- Review the evidence base on erroneous actions related to shift changes.
- Understand the limits of standardizing handoffs in preventing errors at shift change.
- Explain basic assumptions about human expertise and error.
- Learn three heuristics for effective coordination: Reduce, Reveal, Focus.
A 3-month-old infant was admitted with a respiratory syncytial virus (RSV) infection to a pediatric medical unit. Although she was initially stable (without oxygen requirements), her breathing soon became labored, with an increased respiratory rate and subcostal retractions. Providers determined that she would benefit from a higher level of care and initiated the transfer process. This transfer happened to coincide with a shift change for both the nursing staff and the physicians involved.
The off-going nurse assumed that the transfer would take place immediately and signed out her patients to the next nurse before the patient was physically moved. The outgoing physician sent a text page to his incoming colleague with similar sign-out. Approximately 45 minutes later, the unit clerk called the infant's bedside nurse to report that the infant's parents believed their child was in significant distress. The nurse was surprised that the patient had not yet been transferred and, after an initial evaluation, immediately called the rapid response team. The evening physician was also contacted and was equally surprised by the series of events. She had been told that the patient was being transferred "non-emergently" to a unit for closer observation. The patient was transferred to the intensive care unit (ICU), where she made a full recovery, after a prolonged hospitalization.
The hospital performance improvement committee reviewed the case and determined that improper delegation and lack of communication between staff members contributed to the delay in treatment, and that improved policies should be required for shift changes.
Challenges at Shift Change
Shift change updates are widely believed to be a point of vulnerability in complex systems with high consequences for failure. They may also increase resilience by providing an opportunity to detect erroneous assessments and actions made by the outgoing personnel.(1-2) One study reported that sign-out–related problems with internal medicine physicians in one hospital occur at a rate of 7.5 per 100 patient-days of care.(3) Erroneous actions stemming from shift changes have caused a wide variety of problems (4), including inaccurate medication prescriptions (5-6), inaccurate evaluations (6), longer lengths of stay, and increased use of laboratory tests.(6-8) Moreover, transfers during times of reduced physician staffing may lead to lower billing accuracy, longer lengths of stay, and higher patient mortality.
This case highlights a common scenario: a planned patient transfer that gets delayed until after a shift change. The decision to delay a patient transfer is in most situations locally rational. Two off-going nurses, and in this case the outgoing physicians as well, would need to coordinate the transfer, including the update, during a busy period when they are trying to complete activities to avoid handing them off to the next practitioner; the latter is a strong social norm that reduces the chances of the "ball getting dropped." By delaying the transfer until after shift change, the receiving nurse can focus on assessing and stabilizing the patient and avoid an additional handoff update, which can potentially introduce inaccuracies and gaps in the communication exchange. In addition, transfers conducted during a shift change risk the receiving personnel having an interrupted or curtailed handoff update. These risks are traded off against the potential benefits of the additional care provided by ICU nurses and specialized monitoring technology. In the case of an unstable infant patient believed to be in significant respiratory distress, waiting 45 minutes is likely an undesirable weighting of associated risks and benefits.
Standardizing Shift Change Communication
The recommendation of the hospital's performance improvement committee to implement improved shift change policies is a laudable goal. Ideally, the policy would provide guidance to "sharp end" practitioners about how the organization would like to trade off the associated risks for this and related scenarios. To my knowledge, few existing hospital policies fully accomplish this goal. Instead of providing guidance on desired risk trade-offs in particular scenarios, settings, disciplines, and clinical categories, most policies currently recommend a "blanket" approach to standardization of content or format (e.g., Situation-Background-Assessment-Recommendation, or SBAR ). The primary benefit of this approach is to increase the information conveyed about every patient (10) in either verbal or written format. In some cases, this strategy could prove vitally important in supporting anticipatory thinking about how to best provide safe patient care. On the other hand, the opportunity costs stemming from dedicating providers' resources uniquely to the shift change update might potentially decrease the attention focused on the exchange itself, largely by "hiding" insightful deviations from typical narratives ("stories") in a series of uninformative and potentially irrelevant data pieces. For example, see Table 1 for a comparison of a narrative-based update vs. an SBAR-based update in the context of the case presented. The SBAR example is based upon the worksheet tool provided by the Institute for Healthcare Improvement, which was originally designed for nurses calling physicians to report critical situations, but is also recommended for adaptation to other updates.
Moving Beyond Standardization
Given the limits of standardization with regard to complex systems with constant trade-offs in decision-making, additional strategies may be warranted to ensure patient safety. If asked to provide guidance to a hospital improvement committee, a first step would be to synthesize what is available from decades of research in other complex, sociotechnical settings with high consequences for failure.
Basic assumptions and lessons distilled from the cognitive systems engineering literature (11-14) are outlined in Table 2. Overall, they highlight a number of important lessons that can be applied to health care settings.
