Subarachnoid hemorrhage - a Bayesian approach to a neurological black swan

Published on March 29, 2018   35 min

Other Talks in the Category: Diseases, Disorders, & Treatments

0:00
Hello. Today's talk is on "Subarachnoid Hemorrhage: A Bayesian Approach to a Neurological Black Swan". I'm Dr John Bedolla. I'm an Assistant Professor and Assistant Director of Research at the University of Texas Dell Medical School. I'm also the National Director of Risk for a national emergency medicine company called US Acute Care Solutions. If you have any questions, please contact me at jbedollamd@austin.utexas.edu.
0:28
The goals of my talk today are the following: explain Bayes theorem and apply it in medical reasoning, apply Bayes theorem to estimate the probability of a spontaneous subarachnoid hemorrhage in patients presenting with acute headache, and use Bayesian inference to inform risk benefit analysis of testing for subarachnoid hemorrhage and to decide who needs a lumbar puncture to rule out the subarachnoid hemorrhage when the CT is negative.
0:53
Let's start off with some definitions. The definition of Bayes theorem, which is named after Thomas Bayes (1701-1761) is, the probability of an event when factors that are positively or negatively associated with that event are known. A Black Swan, after Nassim Taleb, is a rare catastrophic unpredictable event that in retrospect looks very predictable. Large scale examples of this are 9/11, the stock market crash of 1987. Medical Black Swans are rare catastrophic diseases that present atypically. I'll be using these terms, so I want to define it now. Likelihood ratio positive, so this is a statistical term that can be used in Bayes theorem analysis, and what it means is that given a particular factor is positive, how much more likely is it that a condition you're trying to diagnose or exclude is present? Likelihood ratio negative, given a particular factor is negative, how much less likely is it that the condition you're trying to diagnose or exclude is present? Pre-test probability is something similar to prevalence. It's the probability that a patient has a condition you're trying to diagnose or exclude before you do any testing. This is similar to the concept of prevalence. Post-test probability, this is the probability that the patient has a condition you are trying to diagnose or exclude after you get a test result. So, pre-test is before you do the testing, post-test is after you do the testing.
Hide

Subarachnoid hemorrhage - a Bayesian approach to a neurological black swan

Embed in course/own notes