What Are Likelihood Ratios and How Are They Used

Video describing the role of likelihood ratios in diagnostic testing

[@marcinw4660]

pure gold - you are the godfather of applied medical statistics

[@ln06502]

I am a medical student and your videos are my go to when I want to brush up or learn about biostatistics and the like. They're so simple and easy to follow. They all make sense. Thank you Dr. Shaneyfelt.

[@balthus9265]

The best video about likelihood ratio I’ve ever seen. Short and straight to the point.

[@augustlyon]

This is the first time I actually manage to understand this material. Thank you!

[@xeebot]

Thank you! I've read so many websites - this is the first time I clearly understood this.

[@alvi78]

My Godd!! I only graduated 24 years back and finally I get it.

[@ciarayoung8750]

Thanks Terry - very helpful video, explained clearly and simply.

[@subhashchandradaga1063]

Simplified and well-presented.

[@taladiv3415]

Thank you professor for explaining this to us.

[@luzlabajos3024]

Great material! Thank you so much

[@humma91]

Subscribed due to this video. Brilliant and easy.

[@RobertoMagnoV]

Thanks a lot from Brazil!!!!!

[@deannawennberg3395]

these videos are very helpful!

[@MohamedKandilMD]

Thank you, this is really helpful

[@tm1972]

Unbelievably helpful.

[@user-wr9ks3tf4n]

fantastic !!!

[@pujasinha3421]

Good one..!!

[@rajeshkhanna5161]

clear.useful

[@anwarhussain5962]

very goog , good job

[@nataliadejesus972]

Thanks proffessor

[@anacosic173]

Thank you so much

[@merleverwohl634]

Thanks for the explanation. Is It possible that the Negative Likeli Hood Ratio is - 1? Or does It have to be >/= 0?

[@zuwaina88]

Thank you. So if we are asked to plot out the post test, we need to dcide first what is our aim: role in disease (hence, we should use the positive likelyhood ration) or role put ( negative test)? Can"t thank you enough.

[@salam7905]

thank you so much^-^

[@hokiehenry]

Very nice explanation. In your example of infection I was wondering about using C reactive protein to rule out infection because the CI includes 1. Shouldn't I try to avoid tests that include 1 in the CI?

[@lavincius]

Yes

[@andresm.acevedo9590]

Where can you find the LRs from the most representative tests used in the clinical setting?

[@briant6164]

For the negative likelihood ratios all the confidence intervals cross one, so none of them should be used to rule out infection

[@RobvanMechelen]

seems a correct conclusion in my view. Thanks.

[@drayubkhan]

If as an example the negative likelihood ratio is between - 0.5 to - 1.5 then it does not cross 1.so not necessarily

[@user-us3zv1ux5f]

Hello there, is the likelihood ratio posttest probability equivalent to positive predictive value?

[@UABEBMcourse]

Yes assuming the test was positive and you used the positive likelihood ratio

[@user-us3zv1ux5f]

Terry Shaneyfelt Thank you for the response!

[@eligottlieb]

So you're saying likelihood+ = P(E | H) / P(E | not H), and likelihood- = P(not E | H) / P(not E | not H)? So it's just a Bayesian update.