Multiple handoffs can occur during perioperative care, which can increase the risk for errors and patient harm. This narrative review summarizes research on the benefits, limitations, and implementation challenges of electronic tools for perioperative handoffs and the role of artificial intelligence (AI) and machine learning (ML) in perioperative care.

Handoffs and transitions of care represent a vulnerable time for patients as important information must be shared and understood by multiple people. This study focuses on postoperative nurse handoffs, specifically regarding situational awareness and anticipatory guidance, and the role artificial intelligence (AI) could play in improving handoffs. Five themes were uncovered, including the importance of situational awareness and associated barriers, how AI could address those barriers, and how AI could result in new/additional barriers.

Multiple handoffs occur during the perioperative period, each presenting an opportunity for miscommunication and patient harm. This review uses the Systems Engineering Initiative for Patient Safety (SEIPS) framework to describe the barriers and enablers for improving staff communication pre-, intra-, and post-operative handoffs. Structured hand-offs, checklists, protocols, and interprofessional teamwork were cited as enablers for improved communication.

Cross-sectional research has suggested many physicians experience burnout which can negatively impact patient safety. This longitudinal study evaluated the effect of workload (collected via electronic health record audit) on burnout and medication errors (i.e., retract-and-reorder [RAR] events) of internal medicine interns. Higher levels of workload were associated with burnout; there was no statistically significant association between burnout and RAR events.

… involved with medication ordering errors . … King CR, Abraham J, Fritz BA, et al. Predicting self-intercepted …

Intraoperative handoffs between anesthesiologists are frequently necessary but are not without risk. This systematic review of 14 studies of intraoperative handoffs and handoff tools found that use of handoff tools has a positive impact on patient safety. Additional research is needed around design and implementation of tools, particularly the use of electronic health records to record handoffs.  

… for safe transfer from operating room to intensive care. … Abraham J, Meng A, Sona C, et al. An observational study of …

… fewer information omissions, and fewer technical errors. … Abraham J, Meng A, Tripathy S, et al. Systematic review and meta-analysis of …

Prior research has found that intraoperative anesthesia handovers can increase patient morbidity and mortality. However, this retrospective cohort study, focused on pediatric surgical patients treated, found that intraoperative anesthesia handovers were not associated with adverse postoperative outcomes.  

… environmental, social, and organizational-level factors. … Abraham J, Galanter WL, Touchette D, et al. Risk factors …

This systematic review of the cumulative effect of computerized provider order entry (CPOE) identified significant decreases in medication errors and adverse drug events in inpatient settings but the authors note considerable variation in the magnitude of risk reduction. No significant reductions in inpatient mortality or length of stay were identified. 

Prior research has found that many clinicians do not engage in active listening behaviors essential to safe patient care. This prospective observational study used a mixed-methods approach to better understand listening and question-asking behaviors during residents and nurses handoffs. The researchers did not identify any significant differences between residents and nurses in their active or passive listening behaviors, but they did find that nurses asked significantly more questions than residents.

This simulation study compared computerized medication order entry between two commercial electronic health records and a prototype designed for safe prescribing. Physicians using the prototype had fewer errors compared to either commercial platform, highlighting the need to improve electronic health record usability in order to enhance medication safety.

Opening multiple patients' charts in the electronic medical record simultaneously may increase the risk of wrong-patient orders, a known patient safety hazard. Researchers analyzed intercepted wrong-patient medication orders in an emergency department over a 6-year period and found no significant reduction when the maximum number of charts allowed to be open at the same time decreased from 4 to 2. Similarly, there was no significant increase when the maximum number of charts permitted to be open simultaneously increased from 2 to 4.

Although computerized provider order entry (CPOE) has been shown to prevent some medication errors, implementing these systems may introduce new errors. This observational study analyzed voided clinician orders to determine if they represented true medication ordering errors identified by the canceling provider. Investigators interviewed ordering and voiding providers and reviewed patients' medical records. They found that the positive predictive value of canceled orders representing true medication ordering errors was 93.1%. The reasons for order cancellation were more accurately reported during provider interviews rather than in the reasons reported within the CPOE system. Common ordering errors included duplicate orders and incorrectly composed orders. A past PSNet perspective discussed the impact of CPOE on patient safety.

Although computerized provider order entry has been found to prevent some medication errors, simulation studies have also demonstrated that electronic prescribing platforms can introduce or fail to prevent medication errors. This retrospective electronic health record analysis examined medication orders that were canceled. Weekend and overnight orders were less likely to be voided than weekday or daytime orders. Pharmacist, nurse, and student orders were more likely to be canceled than physician orders. Comparing the clinician-provided reason for voiding an order with the more comprehensive information in the medical record, physicians found that clinicians' reported reasons for voiding orders were largely inaccurate. The authors suggest there is unrealized potential to characterize medication ordering errors using voided-order data.

Cognition has been recognized as a human factor that can contribute to failures in health care. This review examines cognitive aspects of human error that affect patient safety, methods to augment detection of flawed decision-making, and the potential for educational approaches like virtual reality simulation to train physicians to manage cognitive error once it occurs. A Perspective interview with Dr. Pat Croskerry explored the role of cognition in medical error.

The patient safety risks associated with handoffs have been well documented. As a result, multiple investigators have developed standardized tools to improve the quality of information transfer during handoffs. What remains unclear is the extent to which standardizing the handoff process improves patient safety. This systematic review of 36 studies examining the effectiveness of handoff tools found that most tools were not evaluated rigorously and did not specifically assess the effect of standardizing handoffs on patient-level outcomes. Therefore, the authors were unable to reach conclusions regarding the optimal methods for improving the quality of handoffs. Similar problems were noted in studies of checklists, another widely implemented safety intervention, highlighting both the difficulty and the importance of strictly evaluating patient safety interventions.

Direct observation of handoffs in the intensive care unit was used to develop a new framework for improving handoffs that emphasizes the importance of ensuring overall continuity of care.