The Emergency Department is usually the only
part of the hospital that has multiple doctors and nurses working at full speed
24 hours/day. A night shift as an ED registrar rarely allows for a wink of
sleep, and we are constantly chopping and changing between day and evening
shifts when not on night duty, punishing our bodies. Finishing work on time is
essential to maintain any routine (and sanity!) in your life as an ED doctor.
Therefore,
we buy into the shiftwork mentality – it is
the clock that dictates when we leave work, not how long our list of patients
is.
Despite the transiency of patients in the
ED, (most go home, and those that are admitted usually get rushed upstairs so
that the ‘4-hour target’ is adhered to -certainly in the UK, increasingly so in
Australia) a consequence of shiftwork culture in the ED is that an undeniably
large part of our job is ‘handover’.
What
is ‘handover’?
‘Handover’ (or ‘signout’ /’handoff’ in the
US) is a process that occurs at shift change where a doctor finishing their
shift transfers the responsibility of his/her patients to a
receiving doctor who is starting.
Multiple pieces of information require transmission
from one brain to another:
- Patient demographics
- Working diagnosis/differential
- Background
- Physical examination/investigations/referrals made
- Plan/likely disposition
- Expectations of near future – flag management priorities if any deviation from expected clinical path
- Working diagnosis/differential
- Background
- Physical examination/investigations/referrals made
- Plan/likely disposition
- Expectations of near future – flag management priorities if any deviation from expected clinical path
The process of handover varies
significantly depending on institution:
- The people involved:
- One-on-one meeting/conversation between two doctors
- Formal doctors meeting (often called ‘handover’), where cases are discussed by the whole group before the responsibility of a patient is allocated to a receiving doctor - usually overseen by a consultant.
- Other members of the multidisciplinary team (MDT) can be present at handover (nurses, pharmacists, physiotherapsists)
- Where in the process occurs:
- In the central clinical area, usually beside computers
- In a separate area of the department - away from the busy, chaotic clinical area
- At the bedside
- Structure:
- Oral only, with perhaps occasional jotting on scrap paper
- Oral and written (either handwritten, or electronic medical record)
- Written only
- Use of handover mnemonic (SBAR, I-PASS, SIGNOUT etc)
Is
handover a source of error?
A sentinel event is ‘any unanticipated event in a healthcare setting resulting in death or serious physical or psychological injury to a patient or patients, not related to the natural course of the patient's illness’. A sentinel event occurs as a result of an error from a healthcare professional.
A sentinel event is ‘any unanticipated event in a healthcare setting resulting in death or serious physical or psychological injury to a patient or patients, not related to the natural course of the patient's illness’. A sentinel event occurs as a result of an error from a healthcare professional.
The Joint Commission in the US report that 70% of sentinel
events are caused by communication breakdown. Of these, 50% occur during the transition
of care from one healthcare professional to another – i.e. during handover (1).
What
contributes to poor handover in the ED?
Departmental
Factors
The ED is a chaotic working environment
with frequent interruptions. It has been estimated that consultants are
interrupted on average every 9 minutes during the course of a shift, and
residents every 14 minutes (2). An
unpredictable, loud, and disruptive environment is suboptimal for the sensitive
handover process.
When the department is particularly busy,
there is increased pressure on doctors to ‘pick up’ patients. The more patients
a doctor is taking care of (particularly late in a shift), the more information they’ll have
to handover – more opportunity for error (especially if that doctor is
desperate to leave on time).
The culture of the department may dictate
that handover is viewed as an informal process. This creates a fertile environment for error:
- Handover may not receive an appropriate amount time or effort.
- Despite taking nominal responsibility of a handed patient, receiving doctors may not invest as much time/energy as that patient is ‘second hand’ – how many times have you heard a colleague being grilled by a boss about a patient, and defensively responding with ‘well it was a handover..’ – as if that is an acceptable excuse for not knowing all the important details regarding a sick patient they are responsible for.
- The hierarchy of a department might dictate that juniors handover to other juniors without senior supervision.
- Lack of MDT involvement: this can lead to nursing staff not knowing which doctor has the responsibility of the patient at a certain point in time - ‘ambiguous moment of transition of care’ (3).
Individual
Factors
Task saturation/multitasking is commonplace
in a busy ED. Individuals may be so busy that they may not allocate enough time
to give/take adequate handover. A departmental culture with a lax attitude
towards handover compounds this problem.
An individual may be desperate to finish a shift
for personal reasons, and rush the handover process. Again, this problem can be
facilitated by suboptimal departmental culture.
