Understanding and Improving Patient Handoffs · tion, higher costs, longer hospital stays, more...

www.jcrinc.com

Understanding andImproving PatientHandoffsFeaturesEditorial

■ Handoff Improvement: We Need to Understand What We Are Tryingto Fix

Continuity of Care

■ Patient Handoffs: Standardized and Reliable Measurement ToolsRemain Elusive

■ The Veterans Affairs Shift Change Physician-to-Physician HandoffProject

Information Technology

■ Improved Physician Work Flow After Integrating Sign-out Notes intothe Electronic Medical Record

Medication Safety

■ Applying Lean Sigma Solutions to Mistake-Proof the ChemotherapyPreparation Process

DepartmentsNational Patient Safety Goals

■ A Checklist to Identify Inpatient Suicide Hazards in Veterans AffairsHospitals

Letter to the Editor

■ On Using Statistical Process Control Charts to Analyze the Impact ofQuality Improvement Interventions

February 2010Volume 36 Number 2

Improvement fromFront Office to Front Line

“The wide diversity of handoff quality measures

suggests a lack of consensusabout the primary purpose of

handoffs and how best to intervene to improve handoff processes.”

—Patterson and Wears (page 59)

The Joint Commission Journal on Quality and Patient Safety

Handoff Improvement: We Need to Understand What We AreTrying to Fix

Editorial

Too often in the hospital, we hear that important informationor tasks “just fell through the cracks” after a transfer of patient

care. Yet this analogy is wrong. The problem with transfer of patientcare is that it is not a clear pathway with some dangerous cracks thatneed fixing. Patient handoffs in medicine are astonishingly vari-able,1,2 with no prespecified purpose or structure other than to“maintain continuity of care” or to provide enough input to thenext care area to support a good outcome. Most residents are nottrained in handoffs and learn the technique by modeling equallyunskilled peers. It’s no wonder the process is called a “peculiar ritu-al”3—unscientific repetitive behavior with no guaranteed outcome.

In this issue of The Joint Commission Journal on Quality andPatient Safety, Patterson and Wears expose this uncertainty andcomplexity in a review that explores all the objectives that inpa-tient handoffs could accomplish.4 These researchers, mindful ofclinical needs, group these functions into seven categories, or fram-ings, and suggest outcome measures for each framing. Those inter-ested in improving handoffs will realize after reading the articlethat we cannot simply “improve handoffs” with a checklist or sometraining. We must instead review our handoff processes. Relevantimprovement will only come after understanding that differentclinical scenarios require different handoffs. Each handoff must bereviewed, dominant framings discovered, and changes thoughtful-ly designed to fit work flow and be measurable. Several of theauthors’ measures are things that care providers do or rememberduring busy clinical work. Those who use this review’s importantframework in a careful handoff redesign must beware of focusingtoo many interventions on the workers at the sharp end while fail-ing to overhaul handoff conditions and supporting tools that arecreated in the organization’s upper echelons—where more lastingimprovements ought to be made.5,6

An example of one such supporting tool is described in the arti-cle by Anderson et al., also in this issue.7 This software applicationprovides a readable, reliable, and complete, yet not overwhelming,supply of information from an outgoing care provider to theincoming replacement. The authors helpfully mention, in passing,an early failure during testing: “significant problems with func-tionality (that is, difficulties forming the handoff list, propagatingthe DNR [do-not-resuscitate] orders, printing).”7(p.64) However, theproblem is of critical importance. Institutions must remember thatdesigners and implementers do not view electronic systems as clin-

ical users do. Top-down deployment that doesn’t consider real-time work flow results in work-arounds, frustration, and failure.8

The development group wisely stopped the project to reassess andimprove the application before the project became irrecoverable.9

A second key message is the warning about “commission of infor-mation.” Including irrelevant information during a handoff con-verts what should be a helpful quick-reference summary into anunhelpful, overdense recapitulation of progress notes. As othershave commented, merely applying structure to information doesnot improve communication or care quality.10 Finally, althoughnot statistically significant, 50% more survey respondents saidafter implementation that handoffs occurred in a quiet place. It isunknown from this report whether sites deployed the software toolin conjunction with a handoff redesign that included otherimprovements—which, for real progress, as Patterson and Wearswould say, would be re quired. Applying computerized automationto a broken or ill-defined process simply results in a broken com-puterized process.11

References1. Horwitz L.I., et al.: What are covering doctors told about their patients? Analysis ofsign-out among internal medicine house staff. Qual Saf Health Care 18:248–255, Aug.2009.2. Arora V., et al.: Medication discrepancies in resident sign-outs and their potential toharm. J Gen Intern Med 22:1751–1755, Dec. 2007.3. Mukherjee S.: A precarious exchange. N Engl J Med 351:1822–1824, Oct. 28,2004. 4. Patterson E.S., Wears R.L.: Patient handoffs: Standardized and reliable measure-ment tools remain elusive. Jt Comm J Qual Patient Saf 36:52–61, Feb. 2010. 5. Reason J.: Understanding adverse events: Human factors. Qual Health Care4:80–89, Jun. 1995. 6. Sutcliffe K.M., Lewton E., Rosenthal M.M.: Communication failures: An insidiouscontributor to medical mishaps. Acad Med 79:186–194, Feb. 2004.7. Anderson J., et al.: The Veterans Affairs shift change physician-to-physician hand-off project. Jt Comm J Qual Patient Saf 36:62–71, Feb. 2010. 8. Nikula R.E.: Why implementing EPR’s does not bring about organizationalchanges: A qualitative approach. Stud Health Technol Inform 84(pt. 1):666–669, 2001.9. Charette R.N.: Why software fails. IEEE Spectrum, Sep. 2005.http://spectrum.ieee.org/computing/software/why-software-fails/ (last accessed Dec.15, 2009). 10. Murphy A.G., Wears R.L.: The medium is the message: Communication andpower in sign-outs. Ann Emerg Med 54:379–380, Sep. 2009. Epub Apr. 11, 2009.11. Cohen P., Sayer B.: Avoiding software development failure. HCi Journal, 2001.http://www.hci.com.au/hcisite2/journal/Avoiding%20software%20development%20failure.htm (last accessed Dec. 14, 2009).

J

Erik Van Eaton, M.D.

51February 2010 Volume 36 Number 2

Erik Van Eaton, M.D., is Assistant Professor of Surgery and Surgical

Critical Care, University of Washington and Harborview Medical

Centers, Seattle. Please address correspondence to Erik Van Eaton,

[email protected].

Copyright 2010 Joint Commission on Accreditation of Healthcare Organizations

February 2010 Volume 36 Number 2


52

Patient Handoffs: Standardized and Reliable Measurement Tools Remain Elusive

Continuity of Care

Numerous quality improvement projects on patient hand-offs have been conducted in the last decade. In part, these

pro jects have been motivated by hospital and training programaccreditation requirements in the United States. The JointCommission’s 2006 National Patient Safety Goal, whichrequired “a standardized approach to hand-off communica-tions,” is now scored as a standard.1* In addition, the Accredita -tion Council for Graduate Medical Education (ACGME)’slimit on resident physician duty hours,2 effective July 1, 2007,has had the effect of increasing physician handoffs forpatients.3–6 In addition, the World Health Organization haspublished a Patient Safety Solution for ‘CommunicationDuring Patient Handovers’,7 and the Society of HospitalMedicine has endorsed a set of recommendations for hospital-ist handoffs.8

There is a growing awareness that high-quality handoffprocesses are critical to providing safe and effective patientcare.9–15 Approximately 20%–30% of information conveyedduring handoff updates is not documented in the medicalrecord.16,17 Impacts of less-than-ideal handoffs likely includeadverse events,18 delays in medical diagnosis and treatment,redundant communications, redundant activities such as addi-tional procedures and tests, lower provider and patient satisfac-tion, higher costs, longer hospital stays, more hospitaladmissions, and less effective training for health careproviders.19 Several observational, focus group, and survey stud-ies have confirmed that the baseline condition for patient hand-off processes is highly variable in quality and structure.20–28

Given the dramatic increase in activity to improve patienthandoff processes, it is surprising that standardized, reliablemeasurement tools remain elusive.29 We believe that the widediversity of quality measures in the handoff literature suggests

Article-at-a-Glance

Background: Numerous quality improvement projectson patient handoffs have been conducted, yet standardized,reliable measurement tools remain elusive.Handoff Quality Measures Classified by Primary

Handoff Purpose: The literature review, which yieldedapproximately 400 relevant articles, led to the identificationof seven primary functions for patient handoffs, each ofwhich implies different interventions to improve them: (1)Framing 1, information processing is the most prevalent inthe patient handoff literature; (2) Framing 2, stereotypicalnarratives, emphasizes highlighting deviations from typicalnarratives, such as a patient who is allergic to the preferredantibiotic for treating his or her diagnosed condition; (3)Framing 3, resilience, takes advantage of the transparencyof the thought processes revealed through the conversationto identify erroneous assumptions and actions; (4) Framing4, accountability, emphasizes the transfer of responsibilityand authority; (5) Framing 5, social interaction, considersthe perspective of the participants in the exchange; (6)Framing 6, distributed cognition, addresses how a transferto a new care provider affects a network of specialized prac-titioners performing dedicated roles who may or may notbe transitioning at the same time; (7) Framing 7, culturalnorms, relates to how group values (instantiated as socialnorms for acceptable behavior) in an organization or subor-ganization are negotiated and maintained over time. Discussion: The diversity of handoff measurementapproaches suggests a lack of consensus about the primarypurpose of a handoff, as well as about what interventionsare most promising for improving handoff processes.Recognizing that there are simultaneously multiple purposes for handoffs is a critical precursor to qualityimprovement.

Emily S. Patterson, Ph.D.; Robert L. Wears, M.D., M.S.

* For all accreditation programs, the requirement is now Element of Performance 2

(“The hospital’s process for hand-off communications provides for the opportunity

for discussion between the giver and receiver of patient information”) for Standard

PC.02.02.01 (“The hospital coordinates the patient’s care, treatment, and services

based on the patient’s needs”).




confusion and disagreement about what patient handoffs areintended to accomplish. To aid health care personnel taskedwith measuring the impact of changes to patient handoff pro-cedures, we have classified the literature on possible handoffmeasures by seven primary purposes of a handoff. We discusshow these different definitions suggest different types of inter-ventions to improve handoffs, and specifically counsel againstinterventions based solely on the dominant conceptualizationof a patient handoff as an information processing task.

Methods The literature review was conducted in three waves, primarilyin October 2008, with small reviews conducted in October2009 and December 2009 to find recent publications, in elec-tronic databases (PubMED, Google Scholar) using searchterms, using forward and backward citation searches from keyarticles, and by requesting papers from researchers.Approximately 400 relevant articles were identified. Relatedterms discovered during this review included handoffs, han-dovers, sign-outs, sign-overs, turnovers, intershift transfers, inter-shift handovers, shift change transfers, patient transfers, transitionsof care, transfers of care, substitutions, bedside reports, shift reports,shift-to-shift communications, shift-to-shift reports, discharges, dis-charge communiqués, discharge summaries, discharge notes, post-operation updates, interdisciplinary transfers, multiprofessionalhandovers, and admissions. Existing literature reviews discoveredduring the search process30–35 provided many of the referencedcitations.

Handoff: A Working DefinitionOur working definition for a patient handoff is as follows: “Theprocess of transferring primary authority and responsibility forproviding clinical care to a patient from one departing caregiv-er to one oncoming caregiver.” Caregivers include attendingphysicians, resident physicians, physician assistants, nurse prac-titioners, registered nurses, and licensed practitioner nurses. Itis assumed that the patient handoff is conducted between care-givers at comparable levels of experience and expertise and whoare at equivalent levels of a hierarchy (for example, attendingphysician to attending physician). The scope of this article doesnot include handoffs across differential levels of staffing, suchas from a primary caregiver to an on-call provider; handoffs forshort periods of time (for temporary relief of nursing coverageduring a 15-minute break); handoffs across specialties (anesthe-siologist report to postanesthesia recovery room nurse); hand-offs across settings (from the emergency department [ED] tothe intensive care unit [ICU]); or handoffs between profession-

al health care providers and family caregivers (discharging froma hospital setting to an in-home setting).

