Health Policy$ense

Patient Handoffs: A High-Stakes Game of Telephone

As a past, present, or future patient, you might hope that your health care providers are consistently and reliably involved in your care. Just like you, though, your providers need rest and time off. When they go off-duty, responsibility for you is transferred, or "handed off", to someone else.

In hospitals, this handoff process happens multiple times per day for each patient when you consider nurses, doctors, and other care providers. It's a high-stakes game of telephone. Each handoff offers colleagues an opportunity to talk about patient care, which sometimes generates new insights about management. More commonly, though, handoffs lead to vital information becoming garbled or getting lost. This misinformation can cause errors and preventable harm.

The link between handoffs and poor patient outcomes has been a topic of study for more than 15 years. Most of these studies have been small, single-center projects with limited applicability outside the settings where the studies were done. Recently, however, a well-funded multicenter trial of a handoff improvement intervention raised the bar significantly.

With their November 2014 publication in the New England Journal of Medicine, Amy Starmer of Oregon Health Sciences University and her colleagues (the "I-PASS study group") showed that improving the handoffs of pediatric physicians-in-training (residents) decreased multiple types of patient care errors in nine academic hospitals. This large study follows on the heels of a single-center study published by the same group that appeared in the Journal of the American Medical Association in 2013.

Why haven't handoffs gotten more attention before now? There are undoubtedly several answers to this question, but as a handoff researcher, I have a few ideas.

First, handoffs are conceptually simple. "It's not rocket science," said a participant in my study, Handoffs and Transitions in Critical Care, which focuses on operating room to intensive care unit handoffs. At their core, handoffs are simply conversations about patient care. As such, the idea that they are worthy of systematic investigation has taken a while to gain traction. The I-PASS study is likely to help in this regard.

Second, handoffs are complicated. Imagine the conversation you might have with a babysitter before you go out for the evening. (This is, in essence, a handoff.) You have to give the babysitter enough information to do the job while you're gone, but not so much that he or she becomes overwhelmed. The babysitter needs to know that your son is allergic to peanuts, but not that he has a play date in a week. Similarly, a physician handoff might include information about important medications and allergies, but not the tonsillectomy that a 90-year-old patient had at age five. Doing handoffs well is a skill requiring knowledge about what is important, the ability to anticipate what might happen in the near future, and the ability to effectively communicate this information to another person. As with any skill, some people are better than others. In health care, though, we have a responsibility to make sure that everyone meets some minimum standard.

A third important reason that handoffs have gotten short shrift is that they are social phenomena requiring social science techniques for adequate evaluation. As Einstein is credited with saying, "Not everything that counts can be counted, and not everything that can be counted counts." Qualitative approaches such as observation, interviewing and narrative analysis are crucial for understanding handoffs, but many clinician-researchers are not versed in these approaches. Social scientists such as anthropologists and sociologists are therefore key collaborators for studying handoffs. LDI colleagues Charles Bosk and Ross Koppel are sociologists who have worked on handoffs in medicine.  

With this multicenter, well-funded, systematic investigation of handoffs, what new information do we have? The I-PASS group found that certain error types (diagnostic errors and those related to history and physical examination) were reduced, while others (medication- and procedure-related errors) were not. This is not surprising. The pediatric residents whose handoffs were improved are still physicians-in-training, after all. We would expect them to make different types of mistakes, some related to medical knowledge, some related to skill, some to communication. A handoff improvement intervention would only affect those errors related to communication, and this is what was observed.

We also learned that handoff improvement required a bundle of interventions that included lectures, simulation sessions, reminder cards, faculty development and resident evaluation by faculty. Because the bundle was identical at all sites, we don't know what the essential components of the bundle were. We know that education-only interventions don't generally work, but are all elements of the bundle needed? Without a follow-up study dissecting the bundle, it's impossible to know. I would be surprised if this actually happens though, given the time, energy and expense involved in carrying out such a project. A likelier event is that the bundle gets picked apart in the real world with variable impact on effectiveness.

After years on the fringe, then, it appears that handoffs are coming of age in health care scholarship. As the scope and rigor of handoff research improves, we should be able to improve the reliability of communication between health care providers. We will then be able to live up to our patients' expectations of seamless coordination of care.