Individual cognitive bias can play a role
in faulty handover leading to error:
- Receiving doctor trusting erroneous
information – ‘diagnosis momentum’.
- Receiving doctor misinterpreting information due to their own biases. ‘Anchoring’ occurs - focusing on one aspect of the presentation (and ignoring others) because that particular element was important in previous memorable cases (3).
- Receiving doctor misinterpreting information due to their own biases. ‘Anchoring’ occurs - focusing on one aspect of the presentation (and ignoring others) because that particular element was important in previous memorable cases (3).
Doctors may not use a standardized
approach:
- Individual preference/style is often preferred (particularly by senior clinicians who have practiced in a certain way for many years). This creates unpredictable and variable handover content, which can obstruct the transmission off vital information.
- Too much information may be handed over – overwhelming; more difficult to elicit key elements that will potentially require action.
- Poor communication techniques:
- No 'closed-loop'/'repeat back' communication.
- 'Unidirectional' communication (i.e. not bidirectional, where the receiving doctor asks questions and clarifies content) (3, 4).
Wider
issues
An ethos within medicine, particularly
within Emergency Medicine, is to rule out the worst-case scenario for a given
presentation, and then work backwards. We like to hunt for ‘red flags’ that
highlight potentially high-risk patients with a worrying underlying diagnosis.
We know what the red flags are when considering a patient presenting with back
pain or dyspnoea, but we don’t know what the red flags are when considering a handover.
Perhaps a patient who has required consultant input, or has pending imaging
should be flagged as higher risk to the receiving doctor. No evidence currently
exists that shows certain elements of a handover to represent a high-risk scenario.
Handover technique is not in the medical
school curriculum. We learn how to take a detailed history, perform a physical
exam, and then present our findings to a senior doctor, but we don’t learn how
to succinctly structure a handover. This leads to the common problem of very
junior doctors overwhelming a receiving doctor with information (see above).
So… The Paper
“Changes in Medical Errors after Implementation of a Handoff
Program”
This paper from the New England Journal came to my attention by being named a 'Hall of Famer' in Research and Reviews in the Fast Lane.
It was also reviewed by Ryan Redecki over at Emergency Medicine Literature of Note - 'It's a Patient Handoff Miracle'.
What were they Pondering?
Will implementation of a standardised handoff (handover) program reduce medical errors/adverse events/miscommunication?
Will this program increase the workload of residents?
What type of study was it?
Multi-centre prospective intervention study.
Multi-centre prospective intervention study.
Pre-intervention data was
collected for a 6-month period. Then there was a 6-month period of implementing
the intervention. Then post-intervention data was collected for a further 6-months.
Who
were they studying?
They were studying the
residents at nine paediatric residency programs in the US and Canada.
None of the sites studied
had a standardized handover procedure in place.
875 residents participated.
What was the intervention?
Implementation of the
‘I-PASS Handoff Bundle’ – 7 elements:
- Introduction of the I-PASS mnemonic
- 2-hour workshop
- 1-hour role-playing and simulation session (practicing skills from the workshop)
- Computer module for independent learning
- Faculty development program
- Direct-observation tools for faculty to provide feedback to residents
- Departmental culture-change campaign (logo, posters etc)
The I-PASS structured handover tool was integrated into verbal and written
handovers at all sites. Both verbal and written handovers were expected for
every patient.
What
was the Primary Outcome?
2-component primary
outcome:
- Rates of medical errors (overall
errors)
- Rates of preventable
adverse events (errors that led to patient harm)
What
else was being looked at?
The quality of written and
verbal handovers was evaluated. This was performed by recording the inclusion
of key quality elements before and after the intervention.
Key quality elements for
handover:
- Illness severity assessment
- Patient summary
- To do list
- Contingency plans
- Readback by receiver (oral
handover only)
Resident workflow patterns
were evaluated. Did the intervention increase the amount of time residents
spent handing over, and therefore, reduce the time they spent with patients?
How
were medical errors identified?
10, 740 patient admissions
(5516 pre-intervention and 5224 post-intervention) were reviewed for the
presence of medical error.
Possible errors were
identified by a research nurses reviewing medical records, formal incident
reports, or solicited reports from nurses/residents in postshift surveys.
2 physician investigators
interrogated each suspected error and classified it into:
- An adverse event (patient
harm due to medical care) – further classified into preventable (due to medical
error) and non-preventable.
- A non-harmful error (or 'near miss').
- Neither (excluded).
- A non-harmful error (or 'near miss').
- Neither (excluded).