For the specific case of designing policies for shift changes, a productive approach would involve efforts to reduce complexity wherever possible to minimize the potential for erroneous actions. This approach also increases the ability to detect and recover from such actions. Based on research about how to cope with complexity, there are three heuristics (15) for designing effective coordination that can serve as a general guide for shift change updates:
Heuristic 1: Reduce complexity.
- Reduce transfers by:
- Proactively "bumping" patients (e.g., moving them to another unit) in advance of a pending patient's arrival.
- Directly transferring patients to planned specialties (e.g., skip the emergency department of a second hospital).
- Minimizing "boarding" practices in emergency departments.
- Minimizing changes in providers to meet quality of worklife goals for practitioners (e.g., 14-day staffing rather than use weekend coverage; eliminate substitutions to attend meetings or give lectures).
- Assigning patients to nurses who provided care to them recently.
- Delaying tasks/breaks for nursing personnel who are providing on-call care for another nurse until the original nurse returns to the unit.
- Eliminate redundant documentation.
- Eliminate or streamline data entry for use at a later time by distant parties (e.g., billing, legal, quality improvement personnel).
- "Chunk" together related information that can be tailored to a particular perspective (e.g., group together nursing notes with vital sign information).
Heuristic 2: Reveal by rendering work observable.
Provide an overview "at a glance" visual display of the current status of a work process. An example of this approach is the Post Anesthesia Care Unit (PACU) Dashboard developed by Massachusetts General Hospital (MGH) in collaboration with LiveData (Figure).
Although every "at a glance" display will appear different based on the specific context, the information can be described abstractly in order to aid translation to other settings:
- Differences from routine assessments and plans (e.g., contact precautions).
- Relation of real-time events to planned milestones.
- Hazards (e.g., patients with a history of allergic reaction to anesthetics).
- Outliers, threshold crossings, and trends in sensored data.
- Constraints and couplings for alternatives to current plans (e.g., status of readiness of future patients).
Heuristic 3: Focus attention.
- Stratify time and attentional resources during the update based on patient instability and level of uncertainty around a diagnosis and treatment plan (e.g., "Save the sick, do the others quick" strategy ) as well as speeding up and truncating aspects of the handoff in preparation for anticipated events that will likely take precedence (e.g., "New demand, change the plan" strategy ).
- Enable peripheral detection of unexpected events and actions while performing primary tasks, such as by grouping patients under the care of a particular provider in one place and enabling peripheral monitoring via audio cues or centralized displays.
- Highlight the need to expedite tasks, including transfers (e.g., patients in significant respiratory distress being transferred to critical care units).
- Automatically compile data into printed paperwork that can be highlighted, annotated, and discussed verbally during the update.
It could be argued that the heuristics provided here do not directly address all of the issues raised by the specific case. In particular, nothing that is recommended specifically ensures that an infant patient with RSV having significant respiratory distress will avoid delays in being transferred to the ICU following a shift change handoff. In my opinion, the common and human desire to "ensure that this never happens again" following an unfortunate outcome can serve as a barrier to making health care systems safer, less costly, more effective, more efficient, and more patient-centered. In fact, I recommend avoiding basing design decisions strictly on an individual case. An alternative approach is to encourage desired trade-off weightings across a range of routine and exceptional cases that practitioners are anticipated to encounter. This approach explicitly recognizes that some amount of undesired outcomes is unavoidable in a world that does not have the benefit of hindsight knowledge about which risk trade-offs will dominate in a particular case.
- The evidence base for interventions to improve shift changes is weak.
- Alternatives to standardization of shift change update content should be pursued.
- Coordination heuristics suggest productive directions to explore.
Emily S. Patterson, PhD Research Scientist
Institute for Ergonomics
The Ohio State University
Faculty Disclosure: Dr. Patterson has declared that neither she, nor any immediate member of her family, has a financial arrangement or other relationship with the manufacturers of any commercial products discussed in this continuing medical education activity. In addition, her commentary does not include information regarding investigational or off-label use of pharmaceutical products or medical devices.
