Handoffs and Transitions
Approach to Improving Safety
Setting of Care
Annual Perspective 2014
Despite recent efforts to promote clinical integration, the United States health care system remains highly fragmented. From its earliest days, the patient safety field identified transitions of care as an important latent condition for errors and harm. In 2014, a substantial body of new research focused on preventing harms associated with such transitions.
Handoffs are a particular type of transition of care, and they too have garnered considerable attention. The 2003 Accreditation Council for Graduate Medical Education (ACGME) restrictions on resident duty hours, further tightened in 2011, markedly increased the number of handoffs among trainees in training institutions. Interestingly, while there have always been enormous numbers of handoffs in health care delivery organizations (e.g., between nurses at shift change), the ACGME regulations served to focus attention—both within delivery organizations and among researchers—on the importance of improving handoffs of all kinds.
The ACGME is not the only regulatory and accrediting organization whose actions have raised awareness of the hazards of handoffs. In 2006, The Joint Commission mandated that all hospitals develop a standardized approach to handoff communications. Yet there is still no single approach that is recommended or required, partly because the research on best approaches has, until recently, been somewhat immature.
A similar situation has emerged regarding the broader area of improving transitions of care, particularly those associated with hospital discharge in order to reduce hospital readmissions. Over the past few years, policies by Medicare and other payers have created powerful financial incentives to reduce readmissions. However, in the absence of research identifying the best approaches to do so, delivery organizations have tried diverse approaches tailored to the culture, workforce, and patient population of particular institutions. Designing the perfect bundle of interventions remains a work in progress.
This article summarizes some of the key research published in 2014 on improving handoff communications and transitions in care.
For the past decade, the safety literature chronicled a variety of efforts to standardize the handoff process between providers in the hospital setting. The SBAR (Situation-Background-Assessment-Recommendation) and ANTICipate (Administrative data, New clinical information, Tasks, Illness severity, Contingency plans) tools offered steps for clinicians to follow as part of a standardized handoff protocol. Other structured tools took advantage of computerized and Web-based platforms. Although some of these tools were useful, none were subjected to rigorous research that captured key elements of context, clinical outcomes, and potential unanticipated consequences. This changed in 2014 with a seminal study by Starmer and colleagues published in the New England Journal of Medicine.
The tool was known as I-PASS, which stands for Illness severity, Patient summary, Action list, Situational awareness and contingency planning, and Synthesis or read-back. In early 2014, the researchers published a preliminary study that shared the development, implementation, and dissemination of the I-PASS tool and its associated curriculum. In November 2014, the same team published a multicenter study in the New England Journal of Medicine that linked the use of I-PASS to a reduction in errors.
The implementation of I-PASS in nine pediatric residency programs required adoption of the tool and guidelines for both standardized oral and written signouts. The implementation process was carefully developed and applied. It included formal training, faculty development, and a remarkable effort to engage the primary users (residents) in revising their process and workflows. The investigators prospectively studied more than 10,000 patient admissions, using active surveillance strategies to measure rates of medical errors and preventable adverse events. They discovered a 23% reduction in medical errors from the preintervention to postintervention period, accompanied by a 30% reduction in preventable adverse events. The study also found no significant changes in resident workflow, patient–family contact time, time spent by residents on computers, and time spent completing the oral handoffs. These findings represent the most significant advance in the study of handoffs to date, providing a set of best practices that should be adopted and tested in other specialties and settings.
Another 2014 study focused on the specific handoff issue created by the "morning holdover"—a patient admitted by provider(s) overnight who is transitioned to a different primary provider(s) in the morning. Here too, the ACGME resident duty hour regulations have increased the frequency of such handoffs in teaching hospitals, making it a more common safety hazard than in the past. In this prospective study using chart review and direct observation, Devlin and colleagues found that on-call trainees omitted 40% of clinically important issues during the morning handoff and failed to document 86% of those issues in the medical record. Programs with systems in which every holdover patient was reviewed sequentially in a designated location every morning were more likely to transmit the key pieces of information. Much like the impact of the I-PASS study, this study should foster careful review of practices to improve communication associated with morning holdover admissions.
Transitions of Care
The area of care transitions, particularly those associated with hospital discharge, received considerable attention in 2014. Unfortunately, the results of the studies served mostly to reinforce the notion that there is no magic bullet for preventing hospital readmissions. However, the year's literature did shed light on how to best identify patients at highest risk for readmission, a significant advance. Moreover, in light of the substantial financial penalties facing hospitals with high readmission rates, important data emerged supporting the premise that such penalties may unfairly penalize certain institutions, particularly teaching hospitals and hospitals that care for indigent patients. The review below focuses on some of the highest impact studies in this area.
