Transcript Why We Make Mistakes - University of Kentucky

Why We Make Mistakes
Jeffrey Young, MD
Senior Associate Chief Medical and Quality Officer
Professor of Surgery
Director, UVa Trauma Center
University of Virginia Health System
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Understanding Clinical Care
• First step in understanding error
• How do we carry out diagnosis and treatment?
• Where are the opportunities for safety?
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Effective Care Taxonomy
Effective
Care
Cognitive
skill
U N I V E R S I T Y
Technical
skill
O F
V I R G I N I A
Teamwork
H E A L T H
S Y S T E M
Cognitive skill
• Knowing how to accurately assess the state of a
patient
– What data and actions are necessary to get an accurate
assessment
• Knowing the significance of the data reflecting the
patient’s current state
• Matching the patient’s state to the correct schema
• Mentally testing and activating an acceptable action
plan
• Follow-up
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Background
• Medical education
– Even problem based curriculums are not truly
tactically oriented
• Issue – data- analysis- action- reassessment
– Differential diagnosis
• Look at data and create diagnosis list based on
characteristics of conditions and their relation to the
data you currently have available
• Then look at list and decrease number of diagnoses until
you are left with one
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Tactical Assessment
• We are not trained to rapidly assess a situation,
look at the key elements, discard less
important inputs, create an action plan,
implement it , and see if it worked
• Military does this very well, as do other high
risk jobs (aviation, etc.)
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Current Process
• Do a thorough history and physical examination
–
–
–
–
–
–
–
–
Chief complaint
HPI
ROS
Medications
Allergies
Physical Exam
Assessment
Plan
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Assessment/Plan
• Signs and symptoms lead to potential
diagnoses
– Diagnoses lead to investigation
– Through investigation diagnoses are supported or
discarded
– You are left with your most likely diagnosis
– Then you initiate a treatment plan
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Paradigms
• Definitive treatment delayed until all
possibilities are entertained and evaluated
• There is concern that aggressive early
intervention can lead to overtreatment,
incorrect therapies, or complications of
medications and procedures
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Bottom Line
• Some patients have time for thoughtful
evaluation
• Some do not
• First cause of error: not making the above
characterization of the patient correctly
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Technical Skill
• Knowing the indications for a test or procedure
• Knowing how to safely carry out the action, and all
possible adverse events related to that procedure
• Insuring backup and help if needed
• Monitoring the patient during the procedure
• Correctly interpreting the results of the procedure
• Following up to insure absence of adverse event
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Teamwork
• Can not carry out technical act, nor action plan
without other competent individuals or processes
• Need to know how to engage the care team
– Advice members of goal, path to goal, alternative paths if
they arise,
– Framework for communication
– Checklists and Time outs
• CRM principles during action plan
• Follow-up
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Example
• Intern is on call and at 1AM nurse calls:
– “ Mr. Smith who had a stent placed for a AAA
today is having some belly pain and his temp
spiked to 38.7C”
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Opportunity of Error
• Intern
– “Well…how bad does his belly hurt?”
– “He had his procedure today, I’m sure he just has
some atelectasis, make sure he uses his incentive
spirometer”
– “Thanks for the call”
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Error
• Have we given that intern the tools to ask the
right questions and do the right things?
• Have we prepared him to fail in this case or
succeed?
• Would a better understanding of how people
make errors improve his chances of success?
• Can we learn from other high risk industries?
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Experience and Error
• Novices
– Novices (or people inexperienced in that domain)
do not yet have the experience and knowledge to
make correct decisions
• They don’t know what can go wrong
• Don’t know the cues
• Don’t even know what data to look for
– Thus they must depend on an overall philosophy or
mindset toward events to guide their decisions
– If that mindset is dangerous, bad care will result
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Experience and Error
• Experts
– When you are an expert, you’ve gained the
experience and knowledge to properly make
decisions
– You know what data is essential and what is
extraneous
– You can look for cues that put you down the
correct path
– Can we teach this? Or do you have to live it?
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Types of Errors
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Types of Errors
• Slips
– Failures to properly adjust tasks that require little
conscious attention to the characteristics of a new
situation
• Without thinking, ordering an adult dose of a med for a
child
– Correction
• Make it difficult to do the wrong thing
• Error requires more steps and positive affirmation
– “Do you really really want to format C:?”
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Types of Error
• Lapses
– Failures of memory that cause tasks not to be done
– Common in task overload or distraction
– How to avoid?
• Train in high intensity situations
• Clear pre-arranged plans that require little creative
thought and may not be perfect, but are SAFE
• Reminders
• Same corrective actions as for slips
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Errors
• Mistakes
– The selection of incorrect actions by
misclassifying a situation or failing to take into
account all relevant factors in a decision
• Evaluating for nausea and vomiting but not taking into
account new onset DM as a cause
• Attributing cold symptoms to the URI and not realizing
degradation in cardiac function is the cause
– Perfect execution of incorrect plans distinguishes
this from lapses and slips
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Error Generators
• Assumptions
– It must be OK or someone would say something
• Generalizations
– Didn’t happen last time, why should it happen this time?
