Editor's note: Vineet Arora, MD, MA, is Associate Program Director for the Internal Medicine Residency and Assistant Dean of Scholarship & Discovery at the Pritzker School of Medicine for the University of Chicago. Dr. Arora's research focuses on resident duty hours, patient handoffs, medical professionalism, and quality of hospital care. She also writes a popular blog, FutureDocs, focused on issues relevant to physicians in training. We asked her to speak with us about handoffs and patient safety.
Dr. Robert Wachter, Editor, AHRQ WebM&M: What got you interested in handoffs?
Dr. Vineet Arora: I was chief resident when the new ACGME [Accreditation Council for Graduate Medical Education] duty hours were announced for 2003, so I was on our team helping to implement them. From my experience as a resident, I had been the recipient of some bad handoffs and also been involved in some cases that could have gone better if the handoff was better. I think that's how a lot of us get involved in patient safety, through personal experiences. As chief resident, I knew that the residents were going to be leaving earlier in the day, essentially in the middle of the day, when we used to stay until our work was done. We used to define having a good sign-out as actually having no sign-out—where there was no work to sign out. Now we knew that the residents were going to have to leave, and balls were going to be in the air, so we would need to formalize our sign-out process and education.
RW: It sounds like a lot of people got interested in it at the time of the duty hour reductions, but handoffs aren't new. Why did medicine give this issue so little attention until 7 or 8 years ago?
VA: It's a great question. In nursing, emergency medicine, and other shift work disciplines, handoffs are a core part of the work day and tradition. For the rest of medicine, we've always emphasized continuity of care. Even in medical school, early on we talked about the doctor–patient relationship—one doctor to one patient. There's an emphasis on the allegiance to patients you take care of, so much so that, in many fields the handoff couldn't be that professional. You should always be there for your patient. The handoff was more of an afterthought. When I was a resident, I remember staying on in the hospital until midnight on a precall day, because a patient we were about to sign out around 6 PM had called my intern to say that he had the "worst headache of his life." Those are the words that put you into action, and we took him for the CT and yes, it was a subarachnoid hemorrhage. We called neurosurgery and interventional radiology. We stayed there until midnight. When I came the next morning, my intern was late because she was exhausted and we were on call that day, meaning we were there all night. On rounds, our attending said, good job, you did well. So, we were rewarded for staying late, taking care of the patient, and not handing them off. When I trained, a good handoff was no handoff. As a result, no one ever really focused on this issue until it was forced upon us, because the idea that continuity was valued over physician health was the old paradigm. As we start to value making sure that physicians are well rested and safe, the handoff comes to the forefront. We have to accept that maybe it's no longer the single doctor to single patient, but more of a team-based model of care that will work better in this system.
RW: Over the last several years, what have we learned about handoffs?
VA: The major thing we've learned is that handoffs are risky and variable. Several studies have shown that they're not optimal. Several people are starting to apply communication theory to handoffs, realizing that doctors and nurses are human and miscommunications may occur during handoffs, so we need to consider that in our teaching. We're starting to see studies examining how education can improve handoffs. Several studies look at technology or template solutions, and a few studies show technology actually can improve the outcomes of a handoff.
RW: One of the instincts in safety and quality has been to look to other fields to see how they do things that are parallel to what we're doing in medicine. What did we learn as we looked at handoffs in other industries?
VA: Other industries—such as nuclear power plants, transportation, railroad dispatch centers, and NASA—are shift-work industries where handoffs are much more ritualized and standardized. The strategies that work best are when you have the sender and the receiver in the same place, overlapping shifts, standardization of content, minimal interruptions, structured language, and read back—things that are pretty foreign to health care. Talk about just minimizing interruptions; the literature shows that health care by nature is interruption-driven. Our culture in health care is very different from those other industries that have adopted some effective strategies for handoffs. The caution with looking at other industries is that, unlike a nuclear power plant shift and obviously the famous analogy of aviation where things are more regimented and deviations are not the norm, for patient care, I would say that deviations are the norm.
It's hard to predict a patient's course, and there's this huge heterogeneity in the patients we take care of; it's sort of like flying multiple planes at once all going to different destinations, and you can't see where they're going. That's very different from flying one plane. So health care is a more complex environment, and it's sometimes harder to adapt to those strategies. Some recent work has looked at whether these strategies from other industries are actually being used in health care, and they're not. Very few people are using read back even though it's been suggested. Minimizing interruptions and even suggesting a quiet room or environmental fixes—those are very difficult to do in a health care environment.
