@VPrasadMDMPH's Threads
I see a lot of people interested in discussing what will it take for real world observational #RWE data to justify drug approvals.
Let me give you some clues.... [THREAD]
Let me give you some clues.... [THREAD]
1. You will need to prove that real world obs. registries can predict the results of FUTURE or ONGOING trials, and not trials where you ALREADY KNOW THE RESULT
Anyone with basic SPSS or STATA or R skillz can get an obs data set to recapitulate an RCT HR, when you know what to aim for. So it must be studies for which you DO NOT KNOW the result.
It will also have to be many such studies (>100 ish)
It will also have to be many such studies (>100 ish)
Lots of people in oncology want to do less and less RCTs to verify the clinical benefit of cancer drugs because they are more and more impressed with surrogates and biomarkers
Here is how that works... [THREAD]
Here is how that works... [THREAD]
They do less and less RCTs. And keep saying, "the surrogate is improved sooooooo much that we don't need an RCT"
And so we get less and less RCTs.
We will never haven an RCT for some notable FDA approvals because a surrogate in a single arm study used for REGULAR drug approval (aka no confirmatory efficacy trial)
drive.google.com/file/d/1hER5hK…
We will never haven an RCT for some notable FDA approvals because a surrogate in a single arm study used for REGULAR drug approval (aka no confirmatory efficacy trial)
drive.google.com/file/d/1hER5hK…
Under-use occurs in medicine when you show that a person who met inclusion criteria for a study with indisputable and positive benefit DID NOT undergo a procedure AND that was contrary to their wishes/ desires
Over-use occurs in medicine when you show someone underwent a intervention that has been found to offer no net benefit in well done studies OR care not consistent with their wishes OR care that defies common sense, e.g. (jamanetwork.com/journals/jama/…)
The major cause of under-use is probably lack of access to health care, and the major causes of over-use are likely diverse ($$$, psychology, hubris)
Notified 2day our 2012 paper on stress testing after r/o heart attack (w/ trops & EKG) (2012)
drive.google.com/file/d/1umjA12… was cited in 👇
Our paper argued we need RCTs to know which pts ought to undergo such testing
Then I read this. 🤦♂️ It is so grievously flawed. [thread]
drive.google.com/file/d/1umjA12… was cited in 👇
Our paper argued we need RCTs to know which pts ought to undergo such testing
Then I read this. 🤦♂️ It is so grievously flawed. [thread]
Because it is an NBER paper the audience is happy to praise it, might even be published well, but I fear many will not understand why it is not useful.
As I get time, I will tweet a long explanation here
As I get time, I will tweet a long explanation here
First, as background, you need to know that people who present with pain from above the waist to below the jaw (and particularly if in the chest area) in the ER may get an EKG and troponin to "rule out" a "heart attack"
There are so many things wrong with this article.
Sadly it is just more HYPE that misleads the public about the issues
Sadly it is just more HYPE that misleads the public about the issues
This paragraph, where do I start....
1. Insurers pay for many low value drugs (FDA approved OR recommended by expert bodies for off label use).
The off-label evidence is often poor (as we show belwo)
So it is NOT TRUE insurers only pay for FDA approved drugs
*Wrong
bmj.com/content/360/bm…
The off-label evidence is often poor (as we show belwo)
So it is NOT TRUE insurers only pay for FDA approved drugs
*Wrong
bmj.com/content/360/bm…
Recently attended in the hospital, and I left with this thought for trainees
[Thread]
[Thread]
It is easy in the inpatient setting to get caught up in the day to day;
the adjustment of medications, electrolytes
the coordination of tests, procedures and consults
the adjustment of medications, electrolytes
the coordination of tests, procedures and consults
And, don't get me wrong, that is important stuff
Vital stuff
You have to be diligent in thinking about those things
We can't afford mistakes
Vital stuff
You have to be diligent in thinking about those things
We can't afford mistakes
So @oncology_bg points out that some folk don’t understand or care for our paired papers in Jama Im last month.
