Anesthesia Needs to Lead the Way in Safety—Again—through the Universal Adoption of Structured Handoffs

The first published discussions of anesthesia handoffs began several decades ago. That work highlighted both the risks and potential benefits of handoffs, using observational approaches to develop a portrait of some of the critical aspects of exchanging anesthesia personnel. More than 40 yr later, many more studies have examined the link between anesthesia handoffs and adverse patient outcomes.  Only a few have studied interventions to address whether standardizing handoffs can improve patient outcomes as Saha and Segal describe in this issue of Anesthesiology. The report by Saha and Segal should be a call-to-action for all anesthesia professionals to conduct a deliberate, structured process for every instance in which responsibility for care of a patient is permanently transferred from one clinician or team to another.

The evidence supporting the relationship between unstructured anesthesia handoffs and harm is compelling but mixed. Most relevant studies demonstrate a robust association between “complete” anesthesia handoffs—in which care is permanently transferred from one clinician or team to another—and a host of adverse patient outcomes (table 1). A recent report finding harm caused by handoffs called for forbidding any handoffs except in extenuating circumstances.  An editorial refuted that suggestion; we agree with the editorial, i.e., that the benefit of refreshing the vigilance of the anesthesia provider and bringing in a new set of eyes and fresh thinking to any procedure can be an overall positive action, but only if a handoff is conducted in a way that transfers information effectively and allows for the new provider to accumulate tacit knowledge from the one relinquishing responsibility.

There are at least two reasons that understanding the relationship between anesthesia handoffs and harm is challenging when compared to other perioperative care interventions. First, there is not a one-to-one relationship between handoffs and harm. A patient care episode with a “perfect” handoff could be marked by adverse effects, while another with completely ineffective communication might not result in harm. As evoked in James Reason’s “Swiss cheese” model of safety multiple defects often need to align for harm to reach patients; handoffs represent just one potential defect. This relationship makes it difficult to use traditional statistical approaches to identify a relationship between anesthesia handoffs and harm, even using advanced approaches like propensity score matching and analogous methods. Second, there is a marked asymmetry between the benefits and harms of handoff interventions that should cause us to question what level of evidence is needed to spur action. Avoiding handoffs entirely is infeasible given workforce shortages, burnout, and production pressures. On the other hand, the risks of structuring handoffs are trivial compared to the risks of haphazard, nonstructured communication approaches.

The study by Saha and Segal examined 121,077 cases, 40% of which involved a handoff between anesthesia providers in noncardiac cases, during 5 yr at a single institution. Using retrospective outcome data, they examined composite mortality and major morbidity of patients and compared cases before and after the introduction of a structured handoff checklist embedded in their electronic health record. The handover cases had 7.2% adverse outcomes versus 6.2% for the nonhandover cases. The use of the structured handoff increased from 30 to 90%, and the number of adverse outcomes decreased in gradual fashion as handoff use increased. For the reasons mentioned above, it is difficult to definitively translate process changes to harms avoided, but trends over time provide evidence of a causal association.

Like any study, there are imperfections in this one, e.g., pre–post design, and use of a composite outcome. Yet the work by Saha and Segal supports the ideas that (1) handoffs without using a protocol are too often the cause of harm and (2) an initiative to improve handoffs by using a handoff tool embedded in an electronic health record appears to have reduced harm. Combined with many other studies about handoffs, we believe there is good reason to require some formal, structured process for all complete or permanent handoffs. There remains work to be done to provide evidence for the optimal process for handoffs and how they can be tailored to the demands of different situations.

The finding that handoffs are associated with risk and harm bolsters similar evidence and offers a counterpoint to studies suggesting that they are not. Given the strengths and limitations of the current study, we feel the net result should lead us to invoke the safety-first maxim—if there’s good but not absolute evidence that a practice is safer, then doing it should be the norm. The burden of proof shouldn’t be that something is certainly safe but rather that we can’t be sure that not doing it is safe. It was such “negative safety thinking,” demanding that engineers prove that launching Challenger in January 1986 was unsafe, that was a cause of that catastrophe; later investigation pointed to the weakened safety culture that had evolved at NASA (Washington, D.C.). Health care is no model of a high-reliability organization, but we can still strive to embody high-reliability principles like a preoccupation with error. Promoting the use of an electronic health record–based checklist as was done in the study by Saha and Segal requires relatively little time and seems to warrant whatever cost that might entail.