Likelihood Ratios Explained

Enroll in our online course: bit.ly/PTMSK DOWNLOAD OUR APP:
📱 iPhone/iPad: goo.gl/eUuF7w
🤖 Android: goo.gl/3NKzJX GET OUR ASSESSMENT BOOK ▶︎▶︎ bit.ly/GETPT ◀︎◀︎
This is not medical advice. The content is intended as educational content for health care professionals and students. If you are a patient, seek care of a health care professional. Kai explains in depth how to calculate likelihood rations using equations and the nomogram method. Pre-Test and Post-Test probabilities are also covered.
Support us on Patreon: / physiotutors
Visit our Website: bit.ly/web_PT
Like us on Facebook: bit.ly/like_PT
Follow on Instagram: bit.ly/IG_PT
Follow on Twitter: bit.ly/Tweet_PT
Snapchat: bit.ly/Snap_PT

[@fatimaa6172]

thank you very much i will include you in my prayer every day

[@Physiotutors]

Very much appreciated Fatima!

[@62-88]

Thanks a million for your hard work. This clip save my life.

[@gunesmelikekilic745]

Exam tomorrow and you just saved the day!

[@Physiotutors]

Good luck Melike!

[@ArneBroedel]

Thank you for the explanation. I think using the nomogram to determin the post-test-probability is applicable in practice. But the accuracy of the result is totally depend on the therapists "guess" of the pre-test-probability. So what would you suggest to improve the accuracy of estimating the pre-test-probability? I'm pretty sure, that the "inter-tester-reliabilty" of the estimated pre-test-probabilty is really weak for most cases. How to train that skill? Can that estimation be evidence-based?

[@Physiotutors]

Hi Arne,
the challenge in clinical practice is that you don't have much literature on prevalence rates and on the validity of several items of your patient history.
So before a test you can only estimate how high the pre-test probability of a certain disease/pathology is.
Don't even think that a nomogram is applicable.
We find that the likelihood gives us the best estimate of how good a test is to confirm or exclude a certain hypothesis and after using the nomogram or calculating a couple of scenarios you have a pretty good feel of how much your post-test chance is influenced.
At the end of the day we will all have to learn to live with clinical uncertainty - we just have a better idea about how great that uncertainty is.

[@stephenpuryear]

@@Physiotutors Wonderful answer that also applies to other areas of investigation as well as general measurement practices. Thanks for the extra insight!

[@philipk.5951]

The buff angel is on point as expected ;)

[@adilmanzoor7703]

Very informative, and concise. Thank you.

[@Physiotutors]

Cool, happy you like it Adil and thanks for subscribing.

[@9527david]

Really good explanation, thanks!

[@AdotGplusone]

you guys are awesome! really clear and reliable videos, keep it up!

[@Physiotutors]

Thanks a lot Ali, we certainly will!

[@internalmedicine9982]

Thanks

[@ivancarlson953]

Thomas Bayes has entered the chat.

[@wj9165]

is there any information how the basic parameters are provide to calculate likehood ratio?

[@aliazimdaudpota1117]

Excellent !

[@Physiotutors]

thanks!

[@lakesidemission7172]

you are a born teacher 🐣

[@indiannafranke2731]

love you guys!

[@Physiotutors]

Thanks 🙏🏼 appreciate the love

[@XerxesBreakSama]

how did you set the pretest probabilty to 60%. is there any rule?

[@mrd8300]

Very good, thanks for posting

[@Physiotutors]

Happy you like it David! No worries and thx for following us!

[@GauravGiri.]

Video is very nicely presented. Can you guide on how to estimate pre test probability?

[@Physiotutors]

You have to look into the literature how prevalent certain pathologies are in your country and setting. That's a good starting point.
Then signs & symptoms from your anamnesis will make a hypothesis less or more likely but it remains guesswork.

[@GauravGiri.]

@@Physiotutors thank you.

[@sharineskaria1171]

Thanks for making this topic so easy to understand great 👍

[@Physiotutors]

Ur welcome sharine

[@Global_Info]

In this lecture he says LR+ should be greater than 10 and LR- should be less than 0.1. But in practice the maximum values of both Sn and Sp is between 0 and 1 which further implies that LR+ cannot go beyond 10 and LR- cannot go less than 0.1. Kindly clarify this if someone can. Thank you.

[@Physiotutors]

Yes, the first part is correct. LR+ can go to unlimited and is good above 10, LR- can go towards 0 and is good below 0.1.
Sn and Sp can be a percentage (0-100%) or a decimal (0.00 - 1.0).
LR+ = Sn / (1-Sp)
LR- = (1-Sn) /Sp
Just fill in different examples (like shown in the video) and you will see how LR+ can exceed 10 and LR- can be lower than 0.1

[@Global_Info]

Thank you so much. It works with your example.

[@dr.luis.espinosa]

LRs of 2 to 5 and 0.5 to 0.2 generate small (but sometimes important) changes in probability. EBM Manual JAMA evidence 2 Ed

[@nathananwaboawinwum8135]

What is the name for the formula of the PPV and NPV?

[@plexusamsterdam5392]

positive predictive value / negative

[@jifarjemal2235]

how can one sipmply say from history 60% of pretest ????

[@21stgirly]

What is post-test probability?

[@Physiotutors]

The probability (%) of the pathology/disease being present after your test. It either goes up with a positive test or down with a negative test. The magnitude of the pre to post difference depends on the tests properties (e.g the likelihood ratio).

[@freetothunder9011]

i dont understand shit

[@chanphalee3042]

ich wollte nach LR-Test für die lineare Kombination von Parametern suchen...wieso bin ich hier gelandet....

[@chanphalee3042]

ok... hab mir sogar zum Ende angeschaut...

[@sisiw3074]

Gwapo Mo

[@Kerryedrin]

he is hot