Examples of handoffs using this definition are as follows:■ Nursing shift changes■ Physician sign-outs■ Physician-to-physician transfers during a tour of duty to

balance workload■ Nurse-to-nurse transfers during a shift to balance work-

loadFollowing a handoff, the oncoming caregiver assumes the

responsibility for providing care, as defined in Table 1 (page54).

Handoff Quality Measures Classified byPrimary Handoff Purposes On the basis of the literature review, we have identified sevenprimary framings for patient handoffs, each of which has a pri-mary function that implies different interventions to improvehandoffs (Table 2, page 55). This framework is based on oursynthesis of the existing literature, is informed by others’ dis-tinctions,31,34,36–38 and extends our previous framework.39–40 Wehave limited our discussion of handoff functions to one perconceptual frame, although it is likely that there are multiplefunctions for some of the framings. We have classified thehandoff mea sures into these seven functions (Table 3, page 55).

FRAMING 1: INFORMATION PROCESSING

The first framing, information processing, is the most preva-lent in the patient handoff literature. The Joint Commission, indefining the primary objective for a handoff, used this framing,as follows:

The primary objective of a “hand off ” is to provide accurateinformation about a [patient’s] care, treatment, and services,current condition and any recent or anticipated changes.The information communicated during a hand off must beaccurate in order to meet [patient] safety goals.41(p. 31)

The conceptual framing is based on an information process-ing metaphor that has dominated scientific thought for decadesand is represented as a sequence of four mental operations: (1)encoding, (2) comparison, (3) response selection, and (4)response execution. These operations occur inbetween inputs(stimuli) to an individual and outputs (responses) from an indi-vidual and mediate the use of memory and attentionalresources. With this frame, the primary function of the hand-off is to transfer data through a noisy communication channel(that gets noisier with background noise, interruptions, infor-


54 February 2010 Volume 36 Number 2


mation overload, ambiguous language, speech and hearingimpairments, and cultural differences). The primary risk is thatclinical judgments will be made with missing or inaccuratedata.

As indicated in Table 2, interventions to define and stan-dardize essential information are implied by this frame, as areclosed-loop verification techniques for communicating criticalinformation.8 An example of a closed-loop technique is a read-back, which was first required by a 2003 Joint CommissionNational Patient Safety Goal but was recently applied to a stan-dard.1* A variation is to supplement a verbal interaction with apaper or electronic summary that is written by the departingprovider.4,42 Some interventions reduce the data entry burdenby automatically “pulling” data from electronic medicalrecords.43–45 One study suggests that documentation in thepatient chart is reduced when additional handoff paperwork is

employed.46 This finding suggests including handoff summariesin the official patient chart to avoid losing critical information.On the other hand, there is reason to believe that doing this willencourage the creation of a “shadow chart” for handoffs becauseshort-term, sensitive information that is helpful for the transi-tion of care might not be appropriate for a long-term reposito-ry, particularly if it increases risks to professional reputation orliability risks.47

Interventions to minimize interruptions48 are usually basedon this conceptual framing in that reduced interruptions arebelieved to improve information transmission by reducing thebackground noise in a communication channel and also byreducing memory loss of both the departing and oncomingpractitioner. Interventions include conducting handoffs in aquiet, dedicated space away from main traffic area49 and usingnoninterruptive communication strategies, such as dedicatedpagers for noncritical patient requests or delegating answeringthe phone to a clerk, during the handoff update.50,51

As indicated in Table 3, the primary measurement approachwith this framing is to compare the content of the handoffupdate (based on a verbal transcript or analysis of handwrittenor electronic paperwork) against a “gold standard” of essentialcontent.52,53 In addition, distinctions are made between infor-mation transmitted by the departing provider during theupdate and retained information by the oncoming providerafter the update.54,55 Extensions of this approach include weigh-ing some items as more important than others and/or separate-ly reporting critical information and optional information. Forexample, one article recommends splitting information provid-ed by emergency medical service to ED personnel during ahandoff update into (1) information that is essential to stabilizeand initially diagnose the patient and (2) information that isneeded to treat the patient over the long term.56

An extension of this measurement approach is to ratewhether information has been organized in a particular fash-ion,32 such as that based on a Situation-Background-Assessment-Recommendation (SBAR) ordering.57,58 Alternativeordering schemes include physical locations in beds and patientordering (often alphabetic by last name) on information tech-nology systems, by “most important first,”59 by body system, byhead-to-toe ordering, or by patient problems.60

This measurement approach can be further categorized intoessential content that is designed to apply across a wide rangeof settings58 versus content that is tailored to a particular set-ting, such as from an anesthesiologist in the operating room toa nurse in the postanesthesia care unit61 or from a surgicalphysician to a surgical physician.43 In defining what is essential,

* Element of Performance 20, “Before taking action on a verbal order or verbal

report of a test result, staff uses a record and ‘read back’ process to verify the infor-

mation,” for Standard PC.02.01.03, “The hospital provides care, treatment, and

services as ordered or prescribed, and in accordance with law and regulation.”

1. Performing technical work competently

2. Knowing the historical narrative (relevant patient history and

chief complaint)

3. Being aware of significant data or events

4. Knowing what data are important for monitoring changes and

their associated levels of uncertainty

5. Managing impacts from previous events

6. Anticipating future events

7. Weighing trade-offs if diagnostic or therapeutic judgments need

to be reconsidered

8. Planning patient care strategies

9. Performing planned tasks

Other responsibilities following the handoff may include the

following:

10. Alerting others to the completion of interdependent tasks

11. Supervising junior personnel and/or accessing senior

personnel,

12. Identifying warranted and avoiding unwarranted shifts in goals,

priorities, plans, decisions, or stances towards key decisions

13. Relationship building with peers and other social functions

14. Articulating and reinforcing group values

15. Protective functions, including supporting other caregivers who

are grieving for dying patients

16. Involving patients and their family caregivers in decision

making

*Adapted from Patterson E.S., Woods, D.D.: Shift changes, updates, and

the on-call model in space shuttle mission control. Comput Support CoopWork 10(3–4):317–346, 2001.

Table 1. Oncoming Caregiver’s Responsibilities for CareFollowing a Handoff*




there are challenges in striking a balance between comprehen-siveness of information and directing attention toward particu-larly salient issues (which can be easily missed under dataoverload conditions and/or when a standardized format makesit difficult to highlight unusual information). Additional trade-offs include minimizing the work to tailor a list to a particularsetting, coordinating with other hospitals on standards, meet-ing the needs of different disciplines, allowing flexibility forcontingencies, and maintaining lists over time, in particularremoving items from lists to avoid having the lists becomeunwieldy (a common pattern is to frequently add and rarelyremove handoff items).

An alternative approach is to measure content omissions62 or

recalled information.63 A related observational measurementapproach that is more subjective but reduces analytic time andthat can be done in real time by a single observer is to rate con-veyed content on an acceptability scale or a scale that qualita-tively estimates the level of structure in the update.64 Similarapproaches that use self-report from handoff participants askabout satisfaction with handoffs,65 perceived quality,14 whethera transfer was suboptimal66 or poor,67 whether a problem49 orcritical incident could be attributed to an inadequate transfer,68

whether one or more patients experienced harm from problem-atic handoffs in a physician’s most recent rotation,49 andwhether there were any surprises during the subsequentshift.11,13,44

Conceptual Frame Primary Function Intervention Example

1. Information Processing Transfer data through a noisy communication Standardized handoff protocol

channel

2. Stereotypical Narratives Label by stereotypical narrative and highlight Daily goals for interdisciplinary teams

deviations

3. Resilience Cross-check assumptions with a fresh Two-challenge rule for resident physicians questioning

perspective attending physicians

4. Accountability Transfer of responsibility and authority Handover protocol that explicitly assigns tasks to team

members

5. Social Interaction Co-construction of shared meaning Supporting interdisciplinary team communications dur-

ing rounds

6. Distributed Cognition Replace a member of a network of Shared repository artifact (e.g., whiteboard) for aiding

specialized practitioners coordination between caregivers

7. Cultural Norms Negotiate and share group values Guided reflection on handoff improvements during

resident orientation

Table 2. Alternative Functions of a Handoff

Conceptual Frame Quality Measures Examples

1. Information Processing Accurate essential content transferred Information units transferred

Information omissions

2. Stereotypical Narratives Appropriate patient narrative Level of information abstraction

Insightful summary synthesis

3. Resilience Collaborative cross-checking (accuracy of diagnosis, Risk-adjusted mortality and morbidity

quality of treatment plan) Preventable adverse events

4. Accountability Task completion Dropped patients

Inappropriate tasks transferred

5. Social Interaction Respectful interactions, team climate Interprofessional communication quality

Generation of new insights

6. Distributed Cognition Effective coordination of care Technical errors

7. Cultural Norms Educational interventions Comfort with doing handoff

Policies and procedures Number and quality of implemented process

Changes in priorities, values, acceptable behaviors changes

Table 3. Examples of Quality Measures for Handoff Functions




Outcome measures could theoretically be employed withany conceptual framing. Because information processing is thedominant frame in the handoff literature, most of the attemptsto measure outcomes have been done with this framing. Todate, most of the studies have compared naturally occurring “Aversus B” groups rather than randomly selected groups orpre–post comparisons. Examples include additional transfersto a cross-covering physician,69 coverage by a physician fromanother team,9 short-call and cross-coverage,5 and informationtechnology support versus no information technology supportfor handoffs.70 Patient outcome measures include risk-adjustedmortality and morbidity,71 cardiac arrests, preventable adverseevents,70 length of stay (LOS), patient satisfaction, malpracticeclaims, timeliness of care, timeliness of disposition to hospital,boarding times in the ED, staff retention, staff satisfaction withquality of care,65 patient satisfaction, unplanned transitionsfrom acute care hospital wards to the ICU within 24 hours ofa handoff,4 and staff time.72 Note that few studies have specifi-cally addressed the impact of hindsight knowledge bias on out-come judgments, which may reduce the validity of thesemeasures.73

FRAMING 2: STEREOTYPICAL NARRATIVES

The second framing, stereotypical narratives, takes advan-tage of the narrative structure to quickly and effectively com-municate large amounts of information by associating theinformation with default patient narratives.74,75 With this fram-ing, the emphasis during the handoff is on highlighting devia-tions from typical narratives, such as a patient who is allergicto the preferred antibiotic for treating his or her diagnosed con-dition.

A fictional example of a handoff update from a departing toan oncoming nurse is provided in Sidebar 1 (right). An exam-ple of a stereotypical narrative for this patient is “A post-menopausal woman with late-stage breast cancer who mostlyneeds help with self-care activities.” Deviations from thisstereotypical narrative for this patient include “The patient hasa cardiac history that affects her speech ability, has recently losta lot of weight due to frequent vomiting, is at risk of fallingwhen she is mobile, and frequently asks nurses for aid.”