Subtypes of error:
- Errors related to
diagnosis (incorrect, delayed, omitted)
- Errors related to therapy other than medication or procedure
- Errors related to history/physical examination
- Multifactorial errors
- Medication-related errors
- Procedure-related errors
- Falls
- Nosocomial infections
- Errors related to therapy other than medication or procedure
- Errors related to history/physical examination
- Multifactorial errors
- Medication-related errors
- Procedure-related errors
- Falls
- Nosocomial infections
The
results…
Overall
errors?
There was a 23% relative
reduction in the rate of all medical errors (before vs after the
intervention, 24.5 vs. 18.8 errors per 100 admissions, P<0.001).
Significant error
reductions were seen in 6 of the 9 sites.
Adverse
events?
There was a 30% relative
reduction in the rate of preventable adverse events (4.7 vs. 3.3 events per
100 admissions, P<0.001).
Non-harmful
medical errors?
There was a 21% relative
reduction the rate of non-harmful medical errors (19.7 vs. 15.5 non-harmful
errors per 100 admissions, P<0.001).
What
subtypes of error were reduced?
There were significant
reductions in:
- Diagnostic errors
- Errors related to history/examination
- Errors related to therapies
other than medication/procedure
- Multifactorial errors
Did
the quality of handovers improve?
Significant improvements were
found in the quality of written and oral handovers.
All key quality elements for
written and oral handover saw a significant improvement in their inclusion.
Did resident
workflow patterns change?
No significant change in mean
duration of oral handover sessions.
No significant change in
the percentage of time in a 24-hour period spent in contact with patients and
families, and performing handover (written and oral).
Stengths
of this study…
- Multicenter trial.
- Pre-intervention and post-intervention data was collected at the same time of year – safeguards against time-of-year confounding (i.e. difference in resident experience; patient population).
- The physician investigators who categorized the suspected errors were blinded (however the nurses collecting the initial data were not).
- A single-centre study was performed beforehand as an exploratory study – this saw similar improvements in error reduction. Key differences in the intervention bundle:
- SIGNOUT mnemonic used instead of I-PASS
- As part of the intervention bundle, a team-based handover structure was implemented– all handovers took place in the presence of the whole team, in a quiet room.
- Did not include role-playing/sim session, computer module, faculty development program, direct-observation tools for faculty, or a culture-change campaign.
- There is a 'Paper Chase' on EM:RAP discussing this paper.
Limitations
of this study…
- This study was performed in the paediatric setting. It is difficult to know whether we can extrapolate the results to the emergency medicine/critical care setting.
- The study focuses on one-to-one handovers. It doesn’t introduce a team-based handover structure in the intervention bundle (unlike the exploratory study).
- The two physician investigators had only a 70% agreement (Kappa score 0.47) when classifying incidents into an adverse event/non-harmful error/exclusion; and only a 72% agreement (Kappa score 0.44) when further classifying adverse events into preventable/non-preventable. There was, therefore, a considerable amount of disagreement between the investigators, which makes the data less reliable.
- A 'Kappa score' is a statistic that removes the percentage agreement that would occur by chance, and calculate true inter-rater agreement.
- Kappa scores of 0.47/0.44 indicate, at best, a moderate inter-rater agreement.
- The ‘I-PASS handoff bundle’ was a multi-pronged intervention. It is, therefore, difficult to tease out which elements of the bundle made the difference. Was it the introduction of the I-PASS mnemonic? Was it the various educational interventions? Was it the overall change in departmental culture as a result of participation in this study?
- By virtue of participating in this study, the handover process will have been in the spotlight at each of the participating sites, which may have contributed to the results. Will the observed error reductions be maintained when research is no longer being conducted?
Final
thoughts
- Handover is a massive part of our daily practice as ED doctors.
- A suboptimal handover process (no standardization, informal departmental attitude, minimal senior input) creates an environment that is extremely vulnerable to error.
- It has been shown in this study that standardizing the process and educating doctors/raising awareness of the importance of handover safeguards against medical error– i.e. An overall shift in departmental culture is the key.
- Instead of a minefield for medical error, perhaps a healthier way to view handover is as an opportunity to take a step back, review the case and collaborate with the receiving doctor who has a fresh perspective. It is also a fantastic platform for case-based teaching from seniors.
References
1. “Changes in Medical Errors after Implementation of a Handoff Program”
2. ''The multitasking clinician: decision-making and cognitive demand during and after team handoffs in emergency care.''
3. ''Improving Handoffs in the Emergency Department''
4. ''Reducing Error in the Emergency Department: A Call for Standardisation of the Signout Process''
5. ''Rates of medical errors and preventable adverse events among hospitalized children following implementation of a resident handoff bundle.''
No comments:
Post a Comment