Patient Safety Tip of the Week
April 7, 2015 Missing Patients and
Death
Though our first column
on missing patients was 6 years ago (see our July 28, 2009 Patient
Safety Tip of the Week “Wandering,
Elopements, and Missing Patients”) there was considerable renewed
interest in the topic after a patient was found dead 17 days after going
missing at a San Francisco hospital in 2013 (see our October 15, 2013 Patient Safety Tip of the Week “Missing
Patients” and our December 2013 What’s New in the Patient Safety World
column “Lessons
from the SFGH Missing Patient Incident”).
Despite considerable
discussion regarding lessons learned from that unfortunate incident, similar
cases continue to occur. Several are described below:
An elderly patient
suffering from dementia who wandered out of a scandal-hit hospital was found
dead in a ditch a quarter of a mile from the grounds the next day (Duffin
2015). She was said to be a
‘known wanderer’, was able to walk out of the hospital, wearing only a cardigan
and trousers. She was reportedly seen by a member of staff but not stopped.
This was especially problematic since a report, following an inspection in June
2014, stated: ‘The support for patients living with dementia or who may have a
learning disability was inconsistent.’ It also raised concerns about staff
levels, saying there was a ‘heavy reliance’ on agency staff which, in some
instances, ‘affected the delivery and continuity of patient care’.
A 69-year-old man
with diminished mental capacity died from complications of hypothermia three
days after he walked away from an assisted living facility (Bear
2015).
A 58-year-old
man, hospitalized at a San Diego hospital because of head and neck trauma, went
missing from the hospital and was found dead at the bottom of a nearby canyon 5
days later (CDPH
2014). This case has the most to learn from because the CDPH publishes the
statement of deficiencies (SOD) and hospital plan of correction (POC). Because
he was considered at risk for falling, he had both a bed alarm and a video
monitor. The patient had confusion that would wax and wane. One morning the
Video Monitor Technician notified a Clinical Care Partner (CCP) that the
patient was getting out of bed. That CCP went to the patient’s room and found
the patient to be missing. The CCP notified the RN, who notified the Charge
Nurse. The Charge Nurse called the Security pager number twice and pressed a
panic button twice while the CCP and RN went looking for the patient. The CCP,
after searching the nursing unit, went down the elevator to the first floor and
basement looking for the patient. When the CCP returned to the nursing unit,
security was still not there so she contacted them by phone. The panic button
was not working. Response by Security was delayed. Five days later the patient
was found deceased at the bottom of a canyon adjacent to the facility’s parking
structure.
The Director of Security said the pagers should not be used
to contact Security in the case of an emergency. Instead, staff should press
the panic button and dial 6111. There was actually no written process telling
staff how Security should be contacted. In addition, the security technician
had been aware that the panic button on that nursing unit had been non-operable
for 8 days before it was fixed.
On later review of video tapes the patient was seen exiting
the hospital via the front doors, wearing hospital gown, non-skid socks and a
cervical traction device. Staff noted it was not uncommon to see patients
leaving the building in gowns to smoke outside. So when staff ignore such
occurrences it is really a form of “normalization of deviance” where a
deviation of behavior is so common we consider it “normal”. That desire to
smoke is a risk factor for patient elopement. In our December 2012 What’s New in the Patient Safety World column “Just
went to have a smoke…” we joked about a scene in the Denzel Washington
movie “Flight” in which 3 hospital inpatients all came to the same hospital
stairwell to have a smoke. We wondered how often that might happen and that was
soon answered: 18.4% of patients who smoke will smoke at some time during their
inpatient hospitalization (Regan
2012).
The hospital’s Plan of Correction (POC) focused heavily on
the non-operable panic button. They actually found that many panic buttons in
other areas were also non-operable. The hospital ultimately upgraded to a
version of the panic button that allows real-time notification of failure of
communication/connectivity. Security immediately follows up. If a reboot of the
computer fails to correct the problem a technician responds to fix it within 24
hours. Importantly, while the button is non-operable a sticker is placed over it indicating it is “non-operable” and
having instructions on how to
contact Security. Note that we’ve seen similar problems with panic buttons
elsewhere, such as behavioral health units (see our February 4, 2014 Patient
Safety Tip of the Week “But
What If the Battery Runs Low?”). It is clear that any time you have a
critical alert system you must ensure it is functioning at all times so
programs like daily testing are a must.