The Pronovost et al.76 daily goals intervention for interdisci-plinary teams in the ICU uses typical patient narratives anddeviations in a prominent fashion. A similar approach would beto ask an oncoming provider to summarize what was learnedfrom the handoff update and provide an outline of a plan toeither the departing provider or to a supervisor, immediatelyfollowing which the departing person is encouraged to make

any clarifications or provide additional information. Measures for this framing include the accuracy and com-

pleteness of retained information about relevant patient historyand chief complaint54 or the level of synthesis and abstractionof the provided information; one measure is the Recognition-Primed Abstract Decomposition Space (RP-ADS), which wasused by Miller and colleagues to find that nurse handoffsfocused more on data and intervention levels, whereas physi-cian handoffs focused more on diagnoses and expectations.77

FRAMING 3: RESILIENCE

The third framing, resilience, involves taking advantage ofthe transparency of the thought processes revealed through theconversation to identify erroneous assumptions and actions.78,79

For example, during a (fictional) face-to-face nursing handoffupdate at the conclusion of a shift, the departing nurse says,“She is able to get up and go ad lib to the bathroom.” Then theoncoming nurse says, “No, that’s not right. I took care of that

Mrs. X is a new admit that came up around 11:00 last night. She is

a 62-year-old female patient of Dr. Y in the hematology/ oncology

service with weight loss, weakness, and breast cancer. She has

had a lump mastectomy in the past. She said that she lost a lot of

weight, about 40 pounds, over the last course of last month, and

just can’t keep anything down. A very weak-looking lady there. But

she’s able to move around with some minimal assistance. She

does have a history of a left CVA [cerebral vascular accident] and

some aphasia problems. It’s real difficult for her to swallow any-

thing. You have to crush her pills. Her biggest complaint is consti-

pation. I gave her Lactulose. She also had Colace. I took care of it

as far as crushing it up. She hasn’t taken anything for me.

However, she has had about three episodes of vomiting through-

out the night. I did go ahead about 4:30 A.M. [04:30] and gave her

some Phenergan IV [intravenous] push on that. She has a 22

gauge in her right hand and she’s getting D5 and a half normal

saline at 100 an hour. She’s got some urines on the board there

that I wasn’t able to get. I put her on the bedpan and 5 minutes

after I take her off, she’d go ahead and void or have a stool there.

So if you could follow up on that. I have done nothing but been in

that room all night, putting her on the bed pan, taking her off, put-

ting her back on, taking her off, so she’ll kind of wear your

patience thin a little bit, so just kind of bear with her. She’s on a

clear liquid diet, like I said, she’s not really tolerating anything.

She’s supposed to be put on an airflow mattress today and also

have a nutrition consult done and hopefully those will be in the

computer on that. Labs were done on her and other than her just

constant wants and needs, nothing is really going on with her.”

Sidebar 1. Fictional Example of a Handoff Update from aRegistered Nurse on the Night Shift to a Registered

Nurse on the Day Shift




patient yesterday. The physician ordered her to be on strict bedrest because she tried that by herself and nearly fell downbecause she was so weak.” The departing nurse then says, “Oh,I didn’t know that. Now that I think about it, I don’t think shewent to the bathroom all night, but be sure to let the familyknow that today.” Even when errors are not detected, clarifyingquestions can be asked to confirm an understanding, such asfor the oncoming nurse to ask, “So the plan is to discharge thismorning?” and for the departing nurse to respond, “No, firsthis blood pressure has to come down, so you’re probably look-ing at late afternoon at the earliest.”

Interventions to institute check-out procedures that providethe opportunity for the oncoming provider to ask clarificationand error-detection questions, even when the handoff update isaudiotaped or handwritten, is an example with this frame.Although not specific to handoffs, the two-challenge rule forresident physicians questioning attending physicians thatencourages error-detection questioning strategies is a relatedconcept.80 In army aviation, the rule is defined as follows:

The two-challenge rule allows one crew member to auto-matically assume the duties of another crew member whofails to respond to two consecutive challenges. For example,the pilot-on-the-controls becomes fixated, confused, taskoverloaded or otherwise allows the aircraft to enter an unsafeposition or attitude. The pilot-not-on-the-controls first asksthe pilot-on-the-controls if he is aware of the aircraft posi-tion or attitude. If the pilot-on-the-controls does notacknowledge this challenge, the pilot-not-on-the-controlsissues a second challenge. If the pilot-on-the-controls fails toacknowledge the second challenge, the pilot-not-on-the-controls assumes control of the aircraft.81(p. 15)

Measures for this conceptual frame are the number of clari-fication and error-detection questions posed by the oncomingprovider during a handoff82 (and potentially evidence of othercollaborative cross-checking strategies used in high-risk indus-tries such as having oncoming providers initiate new topics and“overhearing” by interdisciplinary colleagues to identify mis-conceptions78) and increases in patient harm. For patient harm,the same outcome measures described for Framing 1, informa-tion processing, would be used. The main difference is thatrather than assuming an increase in harm due to lost informa-tion during the transition,13 there is hypothesized to be adecrease in harm because of the opportunity for the oncomingprovider to critique the departing provider’s diagnosis andtreatment plan, which could allow earlier detection and recov-ery from errors.83

FRAMING 4: ACCOUNTABILITY

The fourth framing, accountability, emphasizes the transferof responsibility and authority that distinguishes a handoffupdate from an information update. Interventions to ensurethat patients are assigned to providers, that providers are awareand have accepted the transfer,84 and that others are aware ofwho is responsible for a patient and reminders to completetasks that have been handed off, such as checking on laborato-ry results, relate to this framing. One redesign of a handoffinstantiated in a protocol format62 included explicitly assigningspecific tasks to team members, such as assigning ventilation tothe anesthesiologist.

Measures for this frame include the number of droppedpatients (defined as patients who are assigned to departingproviders but are not assigned to oncoming providers followingthe handoff update)43; transfer of inappropriate tasks and/orfailure to complete activities before the end of a shift85,86; andhaving legible, accurate, and up-to-date documentation.

One concern with accountability considerations is whetherchanges to handoff procedures are accompanied by a shift inpower.87 Therefore, we suggest avoiding interventions that areintended to shift power from front-end workers to more distantadministrators and regulators, such as requiring additional dataentry in systems to justify deviations from recommended prac-tices such as appropriate beta-blocker use.37

FRAMING 5: SOCIAL INTERACTION

The fifth framing, social interaction, emphasizes how ahandoff allows a co-construction of essential meanings and co-orientation toward that essential meaning on the basis ofknowledge of the perspective of the participants in theexchange.88 All communicative acts include both a contentdimension and a relational dimension, which reinforces socialrelationships between parties. For example, a departing physi-cian in the ICU could provide a handoff update to an oncom-ing physician that describes the plan for determining betweentwo potential diagnoses for the patient, which includes severalblood draws. The nurse could overhear the interaction andrequest that an arterial line be inserted if multiple laboratoriesrequiring blood draws are planned to make it more comfortablefor the patient. The oncoming physician agrees to this requestby the nurse, also mentioning that this would also provide anopportunity to mentor the medical student in the procedure.The handoff discussion between the physicians then continueswhile the nurse can again overhear the conversation, with theoncoming physician suggesting an alternative third possibilityfor the diagnosis that could be ruled out with a magnetic reso-




nance imaging (MRI) test of the lung. The nurse then rejoinsthe discussion and comments that the patient’s lung sounds thismorning during assessment might indicate some problems forwhich an MRI might provide insight.

Interdisciplinary handoffs that include both physician andnursing personnel would be an intervention with this framing.A literature search on interdisciplinary handoffs uncovered noarticles measuring the impact of implementing these kinds ofhandoffs.

Therefore, the intervention example we provide is notspecifically for handoffs, but rather for rounds, where attemptswere made to explicitly support interdisciplinary team commu-nications and to take the different perspectives into account (insome cases including the patient and caregivers).89

Measures associated with this frame are quality of interpro-fessional communication,90 respectful body language,91 the will-ingness to wait to begin the handoff until both participants areready to engage fully in the interaction (for example, the anes-thesiologist waits until the patient’s airway is unobstructedbefore providing the handoff to the nurse in the postanesthesiacare unit), and whether new insights emerge from the handoffcommunication, such as new potential diagnoses or treatmentplans to pursue.

FRAMING 6: DISTRIBUTED COGNITION

The sixth framing, distributed cognition, addresses how atransfer to a new care provider affects a network of specializedpractitioners performing dedicated roles who may or may notbe transitioning at the same time. Updates need to be providedto the network about when the transfer has occurred andwhether the contact information and/or strategy needs to bechanged, along with the individual as well as parallel updates tothe oncoming provider about what oncoming providers havereplaced departing providers.

Interventions with this framing include making accurate, up-to-date contact information easily accessible92; using white-boards (or other “shared repository artifacts”) to aidcoordination between caregivers93; and clarifying who has pri-mary responsibility and authority for a patient, such as by hand-ing over pagers or cell phones for use in a particular role (chargenurse, nurse, on-call physician, code team) at the conclusion ofa handoff update to signal that the responsibility has transferred.

Measures for this frame include time to contact an intendedprovider, number of “re-routes” to other personnel beforeobtaining access to a specialist, the accuracy of information ona shared repository as compared with other informationsources, and technical errors due to teamwork interactions.62

FRAMING 7: CULTURAL NORMS

The final framing, cultural norms, relates to how group val-ues (instantiated as social norms for acceptable behavior) in anorganization or suborganization are negotiated and maintainedover time. Socialization about “how things are done here” dur-ing on-the-job orientation can include practices for conductinghandoffs, and thus training interventions of cultural changeinterventions can spread beyond the directly involved individ-uals. Alternatively, the “hidden curriculum”94 can underminepolicies, procedures, and training if what is documented andtaught differs from what is learned during an apprenticeshipperiod in the local work environment.

Interventions with this framing include providing sufficientorganizational support to conduct handoffs, such as by reduc-ing multiple concurrent tasks,95 by allocating a dedicated timefor handoffs, having adequate staffing and/or supplementalstaffing during the transfer such as with short-term coverage bynurse managers, overlapping shifts and avoiding short shiftswith minimal staffing, having personnel personally hand off tothe oncoming person, and handing off on-call responsibilitieslike responding to codes before the transfer to reduce the like-lihood of having to cut off the update before it is completed.

Another category of intervention is educational initiativeson how to conduct handoffs.96,97 Unfortunately, there is littleguidance on how to measure the impact of educational initia-tives in general. One measurement approach following a tradi-tional one-hour class setting to teach effective communicationskills during verbal sign-out handoffs to medical students andresidents was to do a pre-post within-subject comparison ofperceived comfort with providing sign-out.4 Simulation-basedcourses that teach recommended handoff practices can beassessed as to whether the practices, or “acceptable” modifica-tions, are adopted in the actual work setting using direct obser-vation. Guided-reflection techniques, where audio or videoclips are used to stimulate group discussion, can be judged as towhat previously unrecognized issues and associated suggestionsfor handoff process improvement are identified and/or lead toprocess improvements.98,99 Similar measurement strategies canbe used for regular administrative reviews of transfer process-es,100 the appreciative inquiry technique where the staff wereasked to discuss and build on their most effective handoff expe-riences,101 and new definitions for professionalism that incor-porate the quality of transfer processes.102

DiscussionIn the last decade, numerous quality improvement projects onpatient handoffs have been conducted, and clearly there




remains much interest in improving on the baseline for patienthandoff processes across the industry. Nevertheless, our reviewand classification of the handoffs literature do not enable us tomake recommendations for the use of any particular standard-ized, reliable measurement tool.

We believe that the wide diversity of handoff quality mea -sures suggests a lack of consensus about the primary purpose ofhandoffs and how best to intervene to improve handoffprocesses. We also believe that all of the conceptual framingsthat we have identified are potentially useful, compatible fram-ings and that there might be others that we have missed. Aswork moves forward, clarifying and elaborating alternative con-ceptual framings for the purposes of handoffs should enable usto make progress on developing useful measures to objectivelyassess handoff quality.