(translation; I got away with it last time….)
– Successful folly is folly nonetheless (Jim Hurst, MD)
• Pushing a bad system to the limits
– Without working on making the system better
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Error Generators
• Laziness (not that common)
• And number 1(‘s)!
– Too little FEAR that things can go wrong
• “Fear does the work of reason” – Winston Churchill
• quickly and with very little warning.
• Making CERTAIN that the conditions are stable or
improving before moving on
– Too little FEAR that you don’t know everything
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Common Causes of Clinical Error
• Incorrect triage of problem
– Problem more serious than most people realize
• Insufficient fear of being wrong
• Practitioner has not seen enough clinical situations to
know all possibilities
– Buggy knowledge – they fill in their knowledge gaps with
generalizations
• No follow-up
– Almost all major disasters can be averted by simply going
back and reassessing the patient
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Error Recovery
• Lack of cognitive understanding of condition
or state by all practitioners (Most common
cause of adverse events I see)
– Team does not have enough experience or know
enough to realize what is going on with patient
• Patient in ectopic units
• Specialists not available
• Patient at low capability facility (don’t understand who
is at risk)
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Error Prevention
• Very complex issue
• Systems, education, decision making,
communication
• New knowledge, new techniques and
procedures
• Information systems
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Fear of Being Wrong
70
60
50
40
30
20
10
0
< 4 Weeks ICU
>4 Weeks ICU
Pessimistic
(percent of
subjects)
U N I V E R S I T Y
O F
Algorithmic
V I R G I N I A
H E A L T H
S Y S T E M
What Can We Learn From Studies of
How Experts think?
• Schema (what is the mindset you are using
with this patient?)
– You actually decide what “kind” of patient you
have very rapidly (almost instantly), but you may
not realize it
• Visualize care plan
– In mind, can I see this patient going home the way
they look now?
– Does the patient look like the typical patient who
is admitted for this type of problem?
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Very few major disasters result
from a single error
“Tactical catastrophes are rarely the outcome of a
single poor decision. Small compromises
incrementally close off options until a commander is
forced into actions he would never choose freely” –
Nate Fick
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Situations Where Error May be Unrecoverable
Tenerife: March 27, 1977
Worst Aircraft Accident in History
Factors:
Inadequate technical skills of ATC
Suspect plan (?)
Experienced pilot (top airline pilot at KLM)
Fog
No ground control radar
CRM
Stepped on transmissions
No warning system for active runway
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Diagram
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Errors
• Cognitive
– Decision to take off
– Decision to place both planes on runway
– Decision to refuel
• Technical
– No ground radar
– Communication equipment inadequate
• Teamwork
– Cockpit and control tower teams
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Tenerife
• This accident (more than any other) changed the
culture of air safety
– To be honest, we have not yet had such an incident in
medicine (IOM report?)
• When people realized even the most experienced pilot
could make egregious error, conclusion reached that
human performance must be enhanced with safety
measures
• Sometimes we just do real stupid things
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Experience and Deliberate Practice
– “Good
judgment comes from
experience, and experience comes from
bad judgment “
– “Luck is not method, and neither is
hope. Hard work is.”
– Can we produce safe doctors with
decreasing clinical experience?
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Decreasing Errors
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Role of “Philosophy” of your team
and facility
• What is your “Philosophy of care”?
–
–
–
–
Get as many patients seen as possible?
Increase patient volumes by 10%?
Insure every patient gets recommended care?
Think of the worst thing that could be going on
with the patient and rule it out?
– Save money?
– Avoid unnecessary radiation and testing?
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Philosophy
• Some philosophies may be at odds
– Increase volume vs. provide safe care
– If care is barely being safely provided at present volume,
how can you expect higher volume will allow safe care
without system changes?
• Trauma service functions with safeguards and double
checks for a ICU census of 8—10 and a floor census
of 10-12
– What happens when ICU census increases to 20 and floor
census increases to 25??
– Do you have a contingency plan that goes into effect (like
the military, police and fire do)
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
“Try Harder”
• Just telling people to “try harder” or “make
less mistakes” or “take better care of the
patients” rarely is an effective strategy
• Most people are trying pretty hard and their
output is more dependant on the system they
are working in, not their effort
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Decision Making
• Can teaching medical practitioners traditional
clinical reasoning be detrimental?
• Naturalistic Decision Making
– Gary Klein
– Under conditions of uncertainly, time pressure, and high
risk (medicine), experts do not use analytical methods.
– They use fast and “sufficient” strategies
• In other words they don’t search for the “best” answer, just the first
“acceptable” answer
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
How do experts make decisions?
