How to Appraise a Clinical Trial - Part 2

[@MahdiAldaini]

Im so happy to have found you sir. What an amazing video. I just learned so much. You are positively impacting the life of so many doctors and patients. Thanks you.

[@StrongMed]

Thank you for the kind words!

[@MsP0624]

Your videos are so clear and helpful. Thank you.

[@the1saul1]

This is fantastic (as usual)! Comprehensive and clear. Keep up the good work!

[@zannatul23]

Dear Sir, thank you so much for this video. Very informative and helpful. Really appreciate it.

[@pelesaifostina8002]

Thank you so much God bless you

[@thomask8225]

Thank you very much for the video!!! I have a question regarding the placebo control: How can one appraise the bias given by the side-effects of the treatment which increase the placebo-effect or boost the placebo-effect but which lacks in a placebo control? I think there were critics from Prof Kirsch regarding the RCT's of antidepressants and the lack of an "active placebo" control. Does that mean that the trials were not sufficient controlled or would that be to harsh?

[@pochola4550]

Thank you! What do you think about wait-lists as controls? They are used a lot in psychiatry to provide more interventional data.

[@docsayedi]

thanks Dr. Eric. Are you gonna discuss in detail on how to make sense of the results. P values and confidence intervals in particular?

[@StrongMed]

I'm sorry, but not so much in this particular series. The P value warrants a whole video just for itself. Maybe someday I'll get around to doing a biostats series (or even better, collaborate with someone more knowledgeable in biostats than myself!)

[@wanasyraf5648]

Thank you sir for the knowledge. I have requested access for the printable summary.

[@StrongMed]

This direct link should work: drive.google.com/file/d/0B9SDUwepGWeUSVdJRzE0anpfRHM/view?usp=sharing&resourcekey=0-DqHKrdfOjiP3le4EjPdm4Q

[@shif442]

great video...any chance the that the upcoming videos will conclude the ecg videos?

[@StrongMed]

Not in the immediate future, but they are still coming eventually! Sorry for the wait.

[@MrJukamada]

This is good!

[@dhashanisivaratnam2126]

AWESOME !! Thank you.

[@folumb]

Hi Dr. Strong, is there an example of a poor clinical study that you might point to for a real life application of your method?

[@StrongMed]

I'd actually offer a challenge to the Strong Medicine community to come up with examples of clinical trials that don't have any flaws in either their prior assumptions, methodology, or interpretation of results. (If they are out there, I'm not sure I've read one!) In all seriousness though, this old thread on Reddit talks about a few problematic trials (ROCKET-AF, NINDS, SPRINT, RE-LY, PARADIGM-HF, etc...): www.reddit.com/r/medicine/comments/52cm08/what_are_some_landmark_clinical_trials_that/ Also, the excellent book, Ending Medical Reversal has some great discussions re: flaws in a few major trials.

[@lizbethmcewan]

Thank you.

[@ahmedtarek3782]

thank you so much for this wonderful video,, very clear and straightforward and informative how can i get in touch with you doctor ?

[@markolucijanic1179]

Dear Dr. Strong, what do You think of trials with "multiple primary endpoints" or co-primary endpoints like when both OS and PFS are considered co-primary endpoints in study protocol and in later publication by study authors. If statistical significance is reached for only one of those, should trial be considered positive or negative?

[@StrongMed]

Good question. As long as the co-primary endpoints were both defined as such ahead of time, and the study is adequately powered for both, then that's fine. However, progression-free survival is a poor surrogate endpoint. If a trial looked at both OS and PFS as co-primary endpoints, I would pay very little attention to PFS and focus just on OS.

[@markolucijanic1179]

Thank You for the answer. I agree with Your reasoning, especially regarding OS and PFS. This way authors (sponsors) definitely increase their chance of having a positive trial.

[@ahmetrmznakur]

Strong Medicine Maybe this question will be too hypothetical but I think has some point. Do you think this co-primary endpoints effectiveness still aplly when there is more endpoints than the minumum number that by chance one of them will look significant. For example roughly if there is more than 20 primary endpoints one of them will look significant by luck in the case of significant p value is less than 0.05. Thank you for your exellent videos, you are great.