clinical decision making Flashcards
what is an error *
the failure of a defined action to be completed as intended (error of execution)
or use of a wrong plan to achieve an aim (error of planning)
examples of medical errors (*
incorrect diagnosis
failure to employ indicated tests
error in performance of an operation, procedure or test
error in dose or method of using drug
how much death in US is caused by medical error
3rd biggest cause of death
what is the scale of medical errors in uk
medication errors have acsaused 12000 deaths
may contribute to 0.75-1.5 billion in additional health care expediture
explain the case of WJ - medical error *
he has given chemo every 3 months
vincristine IV and cytosine IT (intrathecally - via spine)
vincristine was accidently injected IT = paralysis - switch off life support
eg showing problem with dr nurse relationships *
problem is we follow morfe senior collegues than the correct procedure
med in cupboard said max dose 10mg
dr phoned (actor) and said draw up 20mg
21/22 nurses gave the high dose - trend is to obey dr
what are causes for medical errors *
a lot have cognitive involvement
extent od diagnositic errors
they account for:
the largest fraction of claims
the most severe pt harm
the highest total of penalty payouts - $38.8 billion between 1986 and 2010
what is the process of clinical decision making *
rare to use formal computations
instead clinical judgement is a combination of heuristics and probabilities
heuristics are rules of thumb, educated guesses and mental shortcuts
good in urgent situations - quick, heuristics are based in pt data and prev patients make decisions subconsciously by pattern recognition
what are the 2 systems for decision making*
hot system - system 1
cold system - system 2
describe system 1 for making decisions *
emotional
'’go’’
simple
reflex
fast
develops early
accenuated by stress
stimulus control
without reflection
describe system 2 for decision making *
cognitive
'’know’’
complex
reflextive
slow
develops late
attenuated by stress
self control
builduing info together
which system for decision making is better *&
need both
describe context in decision making *
dependant on the situation - if urgent action needed system 2 is not going to be effective - need system 1
eg of when we ahve to override system 1 with system 2 *
when we have seen illusions before - system 1 tells us that 1 line is shorter than the other
system 2 tells us from experience that they are both the same length
describe Nisbett and Wilson’s study *
consumer study in shopping mall
4 pairs of tights in a row and consumer had to decide which they liked the best
consumers were significantly more likely to pick far R pair - even though all identical and positioon of tights was chosen and changed randomly
teh consumers could provide justification, and when it was suggested that their choice was becasue the tights were on the R they opposed this
system 1 controlled the choice (R hand dominance) system 2 was blissfully unaware - thinking it was in charge - 1 made the choice and 2 makes rationale after event
describe conformatory bias *
the tendancy to seek interpret and recall info in a way taht confirms one’s preexisting beliefs or hypothesis, often leading to errors
eg looking for evidence that supports your choice to remain in brexit
how does conformatory bias relate to clinical practice *
possibly to blame for ineffective medical procedures that were used for centuries before the arrival of modern medicine - leeching and blood letting - only found evidence that supported this
when evaluating a diagnosis have to test for alternatives - not just things that would confirm your hypothesis
describe the overconfidence in medicine *
podbregar and collegues studied 126 patients who died in ICU- did autopsy so could determine the diagnosis
physians were asked to provide the clinical diagnosis and their level of uncertainty on it
clinicians who were completely certain of the diagnosis ante-mortem were wrong 40% of teh time
what can explain overconfidence in medicine I*
confirmatory bias
why is it understandable that physicians have overconfidence
hard to manage the demand of making decisions, so if you were not confident, you would be anxious all the time and not function
describe the sunk cost fallacy *
future outcome depends on what we have done so far
sunk costs are any costs that have been spent on a project taht are irretrievable eg money spent building a house and drugs used to treat pt with rare disease
thought process is that you cant give up treating the pt now with the expensive drugs - however they really need a new treatment
how should we make decisions with respect to the sunk cost fallacy &*
rationally - only factor affecting future action should be future cost-benefits ratio
but humans dont always act rationally - the more we have invested in the past, the more we are willing to invest in the future
study illustrating whether sunk cost fallacy exists in decisions *
bornstein et al
evaluation of clinical decisions were not influenced by sunk costs
however they did express a sunk cost effect in non-clinical decisions
describe anchoring *
heuristic
cognitive bias
initial starting point determines deviation from starting point eg in sales
individuals are poor at adjusting estimates from a given starting point
