Transcript Morning Keynote Address: Biases & Mistakes in Epilepsy Care

Biases & Mistakes
in Epilepsy Care
Orrin Devinsky, M.D.
NYU Langone Epilepsy Center
Biases in Epilepsy Care: Lessons
of Behavioral Economics
 Diagnostic
Bias
 Prospect Theory
 Law of Small Numbers
 Status Quo Bias
 Availability Heuristic
Biases in Medicine: Kahneman &
Tversky’s Lessons
 Loss
aversion
 Anchoring
 Framing
 What You See is All There Is
What do NBA coaches, mothers
and doctors have in common?
 The
Diagnostic Bias
 1st
round v. 2nd round choice
 Diagnosis to doctor = child to mother
 Reliance
on prior diagnosis
 Failure to consider other disorders
Convulsive syncope
Nonepileptic psychogenic seizures
 Failure to consider diagnostic changes
Prospect Theory
 Decisions
about alternatives with risk
where final outcome risks are known,
people decide on potential values of
losses or gain
 Risk averse (insurance policy)
 Risk acceptance (lottery ticket)
Prospect Theory:
Epilepsy Care
 Felbamate is too dangerous
 Risk of death is <1/10,000
is too dangerous – <1/1500
 Refractory epilepsy is ok, it is what we are
used to
 Surgery



Yearly risk of MVA in 1/8000
Yearly risk of SUDEP in patients with refractory
epilepsy: >1/500
Yearly risk of other epilepsy related mortality in patients
with refractory epilepsy: >1/500
Availability Heuristic
 If
you can think of it, it must be important
 Mental Shortcut: ease of example coming
to mind = value to make judgment about
probability of event
 News of danger – people worry about rare
causes of illness or death that receive
media attention (9/11 and air travel)
 Letter K – first letter or third letter in
average English word? (2x difference)
Availability Heuristic:
Epilepsy Care
 Valproic
acid (Depakote) is a common
cause of liver disease
 Lamotrigine (Lamictal) is a common cause
of life-threatening rash
 Patients & families are driven by prior
experience
10 drugs don’t work, ?11th
 Bad: someone on the web told me…
 Good:
QOL & Availability Heuristic:
A Different View
 QOL
- Defined by patient not MD
 Should patient’s perspective be filtered
through “objective medical lens”? - NO
 QOL is about listening, changing
perspective, and using the patients’
view as ultimate measure of outcome
QOL: Clinical Relevance
 QOL
issues most relevant to chronic
disorders, problems beyond disease
symptoms
 Hypertenstion – b-blockers v. ACE inhibitors
(Experts wrong!)
 Epilepsy is a paradigm of a QOL disorder:
seizures are infrequent, AED effects,
comorbid disorders (depression, migraine) &
psychosocial problems are often chronic
Law of Small Numbers
 Hasty
generalizations from a few
examples
 Initial set of data is usually biased
 Scientists understand power and statistics
in their discipline, but often forget it when
they think outside their discipline
Humans are Anecdote Driven
 We
evolved to understand individual
instances very well, not statistics
 A moving story about a castaway dog or
sick children v. a genocide of ~800k
 Would
you give more for a dog or 100 sick
kids?
 Rwanda v. OJ Simpson – media coverage
 Vaccines
cause autism (NO!)
Humans are Anecdote Driven
 Sabril
(vigabatrin) can cause blindness
 Felbatol (felbamate) can be deadly
 People can become vegetables after
spinal taps
 You only need to hear about one bad
case…and it doesn’t have to be true
 Need to examine the evidence
Failure to Understand
Numbers
 The
medical literature is very confusing,
even for scientists and doctors
 Few doctors and fewer patients have
formal statistical training
 The Monte Hall problem
 AED/blood count/liver tests and Cancer
Screening – America makes political not
wise choices
Status Quo Bias
 Doctors
and patients fall victim
 Doctors accept previous diagnoses
 Doctors advocate treatments that are
‘accepted’ but not ‘proven’
 Patients accept poorly controlled seizures
and/or side effects
 Patients accept ‘communal experience’
although unproven/anedotal
We get used to what
we get used to
 What
do these all have in common?
 Lottery
winners
 Quadriplegics
 Farmers whose roosters rape chickens
 People who eat mediocre blueberries
 Parents of kids with Lennox-Gastaut
Syndrome
Loss Aversion
 People
prefer to avoid losses more than
they seek equal gains
 Roughly two-fold
 Endowment effect: people value
something they own than something of
identical value
 Duke
tickets
Loss Aversion: Epilepsy Care
 Seizure
control is the loss
 The existing drug regimen is safety – the devil
you know
 Gamble: seizure freedom/stable level of
incomplete control v. greater alertness,
memory, mood, bone health?
Loss Aversion: Epilepsy Care
 How
fearful are you of a side effect in a new
drug versus an existing one?
 Doctors like to add medicines more than they
like to take them away
 The gabapentin story
PB
 30
yo woman, refractory complex partial
seizures
 Any side effects? No!
 Converted from phenobarbital to
carbamazepine (Tegretol, Carbatrol)
 Boss observed dramatic improvement in
mood, memory and mental processing speed
and ‘intelligence’
Errors in Assessing Risk