Nevertheless, we and others103 caution against narrowlydefining a handoff based solely on a single conceptual frame,the dominant information processing metaphor for communi-cating information accurately in a noisy channel. The domi-nance of the information processing measures in this reviewand the many calls to standardize the information content ofhandoffs suggest to us that there is a fundamental miscalibra-tion about how much content standardization and contentordering is desirable or even possible. For us, a patient’s courseof illness as compared to stereotypical patterns; detection oferroneous assumptions and plans; the environmental context,including work load and the number and character of “hang-ing” tasks following the transfer; the social context and back-ground of the handoff participants; team interactions; andcultural norms about acceptable behavior would inform thedefinition of what is “insightful” to include in a particularpatient handoff interaction.

References1. The Joint Commission: The 2010 Accreditation Manual for Hospitals: TheOfficial Handbook. Oakbrook Terrace, IL: Joint Commission Resources, 2009. 2. Accreditation Council for Graduate Medical Education (ACGME):Resident Services Menu. http://www.acgme.org/acWebsite/navPages/

nav_residents.asp (last accessed Dec. 8, 2009).3. Vidyarthi A.R., et al.: Managing discontinuity in academic medical centers:Strategies for a safe and effective resident sign-out. J Hosp Med 1:257–266,Mar. 2006.4. Horwitz L.I., Moin T., Green M.L.: Implementation and evaluation of anasynchronous physician sign-out for emergency department admissions. J GenIntern Med 22:1470–1474, Oct. 2007. Epub Aug. 3, 2007.5. Philibert I., Leach D.C.: Re-framing continuity of care for this century.Qual Saf Health Care 14:394–396, Dec. 2005.6. Institute of Medicine: Resident Duty Hours: Enhancing Sleep, Supervision,and Safety. 2008. http://books.nap.edu/openbook.php?record_id=12508 (lastaccessed Dec. 9, 2009).7. World Health Organization: Communication during patient hand-overs.Patient Safety Solutions 1(3), May 2007. http://www.who.int/patientsafety/solutions/patientsafety/PS-Solution3.pdf (last accessed Dec. 9,2009).8. Arora V., Johnson J.: A model for building a standardized hand-off proto-col. Jt Comm J Qual Patient Saf 32:646–655, Nov. 2006.9. Petersen L.A., et al.: Does housestaff discontinuity of care increase the riskfor preventable adverse events? Ann Intern Med 121:866–872, Dec. 1, 1994.10. Singh H., et al.: Medical errors involving trainees: A study of closed mal-practice claims from 5 insurers. Arch Intern Med 167:2030–2046, Oct. 22,2007.11. Borowitz S.M., et al.: Adequacy of information transferred at residentsign-out (inhospital handover of care): A prospective survey. Qual Saf HealthCare 17:6–11, Feb. 2008.12. Christian C.K., et al.: A prospective study of patient safety in the operat-ing room. Surgery 139:159–173, Feb. 2006.13. Kitch B.T., et al.: Handoffs causing patient harm: A survey of medical andsurgical house staff. Jt Comm J Qual Patient Saf 34:563–570, Oct. 2008.14. Ye K., et al.: Handover in the emergency department: Deficiencies andadverse effects. Emerg Med Australas 19(5):433–441, 2007.15. Simpson K.R.: Handling handoffs safely. MCN Am J Matern Child Nurs30(2):152, Mar.–Apr. 2005.16. Evans S.M., et al.: Assessing clinical handover between paramedics and thetrauma team. Injury, in press. Epub Sep. 6, 2009.17. Sexton A., et al: Nursing handovers: Do we really need them? J NursManag 12:37–42, Jan. 2004.18. Risser D.T., et al.: The potential for improved teamwork to reduce med-ical errors in the emergency department. The MedTeams ResearchConsortium. Ann Emerg Med 34:370–372, Sep. 1999.19. Lawrence R.H., et al.: Conceptualizing handover strategies at change ofshift in the emergency department: A grounded theory study. BMC HealthServ Res 8:256, Dec. 16, 2008.20. Behara R., et al.: A conceptual framework for studying the safety of tran-sitions in emergency care. In Henriksen K. (ed.): Advances in Patient Safety,vol. 2. Agency for Healthcare Research and Quality, 2005. pp. 309–321.http://www.ahrq.gov/downloads/pub/advances/vol2/behara.pdf (last accessedDec. 9, 2009). 21. Kowalsky J., et al.: Understanding sign outs: Conversation analysis revealsICU handoff content and form. Crit Care Med 32(12):A29, 2004.22. Solet D.J., et al.: Physician-to-physician communication: Methods, prac-tice and misgivings with patient handoffs. J Gen Intern Med 19(suppl. 1):108,2004.23. Beach C., Croskerry P., Shapiro M.: Profiles in patient safety: Emergencycare transitions. Acad Emerg Med 10:364–367, Apr. 2003.24. Bomba D.T., Prakash R.: A description of handover processes in anAustralian public hospital. Aust Health Rev 29:68–79, Feb. 2005.25. Nemeth C.P., et al.: Between shifts: Healthcare communication in thePICU. In Nemeth C.P. (ed.): Improving Healthcare Team Communication.

J

Emily S. Patterson, Ph.D., is Assistant Professor, Health

Information Management and Systems Division, Ohio State

University Medical Center, School of Allied Medical Professions,

Columbus, Ohio; and a member of The Joint Commission Journalon Quality and Patient Safety’s Editorial Advisory Board. Robert L.

Wears, M.D., M.S., is Professor, Department of Emergency

Medicine, Division of Emergency Medicine Research, College of

Medicine, University of Florida, Jacksonville. Please address corre-

spondence to Emily S. Patterson, [email protected].




Hampshire, UK: Ashgate Publishing, 2008, pp. 135–158.26. Cleland J.A., et al.: There is a chain of Chinese whispers: Empirical datasupport the call to formally teach handover to prequalification doctors. QualSaf Health Care 18:267–271, Aug. 2009.27. Al-Benna S., Al-Ajam Y., Alzoubaid, D.: Burns surgery handover study:Trainees’ assessment of current practice in the British Isles. Burns 35:509–512,Jun. 2009.28. Kennedy R., et al.: Northern Ireland General Surgery Handover Study:Surgical trainees’ assessment of current practice. Surgeon 7:10–13, Feb. 2009.29. Jeffcott S.A., et al.: Improving measurement in clinical handover. Qual SafHealth Care 18:272–277, Aug. 2009.30. Philibert I.: Selected Articles on the Patient Hand-Off. 2007.http://www.acgme.org/acwebsite/dutyhours/dh_annotatedbibliography_patienthandoff_1207.pdf (last accessed Dec. 9, 2009).31. Cohen M.D., Hilligoss P.B.: Handoffs in hospitals: A review of the litera-ture on information exchange while transferring patient responsibility or con-trol (draft) Jan. 28, 2009. http://deepblue.lib.umich.edu/handle/2027.42/61522 (last accessed Dec. 9, 2009).32. Riesenberg L.A., Leitzsch J., Little B.: Systematic review of handoffmnemonics literature. Am J Med Qual 24:196–204, May–Jun. 2009. EpubMar. 5, 2009. 33. Djuricich A.M., Logio L.: Handoffs bibliography for effective patienthandoffs: Improving resident changeover. Res APDIM Spring Meeting, Apr.2007. http://www.aiamc.org/public/AIAMC_Hand-offs_Resources.pdf (lastaccessed Dec. 9, 2009).34. Lardner R.: Effective Shift Handover: A Literature Review. OffshoreTechnology Report–OTO 96 003. The Keil Centre, 1996.http://www.hse.gov.uk/research/otopdf/1996/oto96003.pdf (last accessedDec. 9, 2009) 35. Australian Commission on Safety and Quality in Healthcare (ACSQH):Clinical handover and patient safety: Literature review report, Apr. 2007.http://www.safetyandquality.gov.au/ (last accessed Sep. 5, 2007; update, Apr.2008, available at http://www.health.gov.au/internet/safety/publishing.nsf/content/E0B59E130FA90A50CA2573AF007BC3C8/$File/CHoverLitReview.pdf). 36. Kerr M.P.: A qualitative study of shift handover practice and function froma sociotechnical perspective. J Adv Nurs 37:125–134, Jan. 2002.37. Smith A.F., et al.: Interprofessional handover and patient safety in anaes-thesia: Observational study of handovers in the recovery room. Br J Anaesth101:332–337, Sep. 2008. Epub Jun. 13, 2008.38. Philpin S.: ‘Handing over’: Transmission of information between nurses inan intensive therapy unit. Nurs Crit Care 11:86–93, Mar.–Apr. 2006.39. Cheung D.S., et al., American College of Emergency Physicians Sectionof Quality Improvement and Patient Safety: Improving handoffs in the emer-gency department. Ann Emerg Med, in press. Epub Oct. 1, 2009. 40. Patterson E.S., Wears R.L.: Measurement approaches for transitions ofauthority and responsibility during handoffs. In Patterson E.S., Miller J.(eds.): Macrocognition Metrics and Scenarios: Design and Evaluation for Real-World Teams. Farnham, UK: Ashgate Publishing, forthcoming.41. The Joint Commission: Meeting the Joint Commission 2008 NationalPatient Safety Goals. Oakbrook Terrace, IL: Joint Commission Resources,2007. 42. Wallum R.: Using care plans to replace the handover. Nurs Stand 9:24–26,May 1995. 43. Van Eaton E.G., et al.: A randomized, controlled trial evaluating theimpact of a computerized rounding and sign-out system on continuity of careand resident work hours. J Am Coll Surg 200:538–545, Apr. 2005.44. Flanagan M.E., et al.: Evaluation of a physician informatics tool toimprove patient handoffs. J Am Med Inform Assoc 16:509–514, Jul.–Aug.2009. Epub Apr. 23, 2009.

45. Frank G., Lawless S.T., Steinberg T.H.: Improving physician communica-tion through an automated, integrated sign-out system. J Healthc Inf Manag19:68–74, Fall 2005.46. Gurses A.P., Xiao Y.: A systematic review of the literature on multidiscipli-nary rounds to design information technology. J Am Med Inform Assoc13:267–276, May–Jun. 2006.47. Arora V.M., et al.: Hospitalist handoffs: A systematic review and task forcerecommendations. J Hosp Med 4:433–440, Sep. 2009.48. Solet D.J., et al.: Lost in translation: Challenges and opportunities inphysician-to-physician communication during patient handoffs. Acad Med80:1094–1099, Dec 2005.49. McCann L., McHardy K., Child S.: Passing the buck: Clinical handoversat a tertiary hospital. N Z Med J 120:U2778, Oct. 26, 2007.50. Coiera E.W., et al.: Communication loads on clinical staff in the emer-gency department. Med J Aust 176:415–418, May 6, 2002.51. Dahl Y.: ‘You have a message here’: Enhancing interpersonal communica-tion in a hospital ward with location-based virtual notes. Methods Inf Med45:602–609, 2006.52. Raptis D.A., et al.: Electronic software significantly improves quality ofhandover in a London teaching hospital. Health Informatics J 15:191–198,Sep. 2009.53. Berkenstadt H., et al.: Improving handoff communications in critical care:Utilizing simulation-based training toward process improvement in managingpatient risk. Chest 134:158–162, Jul. 2008.54. Talbot R., Bleetman A.: Retention of information by emergency depart-ment staff at ambulance handover: Do standardized approaches work? EmergMed J 24:539–542, Aug. 2007.55. Bhabra G., et al.: An experimental comparison of handover methods. AnnR Coll Surg Engl 89:298–300, Apr. 2007.56. Jenkin A., Abelson-Mitchell N., Cooper S.: Patient handover: Time for achange? Accid Emerg Nurs 15:141–147, Jul. 2007.57. Haig K.M., Sutton S., Whittington J.: SBAR: A shared mental model forimproving communication between clinicians. Jt Comm J Qual Patient Saf32:167–175, Mar. 2006.58. Porteous J.M., et al.: iSoBAR—A concept and handover checklist: TheNational Clinical Handover Initiative. Med J Aust 190(11 suppl.):S152–S156,Jun. 2009.59. Patterson E.S.: Editorial: Communication strategies from high-reliabilityorganizations: Translation is hard work. Ann Surg 245:170–172, Feb. 2007.60. McLaughlin E., Antonio L., Bryant A.: Hospital nursing: Get an A+ onend-of-shift report. Nursing 34(6):32hn8–32hn10, 2004.61. Slagle J.M., et al.: Simulation training for rapid assessment and improvedteamwork: Lessons learned from a project evaluating clinical handoffs.Proceedings of the Human Factors and Ergonomics Society 51st Annual Meeting—2007. http://www.mc.vanderbilt.edu/cprq/publications/Slagle%20STRAIT.pdf (last accessed Dec. 9, 2009). 62. Catchpole K.R., et al.: Patient handover from surgery to intensive care:Using Formula 1 pit-stop and aviation models to improve safety and quality.Paediatr Anaesth 17:470–478, May 2007.63. Pickering B.W., Hurley K., Marsh B.: Identification of patient informa-tion corruption in the intensive care unit: Using a scoring tool to direct qual-ity improvements in handover. Crit Care Med 37:2905–2912, Nov. 2009.64. Fenton W.: Developing a guide to improve the quality of nurses’ handover.Nurs Older People 18:32–37, Dec. 2006.65. Meissner A., et al.: Nurses’ perception of shift handovers in Europe: resultsfrom the European Nurses’ Early Exit Study. J Adv Nurs 57:535–542, Mar.2007.66. Arora V., et al.: Communication failures in patient sign-out and sugges-tions for improvement: A critical incident analysis. Qual Saf Health Care14:401–417, Dec. 2005.