adjustments are crude and imprecise
`
clinical effect of anchoring *
initail idea for diagnosis serves as the anchoring point eg acute pancreatitis
but if later symptoms/history reveal no alcohol intake and tests show normal level of panc enzymes
clinicians may dismiss or excuse conflicting data - eg the pt is lying/underestimating alcohol, his panc is burned out and lab has made a mistake)
how is probability involved clinically *
many clinical situations involve making decisions on the basis of prob
eg 2 competing diagnoses, alternative treatments
more complicated than give credit for
describe representiveness heuristic *
subjective probability that a stim belongs to a particular class based on how ‘typical’ of that class it appears to be - regardless of base rate probability
can result in neglect of relevant base rates and otehr errors
describe how representiveness heuristic can cause errors*
eg a 60yr old woman has no known med problems, now looks and feels well, reports erlier symptoms of feeling sick, sweaty, clammy, SOB, feeling faint
this doesnt match typical MI which is characterised by chest pain
but unwise to dismiss MI becasue it is common among women of that age and has highly variable presentations - women present differnetly to men
how do you assess conditional probabilities *
a women presents with a lump in her breast - from age and previous records you estimate risk iof cancer is low - 1%
you send her to radiologist for mammogram and radiologist says mammogram is +ve - indicating cancer
if 1000 people testing - using baseline risk of cancer - 990 dont have cancer and 10 do - there are 80% true positives and 10 percent false positives
therfore 99 false positives and 8 true positives using the baseline chance of 1%
therefore only have 8/107 = 7.5% chance of having cancer
explain how the framing effect can influence clinical decison making *
we have an aversion to loss - so if something is loss framed we are less likely to make that choice
eg 200/600 will survive on a treatment or 1/3 probability 600 will live and 2/3 prob none will be saved
- likely to choose treatment that saves 200
400/600 will die with treatment or 1/3 prob no people will die but 2/3 prob all 600 will die
- more likely to choose 2nd because we dont want to choose option that involves loss of 400 lives
how is framing affected by age *
when presented with treatment options described in postive, negative or neutral terms - older adults are more likely to agree with treatment when positively described
effect stronger in older adults
describe teh availability heuristic *
probabilities are estimated on the basis of how easily and/or vividly they can be called to mind
people heavily weigh judgements towards more recent info
individuals overestimate freq occurance of catastrophic events as cause of death compared to strokes etc
how can the availability heuristic affect clinical decisions *
a clinician who recently missed teh diagnosis of an PE in young women with chest discomfort but no other symptoms or risk factors might overestimate PE in similar cases - ordering more CT angiographies - cause further errors
what are the things that can be done to improve clinical decision making *
education and training
feedback
accountability
generating alternatives
consultation
understand and apply statistical principles eg law of large numbers- greater confidence can be
describe how ed and training improves decisions *
integrate teaching about cignitive error and diagnostic error into med school curricular
recognise that heuristics and biases might be affecting judgement even though we might not be conscious of them
reflect on our decisions and what has lead to them introduces a buffer for these errors
how does feedback improve decisions *
increase number of autopsies - learn from previous decisions - brain bank for dementia where there is a lot of diagnostic uncertainty
conduct regular and systematic audits - see if patterns at certain times/clinics
follow up pts - se if you were right
how would accountability improve decision making*
establish accountability and follow up for decisions made
give more weight
how does generating alternatives improve decision making *
establish forced considertion of alternative possibilities - eg the generation and wroking through a ddx and looking for evidence for different possibilities
encourage routinely asking the qn - what else might this be
how would consultation imprve decision making*
seek second opinions
use algorithms - for probability calculations
use of clinical decision making support systems - AI - fuse large amounts of data
how do algorithms help decision making *
an algorithm is a procedure that if follwed exacltly will provide the most likely ans based in evidence
the rules of probability are eg of algorithms
most useful when problem is well defined - excludes many everyday decisions
people have to be taught how to use them
gets complicated when people have inconsistant symptoms and co-morbidities
describe clinical decisiom making support systems*
include a lot of info at any time
wont replace clinical decision making - art of medicine
but will be better at calculating probability
gambler’s fallacy
a logical fallacy involving the mistaken belief that past events will affect future events when dealing with random activities, such as many gambling games.
