Probability: Flashcards
What is probability?
- how likely something is.
- usually on a scale of 0-100% or 0-1.
- it follows that you need to act before you are certain.
- consider probability for better decision-making in the context of uncertainty.
is it possible to rationally update one’s views to reach complete certainty (0-100%)?
- no but you can approach this.
What were the results of the Morgan et.al 2021 paper?
(Accuracy of practitioner estimates of the probability of diagnosis before and after testing).
- respondents overestimated the probability of diagnosis before and after testing.
- this overestimation is consistent with cognitive biases, including base-rate neglect, anchoring bias and confirmation bias.
what are the 2 implications of these results from morgan et.al?
- overestimated probability used in deciding whether to initiate therapy can result in medication overuse and excessive procedures with their associated harms.
- these errors would corrupt shared decision-making with the patients because practitioners need to understand the likelihood of various outcomes in order to communicate them.
what is the problem with how decision-making is taught?
- decision making is often focused on communication skills and not on actually understanding the probability of disease.
what was the worrying observation found by Attia et al (2004)?
there were a number of clinicians indicating pre-test probabilities of 100%.
What does this finding reflect from Attia et al?
- This presumably reflects a cautious attitude, assuming that all patients have the disease until proven otherwise.
When would indicating pre-test probabilities work?
- this would only work if the tests ordered have powerful negative likelihood ratios and if these tests indeed does give negative results.
What does the overestimation of disease risk do?
- leaves clinicians unable to judge false positive test results, and may result in more intervention than necessary, indicating a lack of appreciation for how diagnostic tests influence the probability of disease.
What are biases?
- Not a personal shortcoming but a human trait.
- when we substitute judgements of representativeness for judgements of actual probability.
what is representative?
-the degree to which something is representative of, or similar to, the stereotype.
Outline the steve bias example, is he more likely to be a farmer or a librarian?
- based on the description alone it sounds as if he is a librarian (judgment of representativeness).
- however, we have to take into account other factors such as when this was where this was etc. farmers had a bigger role in society during this time than librarians.
- is there a library in his city, does he read?
Why is being aware of bias important?
- because an uncommon representation of a common disease is more likely than a common representation of a rare disease.
- if this weren’t so then diagnosis would be a matter of pattern matching.
- when you hear hoofbeats think of horses not zebras.
how can the essential keys to disciplined bayesian reasoning be simply summarized?
- anchor your judgement of an outcome on a plausible base rate (pretest probability).
- question the diagnosticity of your evidence.
What is a pretest probability?
- it is the best estimate of a disease probability before you do a test.
- it is a starting place from which to update probability.
what do you have after a test?
- post test probability.
how do we determine pretest probability?
- there are multiple ways to do this but you want to start with a good reference class.
what is the best reference class?
- the set of patients that most closely matches this patient.
What is a basic reference class?
- the prevalence of disease in a population.
what is prevalence?
- the proportion of a population affected by a condition.
what are the pros of using prevalence?
- relatively easy to search for.
- can specify sub-populations to get a more accurate estimate.
what are the cons to using prevalence?
- maybe an underestimate if it is something that people frequently seek medical attention for.
- less helpful for acute medical conditions.
What is a more specific reference class?
- studies that give an eventual diagnosis in patients presenting with complaints similar to your patients.
what are the pros of using a study?
- takes the symptoms into account to provide a more accurate initial judgement.
- takes into account that people tens to seek medical attention for some conditions more than others.
What are the cons of using a study?
- this research is less common to find (harder to find).
- clinical scenarios in research may be different from your own.
What reference classes do we not use?
- incidence in the population.
- lifetime prevalence.
what is the incidence in the population?
- the frequency of disease over a period of time.
what is lifetime prevalence?
- the chances of developing the disease over a lifetime. will tend to be an overestimate.
what is a consideration when using reference classes?
- try to find different reference classes with the useful methods above to represent maximum and minimum estimates.
- considering pre-test probabilities to the illness scripts (epidemiology).
what can explicitly considering probability help with?
- better decision-making in the context of uncertainty.
what can biases such as representativeness in place of probability distort?
judgements of subjective probability.