67. Lee L.H., Levine J.A., Schultz H.J.: Utility of a standardized sign-out cardfor new medical interns. J Gen Intern Med 11:753–755, Dec. 1996.68. Sabir N., Yentis S.M., Holdcroft A.: A national survey of obstetric anaes-thetic handovers. Anaesthesia 61:376–380, Apr. 2006.69. Lofgren R.P., et al.: Post-call transfer of resident responsibility: Its effect onpatient care. J Gen Intern Med 5:501–505, Nov.–Dec. 1990.70. Petersen L.A., et al.: Using a computerized sign-out program to improvecontinuity of inpatient care and prevent adverse events. Jt Comm J QualImprov 24:77–87, Feb. 1998. 71. Young G.J., et al.: Patterns of coordination and clinical outcomes: A studyof surgical services. Health Serv Res 33(5 pt. 1):1211–1236, Dec. 1998.72. Footitt B.: Ready for report. Nurs Stand 11:22 –23, Mar. 12, 1997.73. Henriksen K., Kaplan H.: Hindsight bias, outcome knowledge and adap-tive learning. Qual Saf Health Care 12(suppl. 2):ii46–ii50, Dec. 2003.74. Schank R.C., Abelson R.P.: Scripts, Plans, Goals, and Understanding.Hillsdale, NJ: Lawrence Erlbaum Associates, 1977.75. Srull T.K.: Person memory: Some tests of associative storage and retrievalmodels. Journal of Experimental Psychology: Human Learning and Memory7:440–463, Nov. 1981. 76. Pronovost P., et al.: Improving communication in the ICU using dailygoals. J Crit Care 18:71–75, Jun. 2003.77. Miller A., et al.: Uni- and interdisciplinary effects on round and handovercontent in intensive care units. Hum Factors 51:339–353, Jun. 2009.78. Patterson E.S., et al.: Collaborative cross-checking to enhance resilience.Cognition, Technology and Work 9(2):155–162, 2007.79. Wears R.L., et al.: Shift changes among emergency physicians: Best oftimes, worst of times. Proceedings of the Human Factors and Ergonomics Society47th Annual Meeting, pp. 1420–1423, 2003.80. Pian-Smith M.C.M., et al.: Teaching residents the two-challenge rule: Asimulation-based approach to improve education and patient safety. SimulHealthc 4:84–91, Summer 2009.81. U.S. Army: Aircrew Training Program Commander’s Guide to Individualand Crew Standardization. Army Aviation Technical Manual TC1–210. Jan.1996. http://www.cavalrypilot.com/tc1-210/chg1toc.html (last accessed Dec.9, 2009). 82. Patterson E.S., Woods D.D.: Shift changes, updates, and the on-callarchitecure in space shuttle mission control. Comput Support Coop Work10(3–4), 317–346, 2001.83. Cooper J.B.: Do short breaks increase or decrease anesthetic risk? J ClinAnesth 1(3):228–231, 1989. 84. Williams R.G., et al.: Surgeon information transfer and communication:Factors affecting quality and efficiency of inpatient care. Ann Surg245:159–169, Feb. 2007.85. Wayne J.D., et al.: Simple standardized patient handoff system thatincreases accuracy and completeness. J Surg Educ 65:476–485, Nov.–Dec.2008.86. Ferran N.A., Metcalfe A.J., O’Doherty D.: Standardised proformas

improve patient handover: Audit of trauma handover practice. Patient Saf Surg25:24, Sep. 2008.87. Murphy A.G., Wear, R.L.: The medium is the message: Communicationand power in sign-outs. Ann Emerg Med 54:379–380, Sep. 2009. Epub Apr.11, 2009.88. Eisenberg E., et al.: Communication in emergency medicine: Implicationsfor patient safety. Communication Monograph 72(4):390–413, 2005.89. Uhlig P.N., et al.: John M. Eisenberg Patient Safety Awards: System inno-vation: Concord Hospital. Jt Comm J Qual Improv 28:666–672, Dec. 2002.90. Botti M., et al.: Examining communication and team performance duringclinical handover in a complex environment: The private sector post-anaes-thetic care unit. Med J Aust 190(11 suppl.):S157–S160, Jun. 2009.91. Leonard M., Graham S., Bonacum D.: The human factor: The criticalimportance of effective teamwork and communication in providing safe care.Qual Saf Health Care 13(suppl. 1):i85–i90, Oct. 2004.92. Hiltz F.L., Teich J.M.: Coverage list: A provider-patient database support-ing advanced hospital information services. Proc Annu Symp Comput Appl MedCare, 809–813, 1994.93. Wilson S., Galliers J., Fone J.: Cognitive artifacts in support of medicalshift handover: An in use, in situ evaluation. International Journal of Human-Computer Interaction 22(1 & 2):59–80, 2007.94. Hafferty F.W., Franks R.: The hidden curriculum, ethics teaching, and thestructure of medical education. Acad Med 69:861–871, Nov. 1994.95. Strople B., Ottani P.: Can technology improve intershift report? What theresearch reveals. J Prof Nurs 22:197–204, May–Jun. 2006.96. Arora V.M., et al.: A theoretical framework and competency-basedapproach to improving handoffs. Qual Saf Health Care 17:11–14, Feb. 2008.97. Theorem S., Morrison W.: A survey of the perceived quality of patienthandover by ambulance staff in the resuscitation room. Emerg Med J18:293–296, Jul. 2001.98. Broekhuis M., Veldkamp C.: The usefulness and feasibility of a reflexivitymethod to improve clinical handover. J Eval Clin Pract 13:109–115, Feb.2007.99. Iedema R., et al.: Handover—Enabling Learning in Communication forSafety (HELiCS): A report on achievements at two hospital sites. Med J Aust190(11 suppl.):S133–S136, Jun. 2009.100. Miller C.: Ensuring continuing care: Styles and efficiency of the handoverprocess. Aust J Adv Nurs 16:23–27, Sep.–Nov. 1998.101. Shendell-Falik N., Feinson M., Mohr B.J.: Enhancing patient safety:Improving the patient handoff process through appreciative inquiry. J NursAdm 37:95–104, Feb. 2007.102. Kellogg K.C., et al.: Resistance to change in surgical residency: An ethno-graphic study of work hours reform. J Am Coll Surg 202:630–636, Apr. 2006.103. Turner P., Wong M.C., Yee K.C.: A standard operating protocol (SOP)and minimum data set (MDS) for nursing and medical handover:Considerations for flexible standardization in developing electronic tools. StudHealth Technol Inform 143:501–506, 2009.


February 2010 Volume 36 Number 2


62

The Veterans Affairs Shift Change Physician-to-PhysicianHandoff Project

Continuity of Care

Communication breakdowns are cited as the root cause forthe majority of sentinel events in hospitalized patients,

and the lack of detailed knowledge about patients and poorcommunication of plans to cross-covering physicians is pre-dominant among causes of decreased quality of care.1–3

Communication errors commonly occur during “handoffs,”defined by The Joint Commission as the “real-time process ofpassing patient-specific information from one caregiver toanother or from one team of caregivers to another for the pur-pose of ensuring the continuity and safety of the patient’scare.”4(p. 65) For all accreditation programs, National PatientSafety Goal 2E, in effect through 2009, stated that hospitalsshould “implement a standardized approach to ‘handoff ’ com-munications, including an opportunity to ask and respond toquestions” with “opportunity for discussion between the giverand receiver of information.”4 The Joint Commission has alsorecommended that the receiver of information have opportuni-ty to review relevant patient data, that information be up todate with a process to verify information, and that interrup-tions be limited.4 The Joint Commission’s requirement for astandardized approach to handoff communications is nowElement of Performance 2 (“The hospital’s process for hand-offcommunications provides for the opportunity for discussionbetween the giver and receiver of patient information”) forStandard PC.02.02.01 (“The hospital coordinates the patient’scare, treatment, and services based on the patient’s needs”).5

The Department of Veterans Affairs (VA) has made similar rec-ommendations,6,7 yet standardized methods for patient hand-offs remain rare for physicians.8

Many industries, not only health care, have complex, inter-connected, event-driven, time-pressured, and resource-con-strained systems with the potential for serious consequences forsystem failure. Strategies have been designed to address theseissues.9,10 The current literature suggests that successful handoffsin health care should include the following vital information:contact information for the primary team, complete patientidentification data, active problem list, pertinent past medical

Article-at-a-Glance

Background: Few studies on the safety or efficacy of cur-rent patient handoff systems exist, and few standardizedelectronic medical record (EMR)–based handoff tools areavailable. An EMR handoff tool was designed to provide astandardized approach to handoff communications andimprove on previous handoff methods.Methods: In Phase I, existing handoff methods were ana-lyzed through abstraction of printed handoff sheets andquestionnaires of internal medicine residents atDepartment of Veterans Affairs medical centers (VAMCs).In Phase II, the handoff tool was designed, and the softwarewas tested and revised through user feedback and regularconference calls. Phase III involved postimplementationsystematic abstraction of printed handoff sheets and ques-tionnaires of internal medicine residents. Two VAMCs par-ticipated in abstraction of printed handoff sheets, with fourVAMCs responding to the questionnaires. Results: Handoffs were abstracted for 550 patients atbaseline and 413 postimplementation. Improvements werefound in consistency of information transfer for all handoffcontent, including code status, floor location, room num-ber, two types of identifying information, typed format,medication, and allergy lists (p = .01). The 63 and 51 ques-tionnaires completed pre- and postimplementation, respec-tively, showed improvement in perceptions of ease of use,efficiency, and readability (p < .05) and in perceptions ofpatient safety and quality (p < .01) without causing omis-sion (p < .01) or commission of information (p = .02). Discussion: This standardized EMR–based handoff soft-ware improved data accuracy and content consistency, waswell-received by users, and improved perceptions of hand-off-related patient safety, quality, and efficiency. A final ver-sion of the software was incorporated into the nationalEMR software program and made available to all VAMCs.