1. Cooper JB, Long CD, Newbower RS, Philip JH. Critical incidents associated with intraoperative exchanges of anesthesia personnel. Anesthesiology. 1982;56:456-461. [go to PubMed]
2. Wears RL, Perry SJ, Shapiro M, et al. Shift changes among emergency physicians: best of times, worst of times. Proceedings of the Human Factors and Ergonomics Society 47th Annual Meeting; Denver, CO: October 13-17, 2003. [Available at]
3. Horwitz LI, Moin T, Krumholz HM, Wang L, Bradley EH. Consequences of inadequate sign-out for patient care. Arch Intern Med. 2008;168:1755-1760. [go to PubMed]
4. Singh H, Thomas EJ, Petersen LA, Studdert DM. Medical errors involving trainees: a study of closed malpractice claims from 5 insurers. Arch Intern Med. 2007;167:2030-2036. [go to PubMed]
5. Bell CM, Rahimi-Darabad P, Orner AI. Discontinuity of chronic medications in patients discharged from the intensive care unit. J Gen Intern Med. 2006;21:937-941. [go to PubMed]
6. Moore C, Wisnivesky J, Williams S, McGinn T. Medical errors related to discontinuity of care from an inpatient to an outpatient setting. J Gen Intern Med. 2003;18:646-651. [go to PubMed]
7. Lofgren RP, Gottlieb D, Williams RA, Rich EC. Post-call transfer of resident responsibility: its effect on patient care. J Gen Intern Med. 1990;5:501-505. [go to PubMed]
8. Van Eaton EG, Horvath KD, Lober WB, Pellegrini CA. Organizing the transfer of patient care information: the development of a computerized resident sign-out system. Surgery. 2004;136:5-13. [go to PubMed]
9. Joint Commission on Accreditation of Healthcare Organizations. 2007 National Patient Safety Goals. [Available at]
10. Berkenstadt H, Haviv Y, Tuval A, et al. Improving handoff communications in critical care: utilizing simulation-based training toward process improvement in managing patient risk. Chest. 2008;134:158-162. [go to PubMed]
11. Patterson ES. Structuring flexibility: the potential good, bad, and ugly in standardisation of handovers. Qual Saf Health Care. 2008;17:4-5. [go to PubMed]
12. Wears RL. The error of counting "errors." Ann Emerg Med. 2008;52:502-503. [go to PubMed]
13. Patterson ES, Wears RL. Beyond "communication failure." Ann Emerg Med. 2009;53:711-712. [go to PubMed]
14. Eisenberg EM. The social construction of healthcare teams. In: Nemeth CP. Improving Healthcare Team Communication. Hampshire, UK: Ashgate Publishing;2008:9-20. ISBN: 9780754670254.
15. Patterson ES. Communication strategies from high-reliability organizations: translation is hard work. Ann Surg. 2007;245;170-172. [go to PubMed]
16. Nemeth CP, Kowalsky J, Brandwijk M, et al. Between shifts: healthcare communication in the PICU. In: Nemeth CP. Improving Healthcare Team Communication. Hampshire, UK: Ashgate Publishing;2008:135-153. ISBN: 9780754670254.
Table 1. Comparison of a Narrative vs. SBAR Update about the Patient
|Patient X in room Y is a 3-month-old girl with RSV who was initially stable but recently has been having respiratory issues that are getting worse. You need to transfer her to the ICU as soon as possible.
||For patient X in room Y, the patient's code status is
. The main problem with the patient is respiratory issues. I have assessed the patient personally, and vital signs are BP , pulse , respiration , and temperature . I am concerned about the respiration because it is . The patient's mental status is . The skin is warm and dry. The patient has been on oxygen for minutes. The oximeter is reading . The oximeter does not detect a good pulse and is giving erratic readings. This is what I think the problem is: respiratory distress. The patient seems to be unstable and may get worse; we need to do something. I request that you transfer the patient to critical care.
Table 2. Assumptions and Lessons from the Literature
|Human behavior in professional settings is locally rational.
||Attempts to increase an individual's motivation or attention to detail (e.g., try harder; be more careful) are unlikely to improve performance.
|Expertise is largely defined by context-dependent knowledge of how and when to relax constraints in over-constrained situations.
||Deviations from standard operating procedures are expected.
|All decisions require making trade-offs on competing goals.
||Sacrifices to safety goals to meet economic and production goals are common occurrences that are not usually scrutinized when the outcome is positive.
|Imposing a simple standard on a complex process does not result in simplicity.
||Policies will need to be tailored to particular settings, include contingency plans, and allow for flexibility to deviate in exceptional circumstances.
|Local actors must tailor a distant supervisor's plan to the situation.
||The intent behind mandates and policies needs to be communicated to personnel who carry them out.
|People adapt procedures over time in response to feedback.
||Policies heavily slanted toward safety goals will likely change over time in response to feedback to respond to acute production and quality of worklife pressures.
|Communication is not a (macrocognitive) function, but rather a means to achieve multiple functions in necessarily distributed work.(12-13)
||"Failure to communicate" is not a useful diagnostic concept for system improvement.
|Narratives ("stories") are the primary vehicle for compactly constructing meaning in a shared conversation, rather than a one-way transfer of independent data elements via a noisy communication channel.(14)
||Handoff procedures that emphasize comprehensive verbal communication of low-level data (e.g., vital signs) in a structured format can make it difficult to quickly understand the "gist" of a patient's status and plan.
Figure. Example of "At a Glance" Visual Display: PACU Dashboard.
Click to enlarge.
Figure reprinted with permission from Patient Safety & Quality Healthcare. In: Cieielski S, Daily B, Levine WC. A dashboard for the PACU. Patient Saf Qual Healthc (PSQH). Sept/Oct 2008;5:44-47. Available at: http://www.psqh.com/sepoct08/pacu.html.