But we might make a case that the heavy focus on the panic
buttons was, in fact, too heavy. Panic buttons are a sort of “silent” alarm and
they alert only a very few people and don’t really tell them what is wrong or
what to look for. Some features of this case are eerily similar to the one at
San Francisco General in 2013 (see our October
15, 2013 Patient Safety Tip of the Week “Missing
Patients” and our December 2013 What’s New in the Patient Safety World
column “Lessons
from the SFGH Missing Patient Incident”) though it should be noted this
case actually occurred prior to the SFGH case. Valuable time was lost while a
single person was searching for the patient. And in both cases it appears that
too many people considered it the responsibility of “Security” to search for
missing patients and no true system-wide response ever occurred. That should
never be the case. While “Security” may command and coordinate a search, there
must be a coordinated effort that involves staff from all parts of the
hospital. There is no mention of the search protocol in the SOD or POC
in the San Diego case.
When a patient is
discovered to be missing a brief search of the local unit should be done. If
the patient is not found immediately, the “code” for a missing patient should
be issued. That is called to the hospital operator and announced over PA
system. While most hospitals still use arcane codes for various emergencies,
many are moving to “plain language” codes for these alerts. In the case of a
missing patient a “plain language” code with the description of the missing
patient has the additional advantage that visitors and other patients in the
hospital might identify the missing patient. Once the “code” is announced the
facility should have a predetermined search grid where every area of the
hospital and surrounding grounds has a designated person to search it. A
central command post is set up and staff call in to that post once they have
searched their sector.
Yes, such procedures
are disruptive. All the workers who participate in such searches have other
duties and care for all the other patients cannot be abandoned. It’s just like
responding to a cardiac arrest. But sector searches can be done in an expedited
manner as long as the search sectors are not too large, staff know their
responsibilities, and communication to the command post is simple.
Some facilities use
an alternative system to alert staff to missing patients. They may use text
messages sent out in a “blast” fashion to computer terminals and staff
smartphones, the equivalent of the well-known “Amber alert” when a child in the
community goes missing. This sort of system can have the advantage of including
a description of the patient and perhaps a photograph. But the disadvantages
are that it may take longer to implement and you lose the chance that a visitor
or other patient might identify the missing patient.
In these incidents and other cases of missing patients we’ve
discussed in the past, several themes recur:
- Someone often sees them
but does not recognize they are wandering
- No system to help other
staff identify patients who have been labeled at risk
- Automatic locking doors
may prevent re-access to the unit
- No coordinated response
- Overreliance on “Security”
to find them
- Taking too long to issue a
“missing patient code”
- No assessments for
wandering risk
All types of facilities need to develop policies and
procedures for:
1) doing
an assessment for risk of wandering
or elopement
2) implementing
risk reduction strategies for those
patients at risk
3) performing
a prompt and thorough search when a
patient is missing
Assessment for Wandering
Risk
We refer you to our Patient Safety Tips of the Week for July
28, 2009 “Wandering,
Elopements, and Missing Patients” for a discussion on the assessment
of patients for risk of wandering. Also remember our comments above on the urge
to smoke tobacco as a potential risk factor for wandering and elopement.
Reducing the Risk of
Wandering
So what do you do when you identify a patient as being
at-risk for wandering or elopement? It makes sense to put them in a room where
staff would be more likely to see them exit the room (usually closer to the
nurse’s station). Many floors have one or two rooms that are video monitored, a
logical choice for such patients. Consider having the patient wear a gown that is a different color than the
usual gowns so that all staff would recognize such patient as being “lost” if
encountered in other parts of the hospital. One hospital we worked with used an
off-purple color to flag such patients. The San Diego hospital in today’s index
case began using orange colored arm
bands to identify patients at high risk of wandering. But make sure your
staff all know what those colors mean so they can take action if they see such
a patient somewhere they don’t belong.