Jaclyn Anderson, D.O., M.S.; Divya Shroff, M.D.; Ann Curtis, M.D.; Noel Eldridge, M.S.; Katrina Cannon, M.D.,M.S.; Rajil Karnani, M.D., M.M.E.; Thad Abrams, M.D., M.S.; Peter Kaboli, M.D., M.S.




history, current condition, active and updated medication andallergy lists, code status, anticipated changes in the next careinterval with a recommended course of action, and psychoso-cial concerns that may influence therapeutic choices.9,11–14 Inaddition, some have advocated listing the acuity of the patient,cognitive status, cardiopulmonary status, information onvenous access, discussions on level of care and listing long-termplans in case families have questions overnight, pertinent labo-ratory data, pending tests, consults, and procedures.9,12,14

Previously surveyed physicians considered two main cate-gories of communication failures to be most important: “con-tent omissions,” in which critical information was notcommunicated, and “failure-prone communications processes,”including lack of face-to-face communication and illegiblehandwritten notes that often omit information because of anunwillingness to rewrite information daily. In addition, commis-sion of information, with inclusion of irrelevant information,has also been cited as a cause of poor handoffs.13–15

To help reduce these communication process errors, stan-dardized electronic medical record (EMR)–linked handoff toolsmay be a powerful asset. Yet few EMR–linked handoff tools arecurrently available. Computerized handoff systems have beendemonstrated to reduce poor handoff through improving infor-mation completeness, legibility, and accuracy, and are preferredby residents to handwritten handoffs.15–19 Computerized hand-offs have also been reported to improve patient safety in regardto the frequency of adverse events between cross-coveredpatients and primary team–covered patients, by decreasingpatients missed on rounds and improving perceptions of quali-ty of sign-out and continuity of care.2,20 Moreover, many hospi-tals in the United States use handoff systems that are not secure(for example, e-mail, unencrypted word-processing documents)to save handoff lists from day to day. Furthermore, there aremany barriers to effective handoffs, including the physical set-ting (for example, distractions, background noise), the socialsetting (senior resident communicating with junior resident),language and communication barriers, and constraints ontime,9 with potential loss of vital information occurring witheach handoff.

Although software based on housestaff needs in a solitaryhospital17 and a large hospital system has been shown toimprove patient handoff,20 a single software package has notbeen previously developed and tested in diverse geographicalregions. Therefore, we conducted an evaluation to develop andtest in diverse geographical regions within the VA health caresystem a secure EMR–linked handoff tool to improve the qual-ity of the physician-to-physician handoff in terms of complete-

ness, content quality in terms of consistency of informationtransfer, legibility, data security, between-physician communi-cation, and physician acceptance with and efficiency of thehandoff process.

Methods PHASE I. ANALYSIS OF EXISTING HANDOFF TOOLS

To determine the actual information conveyed in writing byphysicians and ensure that content deemed important to physi-cians was included, we performed data collection and analysis,as now described.

Abstraction of Handoff Sheets. Before implementing thehandoff software, handoff sheets used by internal medicinephysicians rotating at participating VAMCs were collected atthe end of shifts on a voluntary basis as part of a qualityimprovement initiative. Participation in this process wasoptional, with three of the initial seven sites invited in February2006 choosing to participate. Because of the voluntary natureof this endeavor, an attempt was not made to collect all possi-ble handoff sheets during this time period.

Handoff sheets were abstracted to identify the followingvital information: code status, medication lists, allergy lists,complete patient identification and location data (floor, roomnumber), and format (typed or handwritten.) Items could beeither present or absent for each individual patient. Medicationlists were coded as present if any portion of the list was present.Format was used as a surrogate for legibility. Tasks handed offto cross-cover physicians to do during their shift and problemsencountered by the covering physician were also abstracted astext. Anticipated changes in the next care interval with a recom-mended course of action and psychosocial concerns—informa-tion that is frequently conveyed orally without a writtenform—were not abstracted. Although considered importantaspects of patient handoffs, abstraction was not performedregarding primary team contact information, pertinent medicalhistory, current condition, or problem lists, as confirmation ofcompleteness and accuracy of these items was not possible.

Survey of Residents’ Perceptions of Current HandoffMethods. To inform the design of the handoff software, thesoftware development team (including J.A., P.K.) surveyedinternal medicine residents rotating at participating VAMCsregarding their perceptions of current handoff methods beforethe handoff software was introduced. The questionnaire wasdesigned to test physician perceptions of handoff quality andefficiency by soliciting themes present in the literature andthrough modification of previously used surveys. The question-naire included six domains: (1) quality and safety, (2) complete-




ness and legibility, (3) user acceptance and usability, (4) effi-ciency, (5) physician communication, and (6) communicationwith nursing.

The questionnaire responses used 5-point Likert scales andfree-text comments. For purposes of data presentation, “strong-ly agree” and “agree” were combined, as were “disagree” and“strongly disagree,” with “neutral” as a separate category.Anonymous questionnaires (as both paper-based and Web-based instruments) were distributed pre- and postimplementa-tion to a convenience sample of resident physicians who hadrecently been assigned to the medicine wards. Because someresidency program directors were unwilling to share the namesof their residents, we were not able to track the responses or thetotal numbers of potentially eligible physicians to determineaccurate response rates.

PHASE II. HANDOFF TOOL DESIGN AND

DEVELOPMENT

Design. Informed by the analysis of existing handoff methodsand questionnaire responses, current literature, and opinionsrendered by an expert consensus panel of clinicians and infor-mation technology (IT) personnel, the VA Office of PatientCare Services in January 2006 charged a consensus panel—composed of the software design team, programmers, and ITpersonnel—with the design of an EMR–linked handoff soft-ware program to address Joint Commission National PatientSafety Goal 2E. The handoff software was based on a programoriginally developed at the Roudebush VAMC (Indianapolis)in 1999, with continual improvements as indicated by feedbackfrom end users.

Development and Testing. In February 2006, seven geo-graphically diverse VAMCs initially participated in softwaredevelopment and testing. At each site, the medical center direc-tor was contacted, and physician and IT champions werenamed. In April 2006, the Indianapolis VAMC distributed thesoftware to four of the initial seven facilities. (Not all seven siteswere able to implement the software because of restrictions onthe availability of IT personnel and the use of unproven com-puter software applications). Biweekly teleconferences wereheld with all parties to discuss the software, with identificationof functionality issues, troubleshooting of software capabilities,and sharing of solutions. Continuous monitoring at the testsites ensured the tool contained the content desired. At theindividual hospital level, the handoff software was managed bythe IT personnel managing the EMR system.

By August 2006, six of the seven facilities were testing thesoftware. However, because of significant problems with func-

tionality (for example, difficulties in forming the handoff list,propagating the DNR [do-not-resuscitate] orders, printing) allfacilities ceased using this version of the handoff software by theend of the month. In September 2006, a face-to-face meetingwas held with staff from participating VAMCs to discuss howto make the handoff tool better (including making the softwaremore user-friendly and printing more patients per page) andhow to interface better with orders for the VA’s existing EMR,the Computerized Patient Record System (CPRS), while main-taining patient safety. At this meeting, the group agreed onfunctionality features for the revised handoff tool, with theupdated version installed initially at one site in February 2007.After troubleshooting issues at the initial test site, the softwarewas released to the remaining sites at the end of that month. InMarch 2007, after the initial sites were successfully using thehandoff software, word of mouth created interest in the pro-gram, and several additional sites joined testing. In total, 12VAMCs participated in testing prior to approval for nationalrelease in October 2007. Before national distribution, enablingof the software application for use by handicapped persons wastested in April 2008 by two of the initially testing sites and twosites new to the software to ensure functionality in diverse set-tings. During a 26-month period (February 2006 to May2008), the handoff tool was tested at a total of 14 sites to assessusability, ease of installation, or unanticipated problems.Software distribution throughout the United States occurred inJune 2008. All VAMCs were required by the VA central officeto install the software with optional implementation; forcedchange was not thought to increase user buy-in of this softwarepackage, and some facilities already had standardized methodsof handoff in place. In an effort to increase voluntary participa-tion, submission of postimplementation data was not requiredbecause of the burden it would create on the individual sites notinvolved with the initial project.

The majority of VAMCs (the exact numbers are unknown)have adopted the software, with use in internal medicine, sur-gery, and psychiatry. Security of the electronic handoff softwareis password protected using the CPRS system, with routinesoftware management performed by the local IT departmentsmanaging the EMR. Consistent with the current standard,printed handoff reports are to be disposed of in secure fashionafter use.

Tool Features. The handoff software for the shift handofftool, with the on-screen display (Figure 1, page 65) and theprinting display (Appendix 1, available in online article), auto-matically draws information from the CPRS, with mandatoryfields as displayed in Table 1 (page 66) and multiple optional




free-text fields. The system was designed toprevent omission of vital information andimprove patient safety while at the sametime create a product that facilitates theprocess of patient handoff between healthcare providers. The software automaticallyimports demographic data (for example,patient name, social security number[SSN], age, floor location), code status,and medication and allergy lists, and theseitems are automatically updated from theCPRS to reflect current orders. Contactinformation for the primary team (attend-ing, intern, resident, and student, withpager and phone numbers) is updated byend users to reflect continuously rotatingservice schedules and is stored until edited.

The design allows up to four blankfields where pertinent past medical history,current clinical course, “to-do” lists, andother items (for example, specialist con-sults, discharge planning, imaging and lab-oratory data) can be included. The blankfields are free-text boxes, which are site-and specialty-configurable, with a forcing function requiringthe information to be updated within a set period, after whichinformation not updated will be assumed obsolete and auto-matically deleted. Although it is possible to automaticallyimport the assessment and plan from the daily progress noteinto the handoff tool, free-text boxes were used to convey thesedata in an effort to limit printed handoff length. Data automat-ically imported from the EMR is not modifiable by users, andthus error cannot be introduced into the handoff unless ordersare erroneously entered into the CPRS. Medication lists werestandardized to name of drug only without dosage, route, orother information to save space on the printed form and are notmodifiable by users to prevent incomplete information fromappearing complete.

PHASE III. EVALUATION

Following the same process as described above for Phase I,the handoff sheet collection and content analysis was repeatedin March 2007 after software implementation. Three siteselected to participate. The evaluation phase also involvedrepeating the physician questionnaire.

Code Status. Because it is uncommon to order “full code”status, the accepted assumption was to default care to full code

if no code status information was provided. The software devel-opment design team chose to propagate the field with “CodeStatus Not Found” if not otherwise specified. This statementwas intended as a safety reminder to physicians to confirm thepatient’s code status. If there was any mention of code status(for example, full, DNR, “Code Status Not Found”), it wascoded as present.

Participants. Although 12 geographically diverse sites acrossthe United States initially tested the functionality of the hand-off software, only 4 sites (large East coast metropolitan teach-ing hospital, 2 medium-sized Midwestern teaching hospitals, alarge Southern teaching hospital) fully participated in theoptional pre- and postimplementation questionnaires, with 3additional sites participating pre-implementation. Three sitesparticipated in the handoff sheet collection both pre- andpostimplementation; however, only 2 sites (large East coastmetropolitan teaching hospital, medium-sized Midwesternteaching hospital) collected sheets at both pre- and postimple-mentation. To reduce uncontrollable bias, only the sites thatparticipated both pre- and postimplementation in the handoffsheet collection and/or questionnaire were included in theanalysis. However, pooled results from the additional sites didnot significantly differ from the reported results. For missing

Shift Handoff Tool: On-Screen Display for Data Entry

Figure 1. When a provider is working within the Department of Veterans Affairs (VA) electronic med-ical record (the computerized patient record system [CPRS]), a screen is autopopulated by informationpulled from the CPRS, with spaces (shaded boxes) available for optional free-text entry. HPI, history ofpresent illness.




questionnaire responses, the denominators were based on thenumber of nonmissing answers.

Data Abstraction. Data abstraction was performed by twostaff physicians [J.A., A.C.]. Interrater reliability for the first 69handoff tool abstractions showed a high level of agreement � = 0.92), so that abstraction of the remaining 894 patienthandoffs was performed by a single clinician [J.A.] and notrepeated.