RW: Difficult because of cultural resistance or difficult because of logistics and the workload just doesn't allow it?
VA: I should say that they're difficult in some specialties and not difficult in others … and what's very interesting is a lot of it may be related to duty hour limits. Many people do worry about handoffs with shorter shifts for residents. However, it is also true that when you define a start and an end to a shift, it allows you to have a more regimented handoff, such that the handoff is the official end of your shift and you can then apply these strategies. For example, in disciplines like emergency medicine and OBGYN that have structured shifts—often when you think about shift change there, where is it happening? It's happening face-to-face, usually in front of a whiteboard (or electronic monitor), which is actually one of the top strategies recommended in communication science for improving understanding of communication. That's very different from, for example, a medicine resident or a pediatric resident who is on an extended duty hour shift and signing over to somebody else. It's a one-on-one exchange and can be haphazard if the person they're signing out to is in the middle of a procedure, admitting, or in the middle of a busy shift. In those specialties, it's much harder to operationalize two people meeting in the same place. While the old model would be to work until your work was done and then have a very minimal handoff, the new model is probably going to involve a set start time and end time of your work day with very good handoffs.
RW: With sufficient overlap between the person coming and going?
RW: Let me ask you about different types of handoffs. Should there be a difference between the person leaving the day shift to another person who is essentially covering the service at night versus the person finishing a week or a month in a row, handing off to the next person who's going to be primarily managing those patients?
VA: In the handoff community, we're starting to standardize the way we talk about these things; we use the term "common ground." We're starting to understand that we need to risk-stratify handoffs and that not all handoffs are the same. For example, a shift change where the sender is returning is going to be very different from a service change where it's the first time that a hospitalist may be meeting the patient and there will be no opportunity to touch base with the outgoing hospitalist because they're going on vacation to Hawaii. So common ground refers to how much a priori shared knowledge did the two people have understanding each patient. If receivers have already met the patients, maybe they've covered for them the night before, then that's going to be an easier handoff. Perhaps then you can use your time to spend on the patients and the issues where you have little common ground, the new admission, for example, who the receiver has no idea of, that's going to be very different from, for example, a service change or a service transfer where maybe you're transferring a patient to the ICU. That's a very risky type of handoff—the patient is moving. They're critically unstable and the handoff is more than just temporary in the sense that the team may or may not be coming back to them. Another example would be transferring the patient to a different specialty service. So a medicine patient who's going to surgery and is going to stay on their service would be inherently risky.
RW: So we've moved away from the idea that there's one way to do handoffs. It sounds like with this risk stratification there's different expectations of the time it will take and the amount of data that gets transferred?
VA: We're starting to move into that realm. There's work looking at the relationship between common ground, the length of time of the handoff, and the quality of the handoff. Obviously, when you have more common ground, the handoff is going to be easier and probably accomplished in less time. But it's a bit early to say where that research is going. It's a philosophy that many of us who have been doing handoff research started to think about.
RW: It strikes me that there are handoffs that involve one person replacing another person but the patient is static. And another set of handoffs that involve the patient moving from place to place. Did you think about that differently and is the field thinking about those differently?
VA: In fact, that actually has been defined. In the taxonomy that I have been using, there's extra-hospital handoffs, like admission, discharge, or inter-hospital transfers (for example from a community hospital to a tertiary hospital), or intra-hospital handoffs, like shift change, service change, or service transfer.
Then in risk-stratifying all these types of care transitions, there's a series of three questions proposed in a white paper by University Health System Consortium about handoffs: (1) Is the patient physically moving? (2) Is the patient critical or unstable? (3) Is the handoff temporary or permanent? And the fourth question that I've added is, Is this the first time the receiver is hearing about a patient? That's the common ground question. I think if there is a yes to any of these questions, there's an inherent increase in your safety risk. And what are some of those types of care transitions? Well, certainly admission, which includes things like transfer from EMS [emergency medical services] to the emergency department, the emergency department up to the floor ICU, a floor patient going for urgent surgery, and discharge.