Potential exit doors
on the unit should be fitted with appropriate alarms (that are functioning
correctly) and with appropriate signage to keep the door closed. But also
consider the downside of locked doors. Many units have doors that are locked
automatically and prevent access to the unit. So you need to consider what
would happen if a patient wandered out such a door and could not get back in.
Such may have been the case in the SFGH incident previously described. We even
recall once getting a cellphone call from a physician who responded to a
cardiac arrest and got locked in a stairwell and could not get back in to any
unit! Also consider that there are times when your locked doors will be
automatically unlocked (eg. fire alarms) creating an
opportunity for a wandering patient to leave the unit.
Consider keeping the patient in a room with a roommate or
have family members stay in the room. Attention to the patient’s physical needs
(food, water, warmth, pain management, toileting) are important. Letting the
patient walk or exercise under supervision may be useful. Our Patient Safety
Tips of the Week for July 28, 2009 “Wandering,
Elopements, and Missing Patients” had links to several websites
having resources related to management of the patient at risk for wandering.
Construction sites
are particularly vulnerable for a few reasons. First, you often have outside
workers there who are not thinking about patient safety. So they may leave
doors unlocked. Second, construction sites have lots of opportunities for
someone to injure themselves. So make sure you pay close attention to any sites
at your facility with ongoing construction.
Be especially careful during patient transports. One of the items we recommend including in your “Ticket
to Ride” checklist/communication tool for transports (eg
to Radiology) is information about wandering risk.
Ultimately, we will look to technology solutions to help in cases of lost or missing patients. Technology
already allows us to find our car keys, track our dogs, and locate hospital
assets like wheelchairs so technology solutions to track patients are a logical
direction. However, we’ll emphasize that technology will always be only part of
the response to missing patients. It should
never be the sole modality relied upon. The three most obvious technology
tools are GPS, RFID, and Bluetooth. A
VA analysis suggests GPS beats RFID in most scenarios (VA 2013). But
the specific technology chosen will likely differ from facility to facility and
may depend upon the need to integrate with other technology needs. For example,
though GPS probably would be best for locating patients who have left the
facility or hospital grounds, some hospitals may prefer RFID because they are
using an RFID system for inventory tracking. Bluetooth would have limited
applicability. Many of you probably already use low-power Bluetooth for
tracking items like your keys (or maybe your TV remote!). But in the context of
missing patients Bluetooth applications would most likely only be of use in
alerting staff when a patient leaves the Bluetooth receiving area (about the
size of a typical inpatient floor). Similarly, the bracelets used in newborns
to prevent abduction can alert staff when a patient leaves the unit but are of
little help in locating a patient once they have already left the unit.
Bed alarms are
another technology that can alert staff when a patient gets out of bed. Their
biggest applicability is in prevention of patient falls, though their success at
even that has been less than stellar (see our January 2013 What’s
New in the Patient Safety World column “Bed
Alarms Fail the Test”). Our own experience with bed alarms has not been
particularly good. We often see them malfunction or be improperly installed
(see our Patient Safety Tip of the Week “Unintended
Consequences of Technological Solutions”) or be disconnected intentionally
because they are alarming too frequently. In addition, as in the current CDPH
case, patients may still end up leaving the unit despite using bed alarms.
How about low-tech
tools? Multiple facilities, particularly long-term care facilities that
care for many patients with dementia, often use tricks to disguise doors or
otherwise encourage wandering patients to avoid doors. One Canadian hospital
uses a door wrap in its continuing care facility that makes the door look like
a cabinet full of things (Whitnall
2015). That helps prevent patient with dementia and wandering from
going to the doors. Others have used the image of a “black hole” near doors to
discourage wandering patients from going to the doors.