• Look at patient and data
• Fit that data into a schema they have seen before
• Choose a plan based on their previous experience
– This is why inexperience is devastating
• War game the plan and its execution in their head
(think about it, you really do this)
• If plan simulates OK, proceed
– If it doesn’t step back and form another option
• Repeat as necessary
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Centrality of Diagnosis
• Diagnosis has been considered medicine’s central
task, but is this best?
• Having a solid diagnosis can make treatment easier,
but the lack of a diagnosis does not relieve the
necessity to act
• Thus the central task of medicine may be
management, not diagnosis
• We should not say, “what is the diagnosis?” but
“what should we do now?” (Beth Crandall)
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Approach
• Rapid assessment of patients initial presenting
data (clinical and digital)
• Rapid intervention of life threatening signs and
symptoms
– Nothing life threatening
• Narrow to known condition
– Mentally simulate treatment and evaluation
» Proceed with plan
» Follow results
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Situational Tactics
• We don’t teach this well
– How to rapidly evaluate
• Find most important data points
• Assess in relation to other inputs
– Assign priority to actions
– Initiate actions
– Reassess and revise
• You must practice this by running through scenarios
over and over again, or seeing patients with similar
problems over and over again
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Common Emergency Problems
•
•
•
•
•
•
Mental status change
Injury
Septic conditions
Cardiovascular problems
Respiratory problems
When we look at avoidable death, almost
every case fits in one of these categories
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Common Threads in Safely Treating
All These Problems
• ABC’s and Call Help
• Protect from further injury or deterioration
• Rapid exam and assessment of current state and
contributing factors
• Form plan
– Mental war gaming
• Initiate
– May be harder than it sounds
• Follow-up
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Performance by Sessions
0.8
0.7
Percent Completed
0.6
0.5
No Sessions
1-2 Sessions
>2 Sessions
Expert
0.4
0.3
0.2
0.1
0
Initial Evaluation
U N I V E R S I T Y
Secondary Evaluation
O F
Diagnosis
Follow-up
V I R G I N I A
Total
H E A L T H
S Y S T E M
Errors in Trauma and Surgical
Care
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Busy Nights (or Busy Units/Services)
• Tests and double checks fall through the cracks
– 80 hour does not help this
• Triage attention to severely ill
– Leaving less sick to fend for themselves
• Corrective action
– Practice
– Have defined algorithms that you stick to
• Don’t allow people to improvise just because they are busy
– Its actually the worst time for them to improvise!
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Acidosis
• Can be insidious
• Need to screen out occult hypoperfusion and you
need to jump on persistently acidotic patients quickly
– In elderly, persistent acidosis = non-survivor
– In severe head injury, persistent acidosis = skilled nursing
facility
• “The labs must be wrong, ignore them”
• Sometimes people don’t want to face the fact the
patients is heading in the wrong direction
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Inconsistent Neurosurgical Care
• A major problem
• There are few aspects of care everyone agrees with
(despite AANS guidelines)
• Often the most junior attending (or resident) is
saddled with trauma
• Little synergy between Trauma and Neurosurgical
services (often at odds)
• Can we correct?
– Try to get areas of agreement and slowly increase their
scope
– Do the same thing every time
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Elderly Trauma
• Elderly patients with minimal mechanisms
who do not communicate well
– Triage problem
– Treatment problem
•
•
•
•
High risk of respiratory failure
Difficult to get pain free
Interaction with current meds
Underlying disease
• Intervention – focus on the elderly
– Especially those with head injuries
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Conclusions
• To avoid error you should:
– Expect problems
– Think the patient is sicker than they look
– Define evaluation and treatment algorithms
• And make sure your people understand the reasons behind your
algorithms
– If they think its dumb, they wont follow it.
– “Thinking hurts the team”
• In many many situations, this is true
• Inexperienced people improvising often hurts the patient far more
than it helps
– Practice, Practice, Practice
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Some Solutions That Wont Work
• Don’t develop a “wizard” system
– Where the patient are only treated correctly when
the “wizard” is around.
• If your system doesn’t work with the most
inexperienced, dimmest person at the bedside,
then you don’t have a system
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Conclusions
• We need to learn how to handle urgent and emergent
situations from vocations that deal with this all the
time
– Cockpit communication and checklists
– Rapid military response
• Planning and adaptation under intense pressure
• Doing some things every single time, no matter what
– Fire ground
•
•
•
•
Cross training of personnel
Backup
Chain of command
Bringing in more experienced people
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
Systems
• Can not always depend on people to make the
right decision
• Need to have systems that can rapidly
recognize error and intervene
• Need to expect that things will go wrong
(Hope is not a method)
• Need to explain why we have safety processes
– If people don’t understand them, then their mind
doesn’t engage
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M
U N I V E R S I T Y
O F
V I R G I N I A
H E A L T H
S Y S T E M