I find it fascinating because I've done work both in extra-hospital handoffs, admission and discharge, as well as intra-hospital handoffs—a lot of residents spend a lot of time creating and updating their sign-out, and reviewing that sign-out with their cross cover who's going to be covering that patient for 8 to 12 hours until they come back. Then when patients leave the hospital, sometimes there's no phone call to the primary care physician or very little investment made in trying to improve that care transition. With respect to readmissions, there's a lot of great work from the Society of Hospital Medicine and Project BOOST and others, using communication strategies around discharge. So there is a handoff occurring at discharge, and I often think just like you would communicate and sign out to your cross cover, there should be an official sign-out to primary care physicians to let them know that the patient is leaving the hospital and what information they need to know, and how, and where can they go? When we think about effective communication, often people fixate on the discharge summary, and from the communication literature we know that having a conversation is superior. You can clarify questions and be more interactive. I would say we have a long way to go in that realm because we see that discharge is still a very problematic area, particularly around the communication to primary care physicians.
RW: What makes a good handoff from person to person? For example in the hospital, someone handing off to their colleague who's going to cover their patients.
VA: A good handoff usually has three core components: interactive verbal communication, a written communication or "transition record" that's passed off, and the acknowledgement that transfer of professional responsibility took place. We often say handoffs are more than just the transfer of content, it's also the transfer of professional responsibility and the acknowledgement of that, and that verbal communication could take place ideally in person, but obviously that can't happen all the time, so at times it would happen over the phone. Ideally it's a dialogue. The key is that it's an interaction that fosters that shared mental model and common ground so there can be a conversation about the patient as opposed to just the sender dumping out information to the receiver. The written communication is sometimes referred to as the transition record, like a transfer note, a service change note, or an admission or a discharge summary.
Usually the written communication should include those items of information that may become important at a moment's notice. So if the patient became unstable or critical, what you would want to have access to on that written communication? What's the family member's contact information, what are the medications, what is that important piece? That verbal communication doesn't need to be the rehash of the entire written communication, but really focused and targeted for the inpatient handoff. What we say is to focus it on two things: (1) the tasks to be done and (2) the anticipatory guidance, what may happen over the next shift or on-coming shift and what to do about it—and that's usually phrased in the form of if–then statements. The reason we say that's the focus of the verbal communication is that we often see the problem of information overload with handoffs, where basically the sender is just reading their written communication and the receiver is not really listening or is tuning out.
Some of our newest work is starting to look at what a good listener is supposed to be doing during the handoff. With respect to the transfer of professional responsibility, that may be harder to observe. Certainly it's the sense that the incoming person, the receiver, is now going to acknowledge that they are in charge of the care of that patient, and so if they get called, they would act as if it was their patient. A recent model actually highlighted four phases of the handoff. There's a prehandoff, where the sender is organizing and updating their handoff information; the arrival, where the sender stops their patient care tasks to conduct the handoff; the dialogue, the verbal exchange between the sender and the receiver; and then the posthandoff period, where the receiver is integrating that new information while they're assuming care of the patient and then either documenting their new task or taking into account new information that's arriving. So you really sense that it's a very dynamic process, and things could go wrong at any step. By recognizing not only the three components but also the fourth step, you're able to figure out well where you can improve handoffs.
RW: People often frame the patient safety issues of duty hours as better rested residents against more handoffs and that's the safety tradeoff. What do you think has happened to the safety of handoffs since we first began cutting duty hours?
VA: Well, since we began cutting duty hours, I actually believe that the handoff has received a lot more attention. I see that handoffs are markedly better than the handoffs when I was a resident. We're starting to see change where people are really focusing on handoffs, and the residents are receiving education about it. The challenge is that more things are being handed off now of course with duty hours. Are patients worse off with respect to handoffs and duty hours? That's a really difficult question to answer objectively because the handoff research field suffers from not having a handoff-sensitive outcome that we can observe. For some, like readmission or discharge, that makes intuitive sense, but for a specific inpatient handoff, a typical patient in the hospital may have 15, 16 handoffs. So which handoff resulted in that patient going to the ICU or having an adverse event? It's very hard to define. The research hasn't really caught up there. I do believe that one of the reasons that we haven't seen this huge improvement in outcomes that we would expect to see from well-rested residents is because of the tradeoff. So we traded off fatigued residents who knew patients with well-rested residents who maybe were less familiar with the patient.