Performing a Prompt
and Thorough Search if a Patient Goes Missing
Despite our best efforts to identify patients at high risk
of wandering and absconding and implementing risk reduction strategies, patients
will still wander and get lost. If a patient does go missing, doing a thorough
search promptly is crucial. For those missing more than 24 hours, the death
rate can be as high as 50% (VA 2013).
When a patient is
discovered to be missing a brief search of the local unit should be done. Staff
on the unit need to be notified as soon as a patient is missing. A very brief
head count of patients and look in rooms on a unit and any adjacent closets, stairwells
or elevators is typically done but this should last no more than a couple
minutes.
If the patient is
not found immediately, the “code” for a missing patient should be issued. That
is usually called to the hospital operator and announced over PA system. While
most hospitals still use arcane codes for various emergencies, many are moving
to “plain language” codes for these alerts. In the case of a missing patient a
“plain language” code with the description of the missing patient has the
additional advantage that visitors and other patients in the hospital might
identify the missing patient. Once the “code” is announced the facility should
have a predetermined search grid where every area of the hospital and
surrounding grounds has a designated person to search it. A central command
post is set up and staff call in to that post once they have searched their
sector.
Yes, such procedures
are disruptive. All the workers who participate in such searches have other
duties and care for all the other patients cannot be abandoned. It’s just like
responding to a cardiac arrest. But sector searches can be done in an expedited
manner as long as the search sectors are not too large, staff know their
responsibilities, and communication to the command post is simple.
Some facilities use
an alternative system to alert staff to missing patients. They may use text
messages sent out in a “blast” fashion to computer terminals and staff
smartphones, the equivalent of the well-known “Amber alert” when a child in the
community goes missing. This sort of system can have the advantage of including
a description of the patient and perhaps a photograph. But the disadvantages
are that it may take longer to implement and you lose the chance that a visitor
or other patient might identify the missing patient.
Key assigned staff should immediately go to a designated “command center” from which they will
direct the response. Each unit (clinical and nonclinical) will have a specific
predetermined area they must search in a systematic fashion. The command center
must have an overlay grid of the buildings and surrounding areas and be able to
mark off areas on the grid that have been searched. The search teams must have
keys to their search areas since sometimes patients lock themselves into rooms
inadvertently.
We also recommend early outdoor search since a patient can
easily stray far from the building (or into automobile traffic) in a very short
period of time. The search grid
should include outdoor areas like parking areas, open spaces, bushes and
shrubs, and any other adjacent areas where a patient might wander to. A typical
grid sector might be 500 x 500 feet. Typically, the search teams comb a sector
in a standardized direction (eg. south to north).
We also recommend that the local police department be
notified immediately by the operator when the “missing patient code” is called
(don’t forget to include them in your planning process and drills). Many
facilities also use many security video cameras that are monitored centrally.
Security staff may be able to scan those quickly to look for a patient exiting
the building.
What do you do when you find the patient? First, be aware
they are likely confused and be careful not to frighten them. Do a brief
assessment as to whether they may have been injured. Notify the command center
you have found the patient and either return them to their unit or to the
emergency department. They should be evaluated by a physician at that time to
determine whether any injuries have occurred. In the unfortunate circumstance
where the patient is found dead, the scene should be left undisturbed because
the authorities will treat it as a crime scene.
Staff Education and
Drills
All your staff need to be aware of how to respond when a
patient goes missing. That means not only describing their roles on their
initial orientation but doing an inservice at least
annually. Drills are critical for
any event that is likely to be rare but critical when it occurs. Especially with
relatively rare events, it is important that all staff know what to do during such
emergencies and the best way to prepare for those is with drills. Yes, you can
and should include education and training on missing patient alerts during
orientation and annual reorientation but you have to periodically run a drill
to see whether the responses are adequate and timely. During drills one may
also see various nooks and crannies and other areas (eg.
ventilation ducts) that a patient could get into, perhaps leading to some
physical improvements to prevent such dangerous access.
