Suppose I am looking at an estimate $\hat\beta$ in a clinical trial data.

With $a=.05$, a patient with $\hat\beta>1$ is considered to have some medical condition.

A daily example could be a 24-hr blood pressure monitoring. When a person is diagnosed as pre-hypertension, then a physician usually orders a 24-hr BP monitoring to make a final call. The BP monitor usually goes on on 20-minute basis. Unfortunately, even with the 24-hr tracking, there is a lot of noise in any given patient's data.

Back to my original question: if you are looking at a threshold-value for an estimate and need to clean up the noise to reduce the probability of false-positive, what are some standard protocols? Can you provide both theory and how this is done in STATA, MATLAB, Pyhton, R, and the like? You can pick a package that you are familiar with and provide a reference if possible.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.