Let me summarize our findings and let you decide [thread]
Let me summarize our findings and let you decide [thread]
@oncology_bg The paper by Bishal shows that when the FDA gives accelerated approval to a cancer drug— based on a surrogate. They seldom ever verify benefit on survival or quality of life
What is worse is that ~20% of the time they convert to full approval based on a trial using the SAME Exact surrogate
It's always difficult to articulate a systematic way to read and interpret data, but here is my best attempt for cancer screening.
Twitter makes me think this review is necessary (Born from frustration)
[Thread]
Twitter makes me think this review is necessary (Born from frustration)
[Thread]
1 Efficacy
The first question that faces in you in a cancer screening is does the intervention have efficacy, does it improve outcomes healthy people (and remember that is generally the target group) care about
The first question that faces in you in a cancer screening is does the intervention have efficacy, does it improve outcomes healthy people (and remember that is generally the target group) care about
1A The endpoint that matters to healthy people is living longer. For any reason, period. Or living better. Unfortunately, it is v. v. difficult to study living better in these trials. Practically, Living longer it is.
Living longer means reduction in all cause death
Living longer means reduction in all cause death
Discussing/ debating cancer screening is frustrating
Here is why
[Thread]
Here is why
[Thread]
1.Motivated reasoning
Entire specialities of medicine are justified by the revenue from screening and downstream interventions hence massive motivated reasoning to believe
Entire specialities of medicine are justified by the revenue from screening and downstream interventions hence massive motivated reasoning to believe
2. A contest of who really cares about the patients
Frequent comments that one person CARES more about human suffering than another person. (when both parties care, but it is about the data/ facts & what works or doesn't)
Frequent comments that one person CARES more about human suffering than another person. (when both parties care, but it is about the data/ facts & what works or doesn't)
Delighted to announce that I will soon tender my resignation as boring associate professor (yawn, no options!) for a new AMAZING NEW pursuit
I'm HUMBLED to be CMO of MonytorPin, a new wearable startup company!
I'm HUMBLED to be CMO of MonytorPin, a new wearable startup company!
MonytorPin has similarities and differences to all other companies in this space
Here's what sets us apart:
We have a lapel pin wearable that tracks, light/dark cycles, steps, distance, GPS, elevation, and even exhaled breath...
Here's what sets us apart:
We have a lapel pin wearable that tracks, light/dark cycles, steps, distance, GPS, elevation, and even exhaled breath...
That is right, we measure everything important in exhaled breath, and that is a lot of important stuff. Like what?
Lol, you don't know!
Lol, you don't know!
It is OUT!
After 2 years and 7000 person hours, we are pleased to debut the largest set of reversals to date, 396
H/T to Herrera-Perez, Haslam, Crain, Gill
[TWEETORIAL] to come....
elifesciences.org/articles/45183
After 2 years and 7000 person hours, we are pleased to debut the largest set of reversals to date, 396
H/T to Herrera-Perez, Haslam, Crain, Gill
[TWEETORIAL] to come....
elifesciences.org/articles/45183
Out of about 3,000 studies published in the top medical journals (2003-2017), almost 400 of the practices tested were established practices that were no better than existing standards of care
We call these…medical reversals
We call these…medical reversals
This time, unlike prior work, we performed a search for systematic reviews on the topic to BOLSTER the claim
This took a LOT of work, and credit must go to
Herrera-Perez, Haslam, Crain and Gill
This took a LOT of work, and credit must go to
Herrera-Perez, Haslam, Crain and Gill
Paired articles out now in JAMAIM by @bishalgyawali @akesselheim
And @IamEmchee and I with 2 scathing editorials.
Check out this gem from @RichardLehman1
[TWEETORIAL coming here]
jamanetwork.com/journals/jamai…
And @IamEmchee and I with 2 scathing editorials.
Check out this gem from @RichardLehman1
[TWEETORIAL coming here]
jamanetwork.com/journals/jamai…
Let me start with this visual.
A comparison of the complete response rate of GLEEVEC (the 1 parachute in oncology) and all drugs in oncology of which many say we cannot do RCTs because they are "like Gleevec"
Spoiler: they are not. Avg CR rate 6%
jamanetwork.com/journals/jamai…
A comparison of the complete response rate of GLEEVEC (the 1 parachute in oncology) and all drugs in oncology of which many say we cannot do RCTs because they are "like Gleevec"
Spoiler: they are not. Avg CR rate 6%
jamanetwork.com/journals/jamai…
Let me say that again, among all drugs approved based on RR, the median CR rate is 6%
median Sample size > 100
So, you can't do randomized trials because these drugs are like parachutes?
median Sample size > 100
So, you can't do randomized trials because these drugs are like parachutes?