From a few incidents we’ve seen or read about we had the
suspicion that missing patient events were more likely on weekends and
holidays. A VA study actually showed that was not the case, at least
statistically (DeRosier
2005). But keep in mind your drills need to also be done on shifts
other than a regular day shift and you also need to inservice
staff, such as agency staff, that might be temporarily working at your
facility. And that applies to administrators as well! Your administrators
likely have a role all missing patient responses (eg.
communicating with families, police, etc.) so they must know what their role is
on weekends and nights, too.
Despite our best efforts to identify patients at risk for
wandering and elopement and interventions to prevent such, it is impossible to
prevent all such cases. In the San Diego incident, the patient had been flagged
as being at risk for wandering, was put on a video monitor and was even seen
getting out of bed on the video monitor, yet still managed to wander out of the
hospital anyway. Therefore, it is imperative that every healthcare facility and
organization have in place plans and procedures, known by all and rehearsed
periodically, in the event a patient goes missing. Plans that consider missing
patients the sole responsibility of “Security” are doomed to failure.
It’s always a good time for facilities to say “could that happen here?” and do a
thorough review of your policies and procedures for missing patient incidents,
including making sure you do appropriate drills for such incidents. You
probably will be unable to prevent every potential elopement. When one does
occur, do a debriefing session as
soon as possible to identify potential missed clues and other useful lessons.
Then do a formal RCA (root cause
analysis) within a short timeframe. There are always valuable lessons
learned that hopefully can prevent other elopements in the future. But even if
you’ve not already had a patient go missing it’s good to do a FMEA (failure mode and effects
analysis) to determine your potential vulnerabilities. And make sure you do drills, with thorough critiques and
debriefings following the drills. When you do your FMEA, consider also what
happens to locking doors when a fire alarm goes off. You might even consider
doing your missing patient drill immediately following a fire drill.
Our prior columns on
wandering, elopements, and missing patients (many with links to other good
resources):
- July 28, 2009 “Wandering,
Elopements, and Missing Patients”
- February 9, 2010 “More
on Preventing Inpatient Suicides”
- June 26, 2012 “Using
Patient Photos to Reduce CPOE Errors”
- December 2012 “Just
Went to Have a Smoke…”
- April 2, 2013 “Absconding
from Behavioral Health Services”
- April 30, 2013 “Photographic
Identification to Prevent Errors”
- October 15, 2013 “Missing
Patients”
- December 2013 “Lessons
from the SFGH Missing Patient Incident”
References:
Duffin C. Dementia sufferer found
dead in a ditch yards away from scandal-hit Stoke Mandeville Hospital after
being allowed to wander off her ward. The Daily Mail (UK) 2015; 26 March 2015
Bear J. Man dies of hypothermia complications after
wandering away from Longmont assisted living facility. Times-Call (Longmont,
CO) 3/31/2015
CDPH (California Department of Public Health). Complaint
Intake Number CA00357013. 2014
Regan S, Viana JC, Reyen M, Rigotti NA. Prevalence and Predictors of Smoking by Inpatients
during a Hospital Stay. Arch Intern Med 2012; 172(21): 1670-1674
http://archinte.jamanetwork.com/article.aspx?articleid=1389239
VA. VISN 8 Patient
Safety Center of Inquiry, Tampa. Wandering and Missing Incidents in Persons
with Dementia. Updated: October 24, 2013
http://www.visn8.va.gov/patientsafetycenter/wandering/
Wandering Resources.
http://www.visn8.va.gov/VISN8/PatientSafetyCenter/wandering/WanderingResources.doc
Whitnall C. Ross Memorial hopeful door wrap will
discourage wandering patients
Image reflection of
hallway cabinet installed on doors exiting the Continuing Care Program.
MyKawartha.com March 15, 2015.
DeRosier JM, Taylor L. Analyzing Missing Patient
Events at the VA. TIPS (VA Topics in Patient Safety) 2005; 5(6): 1-5
http://www.patientsafety.va.gov/docs/TIPS/TIPS_NovDec05.pdf
Print “Missing
Patients and Death”
http://www.patientsafetysolutions.com/