Thread by @dr_anvith

Importance of learning Probability for Doctors

This simple yet elegant mathematical theorem will help us with the problem of Over diagnosing and Over treating the Patients.

Here is an example, which seems counter-intuitive...

FALSE POSITIVE PARADOX (1/5)

Even a highly accurate test is useless if the disease is rare.

If 10 people among 2 Crore have disease & test identifies with 99% accuracy then here is a shocking fact...

99.995% of +ves will be FALSE !

(2/5)

This is why Every Doctor, actually every educated person should understand Bayesian Probability and how to apply it.

If you have a test that is 99% accurate for a condition that occurs in only 1% of people, then any positive result has a 50% chance of being false

(3/5)

Bayesian probability: P(A|B) = P(B|A)*P(A)/P(B)

Event A is a person having the condition, event B is positive result, then the probability of your subject having the condition given a positive result is

P(A|B) = 0.99*0.01/(0.01*0.99+0.99*0.01)=0.5

(4/5)

Knowing that P(B|A) is 0.99 & P(B) is the sum of the probability of a someone having the condition & getting a positive result (0.01*0.99) plus the probability of them not having the condition and getting a false positive result (0.99*0.01)

(5/5)