STRAIT Simulation Training for Rapid Assessment and Improved Communication

A. SPECIFIC AIMS

Failures of communication may be the most important contributor to preventable adverse medical events1,2. The care of hospitalized patients is marked by numerous transitions in care, including handovers of patient care responsibility at changes of shift3. Poorly communicated handovers can lead to subsequent errors, care lapses, inefficiencies, and adverse outcomes4-7. Yet, few clinician trainees learn the potential risk of these transitions or the strategies to improve patient care during handovers. Our team has shown that a high-intensity simulation-based handover training and quality improvement intervention significantly improved the quality of actual handovers between physicians and nurses in post-anesthesia care units8,9. However, this intervention was time- and resource-intensive. Arora (Project Consultant) and colleagues found that a mid-intensity (and less resource intensive) handover training intervention of medical trainees improved their preparedness to give handovers10.

We now propose a comparative effectiveness trial of a high- versus a mid-intensity change of shift handover training to improve the quality of 1st-year housestaff (intern) changes in care responsibility on the adult and pediatric inpatient wards. After collecting baseline data, interns will be randomized to receive no intervention, a mid-intensity, or a high-intensity intervention. The mid-intensity intervention will include a webinar and a single simulation-based training session that focuses on how to best give handovers11,12. The high-intensity intervention will include the same webinar, more extensive simulation-based training focused on best practices as both Giver and Receiver, interpersonal skills development, and subsequent performance feedback. We will assess the effect of these interventions on the quality of observed handovers, care lapses during the cross-coverage period, and rescue events, including calls to the Rapid Response Team and unplanned transfers to intensive care. Our primary method of discerning care lapses will be the collection and analysis of Handover-related Non-Routine Events (HNRE) through facilitated interviews of handover recipients upon completion of their shifts and of the handover givers on the day they return to work. An HNRE is defined as any event possibly related to the handover that deviates from optimal care for that particular patient given their medical conditions at the time of the handover. The team has extensive experience in rating handover quality, as well as collecting NRE data from both physicians and nurses13-16.

The project has three Specific Aims and three main Hypotheses:

SA1.Compare the effects of mid- and high-intensity simulation-based handover training on the quality of interns’ change of shift handovers.

Hypothesis 1: Interns in the intervention groups will perform more effective handovers than those in the control group.

SA2.Compare the effect of the handover training interventions on the occurrence and severity of handover-related non-routine events (HNRE).

Hypothesis 2: Among observed handovers, patients of interns in the intervention groups will experience fewer and/or less severe HNRE overnight (i.e., post-handover) than will patients whose interns did not receive such training.

SA3.Assess the effect of the handover training interventions on patient rescue events.

Hypothesis 3: Across all change of shift handovers that occur on the study teams (whether or not observed), handovers that are given by interns in the intervention groups will have fewer patient rescue events during the cross-coverage period than will those of interns in the control group. Here “rescue” is defined as an intervention after an event (or complication) that is life-threatening.

Sub-Hypotheses: For each of the above hypotheses, we will examine two sub-hypotheses: A) Performance in the high-intensity group will be better than that of the mid-intensity group (or, for SA2 and SA3, patients of interns in the high-intensity group will experience fewer events than those in the mid-intensity group); and B) The intervention effects will appear early post-training and then attenuate over time as the control group improves its handover performance.

For SA1, the primary measure will be changes over time in trained observers’ ratings of actual inpatient change of shift handovers using validated measures of handover effectiveness. In the high-intensity group only, aggregate results of handover observations will be fed back to interns to reinforce desired behaviors. For SA2, we will measure the effect of the interventions on HNRE occurrence and severity. Handover observers and HNRE interviewers will be blinded to study hypotheses and to interns’ training status. For SA3, we will track all intern handovers (observed or not) and analyze the association of intervention group with certain rescue events that may be related to handover effectiveness – namely, calls to the Rapid Response Team and unexpected escalations of care (specifically, transfer to the intensive care unit).Multivariate models will be used to compare the three groups’ handover effectiveness, HNRE incidence, and rescue events over time, adjusting for important covariates including clinical discipline (medicine or pediatrics), training status of handover giver and receiver, time of year, and the acuity of patients on the day of the handover. These results will inform improvements in clinician-clinician communication, care transitions, and work schedules.

B. BACKGROUND AND SIGNIFICANCE

B.1.Failed Communication is a Major Patient Safety Problem

Contemporary views hold that the root cause of most adverse events are system factors such as dysfunctional organizational processes and faulty teamwork2,17-19. Further, discontinuities (gaps) in care processes are proposed as major causes of adverse events1. This Gap Theory is consistent with concepts about the etiology of complex system failures20-24. In complex systems, clinicians who actually care for patients are responsible for anticipating, detecting, and bridging gaps to assure safe care25. Thus, in contrast to conventional views, practitioners working in error-prone systems actually create safety. Adverse events are then a consequence of flawed processes such as deficient change of shift handovers. If this theory is correct, then efforts to improve patient safety must begin with a detailed understanding of practitioners’ work processes1,26, followed by interventions that aid clinicians in seeing and closing gaps in the process. The present project is based on this conceptual model.