Data Analysis. Pearson’s chi-square test or Fisher’s exact testwas performed to evaluate change in pre- and postimplementa-tion responses for categorical variables. The Student’s t-test wasperformed to analyze continuous variables. All analyses wereperformed using SAS statistical software, version 9.1 (SASInstitute; Cary, NC). The Institutional Review Board at theUniversity of Iowa and the Iowa City VA Medical CenterResearch and Development Committee approved this study.

Results PRE-IMPLEMENTATION HANDOFF TOOL CONTENT

ANALYSIS

Before implementation of the handoff software, 550 patienthandoffs were abstracted (Table 2, page 67). When medicationlists were present, they frequently appeared incomplete by visu-al inspection (for example, no mention of intravenous [IV] flu-ids, only a few medications listed on an acutely ill patient). Noattempt was made to quantify the number of missing medica-tions in a partial list; however, partial lists were counted as pres-ent. Nearly all handoffs were typed and printed for portability

with word processing software, spreadsheets, and handwrittendocuments utilized. The most frequent handoff tasks includedfollow-up of test results or consult recommendations and man-agement of ventilation, blood sugar, and blood pressure. Othertasks included checking on specific patients, performing procedures, confirming medication administration/discontinu-ation, ordering diets, checking line placement, and patienttransfers.

Handwritten comments by covering physicians includedchanges in patient condition, loss of IV access, restraint orders,blood transfusions, falls, laboratory abnormalities, urinarycatheters, medication errors, order clarifications, refusal of ser -vices, poorly communicated plan for patient discharge, partici-pating in family discussions, and failure of tests to beperformed despite orders. One death was noted.

POSTIMPLEMENTATION HANDOFF TOOL CONTENT

ANALYSIS

Postimplementation of the handoff software, 413 patienthandoffs, all of which were produced by the EMR–linkedhandoff tool, were abstracted (Table 2). Physician requests foritems to be done by covering physicians were similar to pre-implementation handoff sheets, as were handwritten commentsmade by covering physicians; however, fewer covering physi-cian comments were noted. Overall improvements from pre- topostimplementation in handoff content regarding consistentpresence of information transferred were found for all vitalinformation categories. Some site-specific variation was noted.

Fields Comments

1. Pertinent past medical ■ Free-text entry that is visible in the handoff tool for the entire inpatient stay and is editable daily

history and current ■ Forcing function requires information to be updated manually within a period of time (e.g., 7 days) or it

clinical course will be assumed obsolete and automatically deleted.

■ Time/date of last update in field included

2. Code status ■ Information imported from CPRS orders

■ Reported as “Code Status Not Found” if no orders are found regarding code status

3. Allergies ■ Information imported from CPRS

■ Reported as “not known” if not entered into CPRS

4. Current medications ■ Imported from CPRS listing active medications alphabetically by name without dosage, route, or other

information

5. To-do list ■ Free text entered daily as needed with same forcing function as Field 1

6. Demographic data ■ Imported from CPRS: full name, SSN, age, gender, floor, team, room/bed, admission date, days of care

7. Header information ■ Team name with name and contact information for attending physician, interns, residents, and students

* CPRS, Computerized Patient Record System; SSN, Social Security Number.

Table 1. Required Fields for the Physician Shift Handoff Tool*




PRE- AND POSTIMPLEMENTATION QUESTIONNAIRES

The results, reported in Table 3 and Table 4 (page 68), aresummarized by survey domain.

Quality and Safety. The four statements related to handoffquality and safety all improved from pre- to postimplementa-tion (Table 4). The percent of respondents who agreed orstrongly agreed that their handoff system “optimizes patientsafety” increased from 44% to 84% (p < .01); only one respon-dent disagreed with this statement postimplementation.Perceptions that the “handoff quality is excellent” and percep-tions of both omissions and commission of information alsoimproved.

Pre-implementation free-text responses mentioned lack ofminimum standards for handoff information and concern withcommission and omission of information. The majority ofthese responses noted that quality of handoff communicationwas primarily dependent on the physicians involved. Weekendswere frequently mentioned as a time for poor handoffs becauseof care provided by physicians unfamiliar with the patients.One respondent mentioned that postimplementation handoffquality was improved because there were “no errors” with auto-

matic importation of data already present in the patient chart.Completeness and Legibility. Respondents stated that a writ-

ten handoff must be easy to read. Before implementation of thehandoff tool, the majority of abstracted patient handoffs weretyped. Notably, the software testing sites not choosing to par-ticipate in the data collection portion of the project reportedmixed handoff methods that involved both handwritten andtyped patient lists. There was a 24% postimplementationimprovement to 90% agreement (p = .03) with the statementthat the handoff is “easy to read.”

Pre-implementation free-text responses related to physiciancommunication indicated that respondents wanted to knowdemographic information, code status, active issues, currentclinical condition, what to do if expected scenarios occur,abnormal and pending tests, medication and allergy lists, recentvital signs, IV access status, and family contacts. With theexception of vital signs, IV access, and family contact informa-tion, all were added to the handoff tool during Phase II throughautomatic data importation, with free-text boxes for pertinentclinical history and a covering shift “to do” list.

User Acceptance and Usability. Common themes included a

Percentage, Chi-square

Site 1 Site 2 Number P value for total

Pre N = 501 N = 49 N = 550 —

Component Post N = 345 N = 68 N = 413 —

Allergy list

Pre 0% (0) 46.9% (23) 4.2% (23)< .01

Post 100% (345) 100% (68) 100% (413)

Room number

Pre 4.4% (22) 100% (49) 12.9% (71)< .01

Post 100% (345) 100% (68) 100% (413)

Medication list

Pre 6.4% (32) 100% (49) 14.7% (81)< .01

Post 99.1% (342) 95.6% (65) 98.5% (407)

Floor location

Pre 64.5% (323) 100% (49) 67.6% (372)< .01

Post 100% (345) 95.6% (65) 99.3% (410)

Code status

Pre 92.6% (464) 100% (49) 93.3% (513)< .01

Post 100% (345) 100% (68) 100% (413)

Complete patient

identification information

Pre 97.0% (486) 100% (49) 97.3% (535)< .01

Post 100% (345) 100% (68) 100% (413)

Typed

Pre 98.4% (493) 100% (49) 98.5% (542).01

Post 100% (345) 100% (68) 100% (413)

Table 2. Presence of Vital Components on Handoff Sheets, Pre- and Postimplementation of the Standardized Handoff Software




Table 4. Pre- and Postimplementation Responses for the Six Questionnaire Domains

Pre-implementation Postimplementation P Value

Training level Percent (Number) Percent (Number)

1st-year resident 61.9% (39) 49.0% (25) < .01

2nd-year resident 20.6% (13) 35.3% (18)

3rd-year resident 17.5% (11) 15.7% (8)

Average time spent at computer to type handoff; minutes (range) 41.2 (8–82) 27.1 (10–50) .20

Average time spent face to face to hand off patients; minutes (range) 13.4 (2–40) 12.7 (1–30) .19

Average number of patients handed off (range) 7.0 (1–15) 6.9 (4–13) .13

Average number of patients received in handoff when on call (range) 13.8 (2–45) 8.4 (1–45) < .01

*All calculations based on non-missing values.

Table 3. Questionnaire Respondent Data Pre- and Postimplementation of the Handoff Software*




need for the handoff to be concise, straightforward, flexible,quickly accessible with minimal typing, paper-based andportable, with simplicity in use. One concern was that animposed handoff tool may be restrictive and not allow flexibil-ity. Pre-implementation, many physicians remarked that theyhad developed their own documents for patient handoffs,which they updated daily. One physician indicated that therewas “no system [for handoffs], it’s whatever each individualchooses to do.” Pre-implementation, multiple respondentsreported lapses in accuracy of manually entered data items fromthe EMR and deficient updating of data, resulting in obsoleteinformation. In general, postimplementation respondents werepleased by the handoff software and commented that it is easyto print and to hand to others and appreciated that handoffdata were saved and easily retrievable. Postimplementationrespondents frequently thanked the team for creating a stan-dardized method of patient handoff with direct importation ofdata from active orders, improving issues with data accuracy.Perceptions of the need for improvement with the handoff sys-tem improved (from 79% pre-implementation to 56.0%postimplementation [p = .05]), as did the perception that “thehandoff system is easy to use.” Although nonsignificant, therewas a trend for improvement in the space for handwritten com-ments.

Efficiency. Free-text responses pre-implementation indicatedthat physicians were accustomed to the current handoffprocess. However, many physicians requested a standardizedhandoff system that automatically imports and updates infor-mation from the EMR to improve efficiency. There was signif-icant postimplementation improvement in the perception thatthe handoff system is “efficient” (p < .01). As shown in Table 3,from pre-implementation (n = 63) to postimplementation (n =51), respondents reported a nonsignificant mean decrease of14.1 minutes spent at the computer to type their handoffs(41.2 minutes versus 27.1 minutes; p = .20) while caring for thesame mean number of patients (7.0 versus 6.9; p = .13), for anonsignificant decrease in computer time from 3.3 to 2.6 min-utes per patient (p = .16).

Physician Communication. As shown in Table 4, postim-plementation, respondents reported improvements in “feelingprepared to care for patients” received in handoff (p < .03) andin understanding “the plan of the outgoing physician” at timeof handoff after use of the handoff software (p < .02). The state-ments “my plan is understood by the receiving physician” and“feeling comfortable handing off patients to the covering physi-cian” each had a nonsignificant increase in agreement.

There was no change in the reported amount of time in face-

to-face handoff communication (13.4 versus 12.7 minutes; p = .19; Table 3).

Nursing Communication. Responses to questionnaire itemsrelated to nursing communication, including the item“Nursing staff understand the plan for the patient at the timeof physician handoff,” were not affected by the new handoffsoftware (Table 4). Free-text responses about making the hand-off accessible to nursing staff indicated some apprehension withconcerns that “nuisance” calls from nursing staff would increaseand that the handoff sheets could be saved to “use againstphysicians” should something go wrong. Respondents were alsoconcerned that if given access to the physician handoff, nursingstaff “may make assumptions regarding the patient and [make]changes in his/her status without contacting [the physician].”Respondents also felt that giving nurses access to the handoffmight create more confusion, especially if care plans were mis-interpreted and then discussed with patients. Interestingly, anapproximately equal number of respondents were in favor ofallowing nursing staff access to the physician handoff.

DiscussionThere is considerable variability in information content trans-ferred during handoffs, and lack of standardization may lead toomission of vital information such as code status and patientlocation. The few studies that have examined the handoffprocess confirm that users are dissatisfied with current handoffsystems, which they find to be variable, unstructured, andprone to error.11,15,21,22

Through abstraction of paper handoff sheets and responsesto housestaff questionnaires, our evaluation demonstrated thehandoff software was associated with fewer omissions of vitalinformation without causing commission of information andcould be configured to automatically import and update datafrom the EMR. The questionnaires suggested that the handofftool was well received by users and that it improved perceptionsof quality and safety of patient care, efficiency, and comfortwith the patient handoff. The variability in percentages of base-line data available on the handoff sheets between sites, as wellas the feedback obtained during conference calls, demonstratesthat the new handoff software can be adopted at diverse loca-tions with overall improvement in handoff quality, even at loca-tions with a standardized handoff method.

For physician communication, there was a high level ofcommunication at baseline, which was essentially unaffected byuse of the handoff software. Because of concern that anEMR–based handoff tool could lead to less face-to-face com-munication, it is notable that no such postimplementation




change was observed. Nor did perceptions of communicationwith nursing change with implementation of the software.These measures provide some validity to the questionnairebecause we would not expect verbal handoff practices or per-ceptions of communication with nursing staff to change on thebasis of use of a standardized handoff software program forphysicians.

