On diagnosing rare diseases

In this post, I want to dive into the interesting relationship between conditional and marginal probability within the Bayesian framework and how it can trick us into misleading conclusions. Let’s assume a patient presents with some symptoms S in an emergency department. The symptoms are highly specific for a very rare disease D.

The statistical hypothesis testing framework

The relationship between power, sample size, effect size and type I error alpha A second very important concept is the relationship of certain probabilities within the framework of statistical hypothesis testing. There are several important concepts in statistics that are closely related when it comes to hypothesis testing. \(\alpha\) is the level of error that is accepted for erroneously rejecting the null hypothesis (although it is true) e rejected.

Why most research findings are false

In his famous paper “Why most published research findings are false” Dr. John Ioannidis proofs that most research findings are false. Given that his paper is quite technical, I will go over his rationale here again. Concepts 1. R: the prior probability The first concept that is introduced is R which is considered to be the relationship between “true relationships” and “no relationships” in all tests in a certain field before any test or study has been undertaken.