"For those planning to perform an observational study, before you begin work, consider why you are doing the study."
Most important suggestion for me. All medical research should be hypothesis-driven. The hypothesis should be structured to be provable and/or falsifiable with the proposed research. Otherwise not worth doing. Retrospective data mining does not usually lead to useful or reproducible results.
I would ask IRBs to reject studies of new fancy drugs Vs placebo when a perfectly good alternative already exists. It's not ethical to withhold the older drug from patients.
Sensible Medicine and its authors (including the returnee’s books/content) are the only reason I feel adept at reading articles and understanding exactly what you are getting at here. Thank you.
This is great. I've been saying this for years but no one is listening. The newest kid on the block is the "non inferiority trial." Like they actually prove anything? Our trainees have not been well served by their medical schools because the schools no longer teach how to think. It's so teach-to-the-test that today's physicians just parrot what BigPharma gives them.
Non inferiority trials seem to be just another way to confuse readers and sell a drug or medical product. I pay no attention to "modern" statistical measures. Critically analyze the validity and relative hardness or softness of each endpoint. Beware of composite endpoints---they are often driven primarily by one of the elements and often are the softest ones. Look for or calculate the percentage incidence of each endpoint (absolute risk reduction) and use common sense to determine whether the difference is of any practical significance. If the ARR figures are not provided, the authors are engaging in deception. Be suspicious of authors that trumpet relative risk reduction figures. The exclusive or primary use of RRRs suggests that the authors are engaging in advocacy rather than science. This kind of basic analysis can be used by anyone of average intelligence.
Wonderful discussion - as always! In addition to your observations, I find that much of what I read are “systematic reviews” of a given topic or procedure. Rarely if ever do these reviews produce a meaningful conclusion or recommendations other than further studies are needed to determine the exact cause etc… Really!!!!
Pet peeve of mine. Outdated citations. I see this weekly, and if infuriates me. An author makes a point, almost a slam dunk, with two superscripts. You click, and they are 15+ years old. I clicked your links relating to how closely RCT and observational trials align (80%): 2001 and 2012. Yes, its Ioannidis, but what i am thinking is with computing power and logarithmic jump in published papers over the last few decades, maybe its better. Or likely, much worse. Most of the crummy papers and viewpoints out should be called out by peer reviewers. They get away with overstatements no person would accept if they read between the lines.
Forget the trials - I go to a doctor and ask him what works the best for your patients , how many patients have you treated , IMHO the science has to catch up with the doctors working in the trenches
Good suggestions.
"For those planning to perform an observational study, before you begin work, consider why you are doing the study."
Most important suggestion for me. All medical research should be hypothesis-driven. The hypothesis should be structured to be provable and/or falsifiable with the proposed research. Otherwise not worth doing. Retrospective data mining does not usually lead to useful or reproducible results.
I would ask IRBs to reject studies of new fancy drugs Vs placebo when a perfectly good alternative already exists. It's not ethical to withhold the older drug from patients.
Sensible Medicine and its authors (including the returnee’s books/content) are the only reason I feel adept at reading articles and understanding exactly what you are getting at here. Thank you.
Dr. Cifu,
This is great. I've been saying this for years but no one is listening. The newest kid on the block is the "non inferiority trial." Like they actually prove anything? Our trainees have not been well served by their medical schools because the schools no longer teach how to think. It's so teach-to-the-test that today's physicians just parrot what BigPharma gives them.
Non inferiority trials seem to be just another way to confuse readers and sell a drug or medical product. I pay no attention to "modern" statistical measures. Critically analyze the validity and relative hardness or softness of each endpoint. Beware of composite endpoints---they are often driven primarily by one of the elements and often are the softest ones. Look for or calculate the percentage incidence of each endpoint (absolute risk reduction) and use common sense to determine whether the difference is of any practical significance. If the ARR figures are not provided, the authors are engaging in deception. Be suspicious of authors that trumpet relative risk reduction figures. The exclusive or primary use of RRRs suggests that the authors are engaging in advocacy rather than science. This kind of basic analysis can be used by anyone of average intelligence.
Strong work Adam! 💪 I’ll spread the gospel.
🙏
But here is what actually happens today. In my email inbox this morning and sent to thousands of providers all over the world:
Medscape Logo
Top Five For You
TOP FIVE STORIES CURATED FOR YOU
Alcohol’s Health Myth Is Over. How to Tell Patients
Micro- and Nanoplastics: A Multiorgan Health Threat
DSM Future: Does Validity Still Matter?
The Massive Clinical Payoff of Full Lung Screening
Vitamin K2 Supplementation May Curb Coronary Artery Calcification in Symptomatic Patients
How much RCT support do you think there is for the four implied recommendations from Medscape today?
Depressing.
Hopefully you didn’t click on any of them😆😆
But here is what actually happens today. In my email inbox this morning and sent to thousands of providers all over the world:
How much RCT support do you think there is for the four implied recommendations from Medscape today?
Wonderful discussion - as always! In addition to your observations, I find that much of what I read are “systematic reviews” of a given topic or procedure. Rarely if ever do these reviews produce a meaningful conclusion or recommendations other than further studies are needed to determine the exact cause etc… Really!!!!
It’s garbage in garbage out-
I so value your contributions!
Happy 4th to you and your family
Thank you RH.
Adam
A systematic review's purpose is to turn 10 bad studies into 1 good one. Good luck with that.
Pet peeve of mine. Outdated citations. I see this weekly, and if infuriates me. An author makes a point, almost a slam dunk, with two superscripts. You click, and they are 15+ years old. I clicked your links relating to how closely RCT and observational trials align (80%): 2001 and 2012. Yes, its Ioannidis, but what i am thinking is with computing power and logarithmic jump in published papers over the last few decades, maybe its better. Or likely, much worse. Most of the crummy papers and viewpoints out should be called out by peer reviewers. They get away with overstatements no person would accept if they read between the lines.
Forget the trials - I go to a doctor and ask him what works the best for your patients , how many patients have you treated , IMHO the science has to catch up with the doctors working in the trenches
I love the idea of providing alternative explanations.