Because physician handoffs are not part of the permanentmedical record, it is not the norm for these written documentsto be available to nursing staff. Questionnaire responses indicat-ed that some physicians were apprehensive about allowing nurs-ing staff access to their handoffs, whereas others welcomed theidea. Emerging evidence suggests that both nursing satisfactionand nursing perception of patient care quality are improvedwith provision of the written physician handoff to nursing staff,without adverse consequences to physicians.23–25 In addition,more than half of physician training programs do not have away to let nurses know about transfers of care.8 Nursing accessto the physician handoff eases confusion related to shift change,improving both nursing satisfaction and patient care. The ques-tion of nursing access should be further investigated and consid-ered with future refinements to the handoff software.

Our findings support the perception that variability adverse-ly affects handoff quality in terms of both omission and com-mission of information.26 Residents were generally concernedwith the variable quality and quantity of information providedin patient handoffs and wanted more concise communication.The design team elected to force users to type pertinent infor-mation regarding patients’ clinical course into free-text fieldsinstead of importing directly from the daily progress note. Thiswas intended to prevent commission of information and exces-sively long handoff reports while creating a flexible systemcapable of including data elements unique to a given patient.Anecdotally, the decrease in handwritten comments after use ofthe handoff software suggests that covering physiciansencounter fewer unexpected scenarios during their shifts. It hasbeen stated that a “disadvantage of electronically linked systemsis the inability to tailor data to the critical data elements for thegiven patient”16(p. 395) Yet the handoff tool design includes thenecessary elements to prevent omission of vital informationwhile at the same time providing flexibility in information provided.

Our findings also support previously published literaturethat existing handoff methods are frequently incomplete anddemonstrates a need for improvement at shift changes. Theamount of time required to prepare and execute the handoffdirectly influences its content,9,21 and although the decrease in

self-reported time to complete the handoff sheet did not reachstatistical significance, both objective measurement of handoffcontent quality and perception of quality improved, with lesstime invested per patient.

LimitationsThere are a number of limitations of our evaluation that shouldbe considered. First, we restricted our analysis to only the aca-demic-affiliated VAMCs for which complete data were present,thus limiting generalizability of the results. Yet, data from addi-tional sites (we collected handoff sheets from two additionalsites and obtained questionnaires from three additional sites)supported our overall conclusions. In addition, all sites thatparticipated in the project provided ongoing feedback duringthe regular conference calls and through e-mail. Another limi-tation is that it is not possible to know if omissions in handoffcommunications discovered in this study led to harm in patientcare. When evaluating pre-implementation handoff sheets, noattempts were made to quantify the accuracy or completenessof information, including medication lists. However, in addi-tion to being more frequently present, medication listsappeared to be more comprehensive with use of the handoffsoftware. An unexpected finding that not all fields automatical-ly entered by the handoff software were 100% present aftersoftware implementation (Table 2) was attributed to program-ming errors in the software that have since been corrected.Finally, although we were able to demonstrate objectiveimprovements in consistency of information transfer, adverseevents directly attributable to poor handoffs are rare and diffi-cult to evaluate and collect, so that our evaluation necessitatedthe use of surrogate quality measures.

SummaryThe new handoff software improved both consistency of infor-mation transfer and perceptions of patient handoff within theVA health care system. Although the tool focuses on the hand-off between physicians at shift change, other potential applica-tions of the software include daily team rounds and en -hancement of communication between nurses and physicians.A final version of the handoff software was incorporated intothe CPRS and made available in June 2008 to all VAMCs. The work reported here was supported by the Department of Veterans Affairs,

Veterans Health Administration, VA Quality Scholars Fellowship Program (Drs.

Anderson, Curtis, Abrams, Karnani, Cannon, and Kaboli). Dr. Kaboli is supported

by a Research Career Development Award from the Health Services Research and

Development Service, Department of Veterans Affairs (RCD 03-033-1). The views

expressed in this article are those of the authors and do not necessarily represent

the views of the Department of Veterans Affairs. The authors thank the participat-

ing VA medical centers: Iowa City, Iowa; Washington, DC; White River Junction,

J




Vermont; Loma Linda, California; Dallas, Texas; and Altoona, Pennsylvania. This

work would not have been possible without the efforts of the software design team,

programmers, and information technology personnel, particularly Charlet Lynn

Cottee, Richard Sowinski, and Daniel Rosenthal, M.D. Previously used physician

surveys were provided by Jeffery Schnipper, M.D. (Brigham and Women’s Hospital,

Boston) and Mindy Flanagan, Ph.D. (Indianapolis VA Medical Center). The authors

also thank the physicians and VA administrators who beta-tested this software and

participated in the data collection.

References1. Mills P., et al.: Teamwork and communication in surgical teams:Implications for patient safety. J Am Coll Surg 206:107–112, Jan. 2008.2. Petersen L.A., et al.: Does housestaff discontinuity of care increase the riskfor preventable adverse events? Ann Intern Med 121:866–872, Dec. 1, 1994.3. Sutcliffe K.M.: Communication failures: An insidious contributor to med-ical mishaps. Acad Med 79:186–194, Feb. 2004.4. The Joint Commission: Handoff Communications: Toolkit for Implementingthe National Patient Safety Goal. Oakbrook Terrace, IL: Joint CommissionResources, 2008.5. The Joint Commission: 2010 Comprehensive Accreditation Manual forHospitals: The Official Handbook. Oakbrook Terrace, IL: Joint Commission

Resources, 2009.6. Department of Veterans Affairs (VA) Iowa City Health Care System: HealthCare Hand-Off Communication. 106–107, Dec. 22, 2006. (last accessed Nov.2007, unavailable outside VA).7. U.S. Department of Veterans Affairs: Findings and Recommendations forImproving Patient Handoffs in the VHA. Dec. 7, 2005. (last accessed November2007, unavailable outside VA).8. Horwitz L.I., et al.: Transfers of patient care between house staff on inter-nal medicine wards: A national survey. Arch Intern Med 166:1173–1177, Jun.2, 2006.9. Solet D.J., et al.: Lost in translation: Challenges and opportunities in physi-cian-to-physician communication during patient handoffs. Acad Med80:1094–1099, Dec. 2005.10. Patterson E.S., et al.: Handoff strategies in settings with high consequencesfor failure: Lessons for health care operations. Int J Qual Health Care16:125–132, Apr. 2004.11. Roughton V.J., Severs M.P.: The junior doctor handover: Current practicesand future expectations. J R Coll Physicians Lond 30:213–214, May–Jun.1996.12. Vidyarthi A.R., et al.: Managing discontinuity in academic medical cen-ters: Strategies for a safe and effective resident sign-out. J Hosp Med1:257–265, Jul.–Aug. 2006.13. Shaw G.: Handoff Rx: Knowing what to say and how to say it.Standardized forms, written and verbal reports help residents avoid potentialproblems during transfers. ACP Observer, pp. 6–11, Oct. 2006.14. Arora V., et al.: Communication failures in patient sign-out and sugges-tions for improvement: A critical incident analysis. Qual Saf Health Care14:401–407, Dec. 2005.15. Ram R., Block B.: Signing out patients for off-hours coverage:Comparison of manual and computer-aided methods. Proc Annu SympComput Appl Med Care, pp. 114–118, 1992.16. Philibert I., Leach D.C.: Re-framing continuity of care for this century.Qual Saf Health Care 14:394–396, Dec. 2005.17. Petersen L.A., et al.: Using a computerized sign-out program to improvecontinuity of inpatient care and prevent adverse events. Jt Comm J QualPatient Saf 24:77–87, Feb. 1998.18. Volpp K.G.M., Grande D.: Residents’ suggestions for reducing errors inteaching hospitals. N Engl J Med 348:851–855, Feb. 27, 2003.19. Van Eaton E.G.: A randomized, controlled trial evaluating the impact of acomputerized rounding and sign-out system on continuity of care and residentwork hours. J Am Coll Surg 200:538–545, Apr. 2005.20. Van Eaton E.G.: Organizing the transfer of patient care information: Thedevelopment of a computerized resident sign-out system. Surgery 136:5–13,Jul. 2004.21. Lee L.H., et al.: Utility of a standardized sign-out card for new medicalinterns. J Gen Intern Med 11:753–755, Dec. 1996.22. Australian Commission on Safety and Quality in Health Care: ClinicalHandover and Patient Safety. May 2005. http://www.health.gov.au/internet/safety/publishing.nsf/Content/clinical-handover (last accessed Dec.15, 2009).23. Behara B., et al.: A conceptual framework for studying the safety of tran-sitions in emergency care. In Henriksen K. (ed.): Advances in Patient Safety.Agency for Healthcare Research and Quality, 2005, vol. 2, pp. 309–321.http://www.ahrq.gov/downloads/pub/advances/vol2/behara.pdf (last accessedDec. 15, 2009). 24. Doyle E.: To keep nurses in the loop, this hospital gave them access to itssign-out system. Today’s Hospitalist, pp. 20–23, Jul. 2006.25. Sidlow R., Katz-Sidlow R.J: Using a computerized sign-out system toimprove physician-nurse communication. Jt Comm J Qual Patient Saf32:32–36, Jan. 2006.26. Philibert I.: Adapting Strategies from High-Reliability Organizations toImprove Patient Hand-offs in Teaching Hospitals (Ph.D. diss.), University ofIowa, Iowa City, 2008.

Jaclyn Anderson, D.O., M.S., formerly Department of Veterans

Affairs (VA) National Quality Scholars (VAQS) Fellow, VA Medical

Center, Center for Research in the Implementation of Innovative

Strategies in Practice (CRIISP), and Department of Internal

Medicine, University of Iowa Carver College of Medicine, Iowa City,

Iowa; is Rheumatology Fellow, University of Nebraska Medical

Center, and Omaha VA Medical Center, Omaha. Divya Shroff,

M.D., is Associate Chief of Staff–Informatics, VA Medical Center,

Washington DC; and Assistant Professor of Medicine, VA Medical

Center, and Department of Medicine, George Washington

University, Washington, DC. Ann Curtis, M.D., is VAQS Fellow,

Iowa City VA Medical Center, CRIISP, and the Department of

Internal Medicine, University of Iowa Carver College of Medicine.

Noel Eldridge, M.S., is Executive Officer, VA National Center for

Patient Safety, Washington, DC. Katrina Cannon, M.D., M.S., for-

merly VAQS Fellow, Iowa City VA Medical Center, CRIISP, and the

Department of Internal Medicine, University of Iowa Carver College

of Medicine, is a Faculty Member, Genesis Quad Cities Family

Medicine Residency Program, Davenport, Iowa. Rajil Karnani,

M.D., M.M.E., formerly VAQS Fellow, Iowa City VA Medical Center,

CRIISP, and the Department of Internal Medicine, University of

Iowa Carver College of Medicine. Thad Abrams, M.D., M.S., is

Core Investigator, CRIISP, Iowa City VA Medical Center, and

Assistant Professor of Medicine, Department of Internal Medicine,

University of Iowa Carver College of Medicine. Peter Kaboli, M.D.,

M.S., is Core Investigator, CRIISP, Iowa City VA Medical Center;

and Associate Professor of Medicine, Department of Internal

Medicine, University of Iowa Carver College of Medicine. Please

address correspondence to Peter Kaboli, [email protected].

Online-Only Content

See the online version of this article for

Appendix 1. Shift Handoff Tool: Display for Printing

8


AP1 February 2010 Volume 36 Number 2


Appendix 1. Shift Handoff Tool: Display for Printing

Online-Only Content8

When patient information is complete, a printout is generated to facilitate the handoff and to be a resource for the covering physician.


Date post:	06-Jun-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Understanding and Improving Patient Handoffs · tion, higher costs, longer hospital stays, more...

Documents