Clinical context is also very important. Sleep research has done a very good job defining what clinical tasks are most sensitive to sleep deprivation. They define these "vigilance tasks," things that require a lot of attention and vigilance like monitoring an EKG, placing an arterial line, things that are very susceptible to sleep loss. I think we need to start defining the handoff-sensitive tasks or "familiarity tasks." Where is it really important that familiarity with a patient is important? Even in the best case scenario handoff, it is unclear if we are able to equalize common ground such that the familiarity decrement is zero. My guess is that's not the case, and that if a handoff occurs, there's still going to be some loss of familiarity with the patient.
We've been actually asking people to make tradeoffs between a fatigued postcall intern who admitted a patient versus a well-rested covering intern who admitted a patient for a variety of clinical situations to get at some of these tradeoffs, and whether people recognize these tradeoffs. Our data does show that some people do recognize the tradeoff. For example, while everyone wants the well-rested intern to put in an arterial line, the results were very different for having a family meeting for an end-of-life discussion or discharging a patient. When familiarity matters, you might want to have the fatigued postcall intern who admitted a patient. As we get more savvy thinking about how to do this research, we can also think about how to use that to more effectively design duty hours that not only help residents but also are more fine-tuned for these patient safety issues. For example, perhaps there are moments where familiarity outweighs fatigue, and there are other moments where fatigue outweighs familiarity.
RW: Would the argument there be that, rather than having a hard stop at the end of the shift, for some of those context-sensitive handoffs someone might actually be allowed to stay longer? It sounds like the new ACGME rules could have some wiggle room there.
VA: Yes, that's very consistent with this idea, and as you allude to, the ACGME 2011 Common Program Requirements for duty hours now say that a resident can exceed their duty hours for the care of a single patient, but the clinical context of that case that justified staying needs to be extensively documented. But we heard this a lot, that there are times when residents wanted to stay and felt that they were doing a disservice to their patients by leaving. Again, I don't believe that this should be the norm, but rather an exception when needed. So a one-size-fits-all approach may not work at all times. We need to think about what makes the most sense for that patient at that moment.
RW: Talk about the latest ACGME revisions and whether you think they made the right calls in terms of what you know about training and safety and handoffs?
VA: I should disclose that after the IOM Report [Institute of Medicine] that helped inform the duty hours task force development of the current 2010 guidelines, I received funding from the ACGME to perform a systematic review of the literature with colleagues at Mayo and Medical College of Wisconsin that showed that there is not enough literature in real-life residency settings testing the proposed shift lengths.
This is an unfunded mandate—and we know that the cost of implementation in some scenarios could be more than a billion dollars—at a time when teaching hospitals are already having a lot of issues regarding financial viability and federal funding for residency education dollars is at risk. It is likely that tradeoffs will be made in order to comply with duty hours and maybe there will be cuts somewhere else that impact residency education or patient safety, which is a concern. For safety, it's also important to consider more than just hours, like the type and amount of work residents do. The implementation has to be better than doing the same amount of work in less time.
Apart from hours, I think one of the most interesting things about the 2011 rules is the focus on supervision. The reason I think that's interesting is because the case that was really the sentinel event for resident duty hours in New York was a lot more about supervision than about duty hours. The duty hours and residents' fatigue is what caught the attention of the public and led to the focus on duty hours today. Even Dr. Bell (who convened the Bell Commission) was quoted as saying that supervision, and not duty hours, is the problem. We've often wondered well, did supervision get lost? I think that supervision was definitely found again with this rendition, and the focus is strongly on supervision and on graduated responsibility. We've always had interns—our most inexperienced residents, pretty much right after medical school—be on call in their first year when they have a huge learning curve and they're least efficient and most susceptible to workload and sleep deprivation. There is good literature to highlight that first-year residents are more sensitive to patient safety issues. So graduated responsibilities and the focus on supervision make sense.
RW: Anything else you wanted to talk about?
VA: I would add one more thing, which is that buried in the 2011 ACGME duty hour recommendations, it does say that programs should all implement handoff education and monitoring, which is a new requirement. I think we're going to see a lot more work in how to train residents and students to do handoffs, as well as how we monitor that such education is turning into better practice and improved patient outcomes. Some of our latest work has been working on using high intensity and low intensity simulation to see whether we can make some improvements in performance as well as patient outcomes.