Failed communication was the most frequently cited cause of medication errors reported to The Joint Commission from 1997 to 2003 and remains a root cause of many preventable adverse events3,6,27. A study in critical care found that communication failures between team members accounted for 37% of all errors reported during a 4-month period2. In research done by the PI, communication problems contributed to 20 of the 98 (20%) non-routine events reported by anesthesia providers16. In another study, communication or coordination issues played a role in about 17% of 175 actual operating room events captured during more than 1000 hours of direct observation and videotaping28. It has been proposed that effective clinician-clinician communication is essential for establishing a culture of safety29,30.

The care of hospitalized patients is marked by numerous transitions in care, including handovers of patient care responsibility at changes of shift3. For patients cared for by interns and residents, dangers posed by poor communication may be amplified since the implementation of resident duty hour restrictions in July 2003 has increased handover frequency31. Research demonstrates that change of shift handovers and periods of cross-coverage by a physician who is less familiar with the patient represent vulnerable points in hospital care during which errors and adverse events are common. In an anonymous survey study of medical interns, cross-coverage was a strong independent predictor of a preventable adverse event (Odds Ratio of 6.1; 95% CI 1.4 to 26.7)32.Specifically with regard to handovers, Arora and colleagues7 reported that 25 critical incidents, all related to resident handovers, occurred in the care of 82 patients. Further, 27% of written handovers contained discrepancies in the medication list, compared to the hospital chart. In a survey study, 59% of responding medical and surgical housestaff reported that one or more patients had been harmed in their most recent rotation due to handover-related problems; 12% reported that the harm was major5. Interns are at the greatest risk since they are clinically inexperienced (especially early in the year) and most have had no prior training in how to handover patients33.

In a just published study of more than 400 housestaff change of shift handovers, 9% resulted in an error or incident and 14% were associated with a “surprise”34. An example of a surprise is when a cross-cover intern is called about a patient who was not signed out to her. The highest risk of handover-associated error was with cross-coverage (OR=4.77 95% CI 2.42-94.1) and, if the cross-covering intern was also on call, the risk increased further (OR=6.2). In Philibert’s study, outgoing residents completed a survey just after the handover and the relieving residents completed a survey immediately after their duty shift. We will use a similar method in this proposal except a trained interviewer will administer our handover-related non-routine event survey tool, we will obtain information from the original handover giver on the day after the handover and again 2 days later (thus capturing events occurring as a result of cross-coverage), and will also track “rescue events”.

Silber and colleagues described the notion of “rescue” in their seminal study on “failure to rescue” as an indicator of hospital quality35. While a “failure to rescue” is defined as “death after a complication,” a rescue event can be defined as an intervention to prevent death after an untoward clinical development36, for example, an opioid overdosage. In this example, measures of rescue would be naloxone administration, need for airway management, or escalation of care37. In the present study (Aim 3), we will use Rapid Response Team interventions and transfers to the ICU as objective indicators of rescue events.

B.2.Basic Structure and Content of an Inpatient Change of Shift Handover

In the inpatient setting, handovers most often occur at the end of the intern or resident’s workday, when patient care responsibilities are transferred to a colleague who is on call and remains in the hospital overnight. Typically, 3 to 8 patients are handed over at a time, corresponding to the census of the departing intern. In addition to providing information about patients’ clinical course and condition, handovers often include discussion of tasks that need to be performed, such as following up on a pending laboratory test or re-evaluating a patient’s clinical condition12. Handovers consist of at least three phases: preparation (by both parties), the actual handover (e.g., interpersonal interaction), and post-handover management by the receiving clinician38. An additional middle phase, patient arrival in the new location, occurs when the handover is associated with a physical transition of care, such as transportation from the Emergency Department to hospital ward. Handovers are critically important for establishing a shared mental model around the patient’s condition. Successful handovers avoid unwarranted shifts in goals, decisions, priorities, or plans, including missing tasks or redoing ones performed by the previous clinician39.

The SBAR (Situation-Background-Assessment-Recommendation) technique, endorsed by the Institute for Healthcare Improvement (IHI),provides a structure for team communications about patients’ conditions and a standardized frame for handover-specific communications40. Although never rigorously evaluated, SBAR is purported to promote communication that allows clinicians to set expectations for continued patient care while encouraging collaboration, teamwork, and a culture of safety. In 2005, Vanderbilt University Medical Center (VUMC) adopted SBAR as its organizing theme for all patient care handovers. Therefore, in this project, SBAR’s structure and elements will be taught to all interns.

At Vanderbilt, specific core content, provided in the SBAR format, is expected to be included in every inpatient handover (see Table 1). This template and our general approach are based on the recommendations of a national Task Force of the Society of Hospital Medicine11 which was led by Drs. Kripalani and Arora. As described below in the Methods, regardless of the intervention group, all interns will receive didactics and written materials during their orientation about proper handover performance according to medical center standards. They will be instructed to prepare for handovers and that a written (generally electronic) tool will be used. Additional VUMC expectations are that an interactive verbal exchange should always occur, all patients should be included in the handover but priority should be given to ill patients, and emphasis should be given to the A(ssessment) and R(ecommendations) parts of the SBAR which, in published research, most often are neglected4,6.