Surgery is too dangerous
 Living with chronic epilepsy can be dangerous
Changing medications is too risky
 Change can be risky; No change can be risky
 The grass is browner on the other side
 Breakthrough seizure
 Living with chronic side effects has risks
We accept the negatives we think we know but fear
the change to make them better
Do no harm, but judiciously assess risk
Anchoring
 Over-reliance
on a specific piece of
information
 Our decisions are tied to arbitrary anchors
 Dan Ariely – write down last 2 digits of
your SS#; now lets auction wine or
chocolate
 Attentional anchor – who is happier?
Californians or mid-Westerners?
Anchoring
 Patients
and doctors often allow one piece
of information to dominate their decision
on a topic that is complex
 What
we heard last about a drug or treatment
 Nickname – Dopamax
 Single side effect – weight gain (Valproic acid)
Frames & Framing
– scheme of interpretation using
stereotypes, anecdotes and accepted
‘norms’ that people use to understand
and respond
 Framing – how information is packaged
dramatically influences how we respond
to it. Presenting the same data in different
frames leads to very different
interpretations.
 Frames
Frames & Framing:
Epilepsy Care
 Many
patient see memory problems as
primarily due to medications when they
are often an effect of epilepsy
 Framing – 80% of children on
levetiracetam (Keppra) have no significant
behavioral problems v. 20% of children on
levetiracetam have significant behavioral
problems
Failure to Understand Framing




“Surgery is 99.95% safe” is very different than
“Someone died from surgery” or “1 in 1500 die”.
 Substitute benign brain tumor for epilepsy surgery
Mentally invert presentations to better understand
pros and cons
Patients must trust their doctors, but they must also
assess their doctor’s bias and their own
The neurosurgeon, the radiation oncologist & the
neuro-oncologist
What You See is All There Is
(WYSIATS)
People make decisions based on limited
data by using available information and
ignoring information that is not available
In Epilepsy: assume we understand causes
of seizures when we may only have 1020% of the data
Missing Mood Disorders
 All
epilepsy patients at increased risk
 Patients must tell; doctors must ask –
both often fail
 Refractory epilepsy
Greater contributor to impaired
Quality of Life than seizures
Depression in up to 50%
Suicidal ideation - 20% in past 6 mos
Majority are untreated
Two Great Lies in
Epilepsy
Seizures don’t hurt the brain
They cause structural and functional
impairment that can progress over time
Seizures are never fatal
SUDEP
Sudden Unexplained Death
in Epilepsy (SUDEP)
General population (2–3)
Epilepsy incidence population (5)
Epilepsy prevalence population (7)
Patients in clinical trials (30–50)
Patients undergoing vagus nerve stimulation (41)
Patients referred to epilepsy centers (50–60)
Surgical candidates (90)
Surgical failures (150)
Missing The Big Picture


Focus on person, not diagnosis
 Listen, beyond the words to feelings
 See their world: situations influence health
 Look patient in the eyes
 Speak with family and friends
Therapies are limited by medical box
 Therapists - cognitive, psychological, etc
 Pragmatic approaches (sometimes key!)
 Compliance
 Sleep hygiene
 Memory lists
The Dangers of Expert
Consensus





MRI offers no real advantage over CT in epilepsy
diagnosis - 1986
Ketogenic diet is not effective - 1990
Felbatol (felbamate) is extremely safe – 1993
Experts convince themselves, other doctors and
patients
Demand evidence or humility
Failure to Reassess
 Disorders
change and evolve
 New situational factors arise
 Need to keep a fresh perspective
 Need to cast a broad differential
diagnosis and consider a broad
therapeutic strategy
 What was is an excellent but sometimes
dead-wrong indicator of what is
Doctors and Patients
Move in Packs
 Doctors
are influenced by peers, thought
leaders, marketing – they are as susceptible
to status quo, texts (eg, JME, absence)
framing as are patients
 Doctors in different medical centers, cities,
and regions have different practices
 Patients strongly influence each other –
support groups, internet, etc
Failure to be Humble
 Most
people don’t enjoy admitting that
they don’t know something
 Doctors are expected to have answers,
to have therapies, and if they are
honest, people go to other doctors or
alternative therapists – catch 22
 Tell a white lie or admit ignorance?
Common Errors in Therapy
 Wrong
diagnosis
 Wrong medication selection
 Failure to use medications systematically
 Start
low, go slow
 Consider time of doses v. seizure & side effects

Benign Rolandic Epielspy
 Consider

strategies to reduce side effects
For dizziness – oxcarbazepine (Trileptal) after solid
breakfast, not empty stomach
 Failure
to document changes carefully
 Nonadherence (noncompliance)
Fatigue: Diagnosis and
Causation
 Premature
exhaustion in mental or physical
activities, weariness, lack of energy
 Common in epilepsy patients
 AEDs
 Other
drugs (eg, psychiatric drugs)
 Seizures

Epilepsy wave activity
 Depression
 Sleep
disorders
Final Thoughts
should make sense – separate
emotional/gut and rational/reflective
 Understand what you do and why
 Be an active partner in care
 Be skeptical
 Be positive, think healthy
 Things