Owning our off-ramps.

In this issue of JCHIMP, Meade et al., publish the results of qualitative analysis regarding a second-year rotation in tracing the outcome of discharged patients. They report that their residents develop remarkable insights into the types of failures and miscommunications that plague our discharge processes. This perspective piece places this article in the context of literature seeking to understand why these problems are endemic and how we must prioritize efforts to address and prevent them.

Imagine medical residency as a school for the crew of a car ferryboat. Their teachers carefully monitor each step of their work from loading cars and trucks, directing them into a proper slot, pulling away from the docks, navigating past rocks and sailboats, and then docking and carefully unloading. Each student learns each step theoretically. They practice the tasks. They are corrected when they make mistakes. Their teachers verify their competence as they progress through more complicated roles.

Then imagine that there is a defect in the mechanism that sets the height of the off-ramp. Only the students are never guided to study the effect of their actual off-loading process. The random foot high gap is invisible from where they stand. They wonder at the growing pile of wrecked cars mysteriously collecting on the dock. From the perspective of a teacher sitting on the ferryboat, the students work hard and perform the tasks to specification. The off-ramp looks fine and is as good as it comes.

In this issue, Meade et al., publish a qualitative review of a rotation where residents trace patients from acute to post-acute settings. This program captures a significant learning experience. The residents are reporting back that the landing ramp is unpredictably off the dock. With a simple 2-week intervention, exposing their second-year residents to a checklist of typical care transition failures, they are triggering transformative reflection. The residents independently report care fragmentation. They report an ‘Aha moment’ as they discover the quantity of errors and non-communicative discharge summaries. They imagine how the transition could be better managed. They become aware of their responsibility and where to focus their attention. They become highly engaged actors in changing the system. Every program director should want to reliably recreate these lessons.

As much as a report on the psychological impact of a helpful educational innovation, though, this paper serves another function. It underlines just how vulnerable our discharge systems are. Why is it so easy for a second-year resident to find these failures? Why are our first-year residents and their supervisors complacent? What do we need to do to ensure that a first-week intern understands the risks their patients are facing; that they have clear standards to follow at discharge and ensure sufficient supervision for effectiveness? How can we better understand and own our discharge off-ramps?

The heart of graduate training is the feedback loop. As a trainee applies knowledge and skills, the supervisor monitors and corrects. Within days of starting, interns rapidly improve their performance responding to the corrections. The improvement in care, seen for admissions and daily visits, although more or less deliberate in different institutions, appears to develop intrinsically in daily rounds. In contrast, when discharging a patient where is the feedback? Part of the problem is the complexity of any communication in post-acute care management. Kripalani (1, 2) surveyed the literature in 2007 and found extensive evidence for breakdowns in discharge communication. From 66 to 88% of primary care physicians reported caring for patients without detailed discharge information. Between 45–75% of patients were reported as having tests that were pending at discharge, but almost 60% of primary providers were unaware of results. Practically, very few of these failures are communicated back to the hospital team. Medicare’s financial penalties for readmission may change this by creating an incentive for improving the feedback loop. So far, this has been elusive. A recent review of a bundle of interventions (3, 4) in reducing readmission in 14 community hospitals was remarkably discouraging. They modestly reduced readmissions but only in proportion to a reduction in total hospitalizations – suggesting an impact on preventing unnecessary admissions rather than truly improving the discharge process.

Perhaps most optimistic are the results of the Society for Hospital Medicine’s Better Outcomes by Optimizing Safe Transitions (BOOST) project (4, 5). They hypothesized that the poor results of previous trials have resulted from the interaction between the local context of variability among providers, patients, and resources, and the universal resistance to change. In other words, because the pattern of discharge failures is specific to each institution, no universal intervention can be effective other than having each system study itself and create effective feedback loops. Their implementation focused on the practical advice needed to engage local staff in identifying needs and measuring the impact. They provided a menu of evidence-based tools and education materials, and a mentoring expert to help sustain the process. On average their institutions only used 3 of 12 potential interventions, and yet they achieved a 2% reduction in readmissions. Interestingly, BOOST observed no difference between academic and non-academic institutions, confirming that discharge failures appear to be a universal constant, rather than simply a training failure.

By owning the discharge process, we step away from the absurdity of not seeing the elevated off-ramp and not seeing a connection to the wreckage. BOOST is showing us that we can do a lot more to anticipate discharge problems, respond to them, and learn from them. Some hope that the focus on readmission will ultimately unmask the degree of dysfunction in our outpatient systems (6). Hospital systems with a BOOST-like team will begin to close the loop, providing data to our discharging teams about what our patients are experiencing. We will be unable to ignore the fragmentation and will learn to provide more seamless care. Of course, we are not responsible for every conceivable error downstream. Just as our metaphorical ferry crew has limited responsibility for confusing signs at the ferry terminal, or pot holes in a highway ramp, our readmission committees need help grappling with outpatient care. Nevertheless, we can do more to professionalize the transition. In admitting a patient, we fully understand that tracking every patient’s outcome is crucial for care and for education. No less, we need to emphasize the same with every discharge.