic errors in at-risk patient populations if adopted.

In this qualitative study, researchers interviewed 40 clinicians in high- and low-performing hospitals to better understand the barriers to effective organizational learning from medical errors. Findings from these interviews suggest that the primary barriers to active learning stem from social issues post-reporting – e.g., lack of trust or proactive engagement from management. The authors highlight the importance of fostering an organizational culture that encourages cooperation and collaboration between management and clinicians.

Computerized provider order entry (CPOE) systems have the potential to reduce error, but their poor CPOE design, implementation and use can contribute to patient safety risks. In this study, researchers found that restricting the number of concurrently open electronic health records did not significantly reduce wrong patient selection errors in their hospital’s CPOE system.

Physicians commonly experience adverse psychological outcomes after being involved in an adverse event. This mixed-methods study of health care providers in the Department of Obstetrics & Gynecology at one large academic hospital found that three-quarters of providers experienced symptoms of traumatic stress after involvement in an adverse event. Respondents reporting anxiety were more likely to be interested in peer-to-peer support programs, whereas those reporting symptoms of guilt preferred debriefing sessions.

Root cause analysis (RCA) is a common method to investigate adverse events and identify contributing factors. To expand resident understanding of and participation in RCA, the authors developed simulated RCAs that were applicable to a broad set of specialties and included other healthcare professionals whose disciplines were involved in the event (e.g., nurses, pharmacists). After participating in the simulated RCAs, there was an increase in trainees understanding of RCA and intent to report adverse events.

Medical residents are frequently involved in difficult patient conversations, including error disclosure. This paper describes the development and implementation of a novel communications/didactic skills training program for OB/GYN residents. Immediately, and 3-months after training, residents indicated an improvement in their communication skills.

The goal of the patient centered medical home (PCMH)  model is to reorganize primary care to provide team-based, coordinated, accessible health care. This study used a consensus process with input from a physician panel to examine ambulatory patient safety concerns (e.g., medication safety, diagnostic error, treatment delays, communication or coordination errors) in the context of the PCMH model and explore variability in the implementation of patient safety practices.

Research has shown that patients admitted to the hospital on the weekend may experience worse outcomes compared to those admitted on weekdays (the ‘weekend effect’). This editorial highlights the challenges to empirically evaluate the underlying mechanisms contributing to the weekend effect. The authors propose viewing the weekend effect as a proxy for staffing levels and the influence of other factors influencing outcomes for patients admitted on weekends, such as patient acuity, clinician skill-mix and access to diagnostic tests or other ancillary services.

Patient suicide in all settings is considered a never event. Nurses caring for the patient may experience negative psychological symptoms following inpatient suicide. This review identified five themes based on nurses’ psychological experiences: emotional experience, cognitive experience, coping strategies, self-reflection, and impact on self and practice. Hospital administrators should develop education and support programs to help nurses cope in the aftermath of inpatient suicide.  

Missed sepsis diagnosis can lead to increased morbidity, mortality and length of stay. Using administrative data, this retrospective study estimated the risk of potentially missed pediatric sepsis in several emergency departments. Approximately 8% of pediatric patients admitted to the hospital with sepsis experienced a treat-and-release emergency department visit within the prior 7 days. Administrative data can be helpful for hospitals in identifying cases that require detailed record review as well as evaluating the impact of sepsis alerts and bundles.

The Hospital Star Rating system was implemented in 2016 to aid patients in identifying safe, high-quality hospitals. The system has evolved over time to address applicability concerns. This commentary highlights changes in the latest reiteration of the program and discusses challenges in its use.