First, a randomized trial by Goldman and colleagues conducted in an urban safety-net hospital tested a peridischarge nurse-led intervention to reduce readmissions for ethnically and linguistically diverse older patients. Despite the discharge support program, the investigators found no difference in readmission rates between the support program and usual care groups, and a trend toward increased emergency department visits in the intervention group. A second randomized trial by Dhalla and colleagues developed a virtual ward model of care to provide interprofessional care for community-dwelling patients identified as high risk for readmission or death when being discharged from the hospital. Once again, despite a high-intensity set of interventions that included home visits and care coordination, there were no differences between the virtual ward and usual care in readmissions or death. While past studies (particularly RED [formerly "Project Red"] and the Care Transitions Intervention) found meaningful improvements in outcomes, the Goldman and Dhalla studies highlight the need to customize certain interventions to a diverse set of patient populations and settings. To meet that goal, AHRQ published a toolkit in 2014 designed to help hospitals reduce preventable readmissions among Medicaid patients.
Moving to the challenges of medication safety in care transitions, two studies published in 2014 further reinforced the importance of medication reconciliation and simplifying complex medication regimens. In the first, by Armor and colleagues, pharmacists who saw patients soon after hospital discharge uncovered significant discrepancies between patient-reported medication regimens and those listed in their electronic health record at the time of discharge. The adverse and potential adverse drug events were most commonly due to nonadherence or underuse, untreated medical problems, and lack of therapeutic monitoring. A second study by Schoonover and colleagues found that complex regimens (as identified with the Medication Regimen Complexity Index) were associated with much higher rates of postdischarge adverse events and 30-day readmission rates. Similar to the Armor study, this one found discrepancies between hospital discharge and home medication lists in nearly 90% of cases. Both studies indicate that creating a new approach to reconciling patients' discharge medication lists and decreasing the complexity of medication regimens may go a long way toward improving postdischarge outcomes. Moreover, the studies help pinpoint patients at highest risk for adverse drug events and readmission, patients who may benefit the most from targeted interventions that reduce their risk for undesired outcomes.
Building on efforts to help identify the highest risk patients during care transitions, a new study by Glance and colleagues examined major surgical complications as a risk factor for readmission. This retrospective cohort study of nearly 150,000 admissions captured within the National Surgical Quality Improvement Program (NSQIP) registry found that 78% of patients with any postdischarge complication, 12% of patients with only in-hospital complications, and 5% of patients without any in-hospital complication experienced an unplanned readmission. Using a risk-prediction tool, patients at very high risk for major complications had 10-fold higher odds of readmission than those at very low risk for complications. Use of the NSQIP complication risk index may allow prospective identification of patients at high risk for unplanned readmission.
Given the continued challenge in demonstrating best practices associated with improving care transitions, it was also an important year to reflect upon the policy implications of the Federal Hospital Readmissions Reduction Program. While Joynt and Jha published a seminal study in 2013 demonstrating that large hospitals, teaching hospitals, and safety-net hospitals were more likely to receive payment cuts under the readmissions program, it remained unclear how best to risk adjust for the key variables. Jha argued in a 2014 commentary that adjusting quality-performance scores based on the socioeconomic status of patient populations was required to avoid simply penalizing facilities that care for the poorest patients. The National Quality Forum (NQF) released a report this past year advocating for a similar approach. It concluded, "As healthcare moves toward increasing use of financial rewards for better quality and financial penalties for worse quality, use of measures that result in incorrect conclusions about quality poses a substantial risk for penalizing healthcare organizations and providers who serve more disadvantaged populations." The NQF's recommendations to the Centers for Medicare & Medicaid Services asked for immediate action, which should stimulate continued policy debates in the coming year about the federal readmissions program.
The year 2014 was an exciting one in the area of improving handoff communications. I-PASS not only provides an evidence-based bundle of interventions to adopt and spread in other settings, it also offers proof of concept that rigorously designed, thoughtfully implemented interventions and studies are possible, even when attacking complex, multidimensional harms such as handoff and transition errors.
On the other hand, 2014 was also sobering in that, despite considerable effort, we have yet to identify an intervention or program that will predictably improve care transitions and reduce readmission rates in diverse settings. New federal policy on electronic health record requirements to enable sending and receipt of transition of care information across providers (individual and hospital) also took effect in 2014. The impact of this policy will be important to follow in upcoming years as both requirements and systems evolve. Perhaps the key message from this research is that we should stop looking for a magic bullet. Instead, the year's literature should remind us that progress will require a methodical and systematic effort to identify high-risk patients, an appreciation of the importance of context in applying interventions in different settings and for different patient populations, and a redoubling of our efforts to learn from the interventions that worked as we had hoped, as well as from those that did not.