Table 1. Information content of each patient’s handover in the SBAR format[§]

Situation(“SOS” – very succinct who, what, where and why)

•A Summary in One Sentence that contains the following: 1) Patient name, MRN, age, gender, weight (pediatrics); 2) Patient location; 3) Responsible team/attending physician with contact number(s); 4) Summary of current diagnosis or problem; 5) DNR status (as relevant); and 6) NPO status (as relevant)

Background(The relevant details of the patient’s history and hospital course)

•Pertinent history (including pre-admission status, initial presentation, ongoing medical problems, etc.)

•Pertinent findings and results (physical signs, laboratory tests, diagnostic studies, etc.)

•Pertinent and recent procedures or changes in therapy

•Current medications and other therapies (IV fluids, blood products, etc.)

•Allergies

•Family history and status (as relevant, especially for pediatrics)

Assessment(The patient’s current situation and their expected care trajectory over the next hours/days)

•Current diagnosis and condition including response to treatment and key issues/concerns

•Expected over-night course (predictive statements), issues, and concerns

Recommendations(What the handover receiver needs to check, do, and worry about)

•To Check (Conditions requiring clinical follow-up, test results to check, pending consults)

•To Do (Tests to order, procedures to perform, people to call)

•If–Then (What to do given different changes in condition, findings, test results, events)

Numerous qualitative descriptions of clinical handovers exist,12,41,42 and handovers between inpatient house officers have been widely studied43-46. The introduction of resident work hour restrictions and the consequent increase in shift changes increased interest in improving physician handovers, commonly by introducing technology44,47,48. The Joint Commission currently requires hospitals to implement a standardized, interactive approach to handover communications49. More recently, the Institute of Medicine has recommended that all residents receive formal education on hand-off strategies50. Nevertheless, handovers in the inpatient setting are seldom studied in a rigorous manner.

Table 2 provides a taxonomy of communication failures (modified from Weinger51 to incorporate work by Arora6 and Horwitz4) that will serve as our framework for categorization of handover-related non-routine events (HNRE). A handover-related NRE (HNRE) is any event that deviates from optimal care for a handed over patient that was in any way contributed to by a deficiency in that handover (examples of handover deficiencies or failures are provided in Table 2). Such failures are common, suggesting the merit of designing and testing interventions to reduce them.

Table 2. Taxonomy of Handover Failures with Examples (see also Appendix 7).

Type of Handover Failure Example

Failure to Prepare {Both Handover Giver and Receiver}

Fail to prepareGiver does not have most recent patient information at time of handover.

Failure to Report {Giver} or Comprehend {Receiver} Critical Content

Recent or scheduled event(s)Giver does not mention that Mr. White is NPO for surgery tomorrow.

Code statusGiver does not inform that Mr. Wu, who is unstable, is DNR.

Medication(s), other treatment(s)Medication list is not updated before handover.

Tasks to performGiver does not ask receiver to follow-up recommendations of Infectious Disease consultants, who are still rounding.

Anticipated problems &/0r guidanceGiver doesn’t note Sally Dunn may have pain, which responds well to ibuprofen.

Patient omitted from handoverMrs. Jones not included because giver discharged Mrs. Jones, though her ride won’t come for 4 more hours.

Failure to Report Work or Team Situational Factors {Giver}

Relevant team absences/issuesGiver does not report that she is off tomorrow and her resident will be covering.

Team task or goal conflictsGiver does not report that resident and attending disagree on working diagnosis.

Failure to Assure Optimal Handover Process {Both Handover Giver and Receiver}

No face-to-face communicationReceiver is eating in cafeteria, so giver quickly runs through patients by phone.

Don’t use standard documentationGiver jots a few notes on scratch paper, does not use electronic handover tool.

Provider-Related Factors {Both Handover Giver and Receiver}

Stress/workloadReceiver has 3 admissions waiting and another patient who is unstable.

Deficient judgment or knowledgeReceiver assumes that following up on Mrs. Blake’s CT can wait until tomorrow

Logistical/System Factors

Policies or proceduresDuty hour regulations require that giver leave hospital at 1 pm, before he has had a chance to follow-up on Sonia Reed’s test results.

Technology problemsPrinter is not working; fail to use electronic handover tool.

B.3.Interventions to Improve Handovers

Initiatives such as voice recorded handovers52, goal sheets53,54, whiteboards55,56, structured communication processes57,58, and interpersonal skills training9,59 have all been implemented to try to improve clinical handovers. However, the comparative effectiveness of these interventions is unclear due to conceptual inconsistencies60,61 as well, in many cases, to methodological shortcomings62,63 or lack of meaningful outcome measures. A recent systematic review of inpatient handovers (performed by Drs. Arora, Kripalani, and colleagues) found no rigorous studies evaluating the effect of such training on the quality of handover performance, nor did any controlled studies evaluate the impact of handover interventions on patient outcomes11. In the proposed study, we will directly compare two handover interventions’ effects on actual clinical behavior and patient outcomes.