Why I did not attend < ACADEMIC EVENT >
Especially if it occurred on a weekend or evening
[A minor rant of my honest thoughts]
Especially if it occurred on a weekend or evening
[A minor rant of my honest thoughts]
I often get asked "Are you going to AACR?
ASCO GI symposium?
JJJ foundation?
NNN forum?
CC conference?
alt: "You should have come to YY fair"
And, this is not #humblebrag edition b/c it isn't just about things you get invited to & it is not unique to me
ASCO GI symposium?
JJJ foundation?
NNN forum?
CC conference?
alt: "You should have come to YY fair"
And, this is not #humblebrag edition b/c it isn't just about things you get invited to & it is not unique to me
Even without formal invitations, I suspect the average academic could travel more than an NBA basketball player in midseason just by stringing together events that one can sign up for
And, there are even more events & things one may be invited to
And, there are even more events & things one may be invited to
JUST OUT!
My colleagues & I offer a Modest Proposal to aid the most zealot proponents of cancer screening!
You know the wise folks who think overdiagnosis and false positives are overblown, and there is no harm greater than missing a localized ca.
Care for a TWEETORIAL?
My colleagues & I offer a Modest Proposal to aid the most zealot proponents of cancer screening!
You know the wise folks who think overdiagnosis and false positives are overblown, and there is no harm greater than missing a localized ca.
Care for a TWEETORIAL?
In recent years, the USPSTF has BACKED DOWN in cancer screening.
Annual -> biannual & 40--> 50 for mammography
PSA gets tossed around, yes, no, only after a discussion; not if over 70
But, backing down from screening makes some very mad.
(no, it has nothing to do with $$$!!!)
Annual -> biannual & 40--> 50 for mammography
PSA gets tossed around, yes, no, only after a discussion; not if over 70
But, backing down from screening makes some very mad.
(no, it has nothing to do with $$$!!!)
They argue that in balancing the benefits and harms of screening; there is no harm greater than the potential to MISS a curable cancer
They actually say things like this... (real quotes)
They actually say things like this... (real quotes)
These Duty Hour trials are a great to teach many many flaws in non-inferiority trial design.
From a scientific standpoint, they are quite poor.
They resemble the seeding trials used by the industry to bolster preconceived notions
A few points
From a scientific standpoint, they are quite poor.
They resemble the seeding trials used by the industry to bolster preconceived notions
A few points
1. Lousy endpoint
All trials must be judged by the primary endpoint. In this case it is a 1% NI margin at 30 days.
All trials must be judged by the primary endpoint. In this case it is a 1% NI margin at 30 days.
Medicare mortality at 30 days is roughly 12%
Accepting up to a 1% increase in mortality, so residents can work longer shifts is total madness.
Accepting up to a 1% increase in mortality, so residents can work longer shifts is total madness.
How Doctors Think [Thread]
You have an intervention X that people use for condition Y
X rose to prominence based on a little bit of bio-plausibility & some anecdotes
& that is about it
You have an intervention X that people use for condition Y
X rose to prominence based on a little bit of bio-plausibility & some anecdotes
& that is about it
After many years (and only if fortunate), a group of brave investigators test X in condition Y in a RCT
It is totally negative
Nothing improved at all
Some say, woah, lets rethink X
It is totally negative
Nothing improved at all
Some say, woah, lets rethink X
But many start saying, "yeah but"
Yeah but, they didn't do the RCT right
Yeah but, it still has a role with appropriate patient selection
Yeah but, it should still be used for pts with severe Y
Yeah but, the trial did not enroll the patients with most severe Y
Yeah but, they didn't do the RCT right
Yeah but, it still has a role with appropriate patient selection
Yeah but, it should still be used for pts with severe Y
Yeah but, the trial did not enroll the patients with most severe Y
Going to slowly tweet out 15 of the most interesting things I read, listened to and saw in 2018. With oncology/ health policy/ evidence themes
Starting with Same Dataset, Different Interpretations by @BrianNosek nature.com/news/crowdsour…
Starting with Same Dataset, Different Interpretations by @BrianNosek nature.com/news/crowdsour…
The point in this paper is if you give researchers the (a) exact same data set and ask the (b) exact same question you get a variety of answers.
Some point out that *some* of the point estimates are consistent....
Some point out that *some* of the point estimates are consistent....
Of course that is true, but other teams are reporting qualitatively different conclusions: some say no evidence of bias, others say bias
And, imagine a second what happens if...
And, imagine a second what happens if...
Cute & funny paper in the @bmj_latest
But unfortunately it furthers a narrative that is detrimental to medicine and scientific progress. Many may not see why.
Allow me to explain.. [THREAD]
But unfortunately it furthers a narrative that is detrimental to medicine and scientific progress. Many may not see why.
Allow me to explain.. [THREAD]
There are many people in medicine who think that the intervention they use must work. Some of these are invasive and costly interventions that have modest to weak bioplausibility, but no real trial evidence
The history of EBM shows that many of these interventions fail upon rigorous testing. So if you are a proponent of an unproven intervention, how can you avoid the call to test it rigorously?
This series by @charlesornstein and @katiethomas about financial conflicts in medicine is devastating; Should win the @PulitzerPrizes
Some thoughts
[THREAD]
nytimes.com/2018/12/08/hea…
Some thoughts
[THREAD]
nytimes.com/2018/12/08/hea…
In addition to outright failure to disclose, the profession treats disclosure as a token gesture
Aaron Boothby former @OHSUSOM student found 38% of conflict slides at ASCO flash faster than humans can read!
drive.google.com/file/d/1rBh7z7…
Aaron Boothby former @OHSUSOM student found 38% of conflict slides at ASCO flash faster than humans can read!
drive.google.com/file/d/1rBh7z7…
@StellaYooMD and @awadahmedMD find virtually identical findings in a different specialty
Why? Because docs don't give a #^$-- its a token disclosure
ascopubs.org/doi/abs/10.120…
Why? Because docs don't give a #^$-- its a token disclosure
ascopubs.org/doi/abs/10.120…
This is unbelievable
In JAMA, Ed Benz replies to @jsross119 paper with many incorrect statements
Shall we review?
jamanetwork.com/journals/jama/…
In JAMA, Ed Benz replies to @jsross119 paper with many incorrect statements
Shall we review?
jamanetwork.com/journals/jama/…
1. cancer has several established surrogate markers?
Uhhh, actually, most correlations btw surrogate & survival are lousy, especially in metastatic disease which has most approvals
drive.google.com/file/d/1FFkD2k…
Uhhh, actually, most correlations btw surrogate & survival are lousy, especially in metastatic disease which has most approvals
drive.google.com/file/d/1FFkD2k…
This is even worse for US FDA cancer drug approvals where most surrogates used are based on poor correlations or other low levels of evidence, aka not established
*true even for regular approval in violation of stated standards*
*true even for regular approval in violation of stated standards*
These two are the worst offenders for churning out garbage nutritional epi papers for years & building empires of BS, and no surprise they do not really understand the criticism, why the answer may not b RCT, but that so much of what this work is no better than nothing at all
They cherry pick the trans-fat story, and omit the thousands of other papers they put out that are utter garbage
Where is the full throated defense of nut consumption and prostate cancer
Yes, I appreciate the gesture by FDA, but in this brief thread I will explain why it is pretty much useless.
THREAD
THREAD
So... basically... We will use PFS or ORR in all these cancers for some traditional and some accelerated approval!
DEVASTATING Letter to the Editor in @BloodJournal for a paper published in @NEJM which (Wrongly) claims brentuximab AVD is superior to ABVD.
Care for a TWEETORIAL?
FYI I have a commentary under review that trashes this bad, bad trial but makes totally different points!
Care for a TWEETORIAL?
FYI I have a commentary under review that trashes this bad, bad trial but makes totally different points!
This is an AWFUL trial used to expand market share of a UBER costly drug Brentuximab and not actually answer the question of whether patients are better off..... In other words, par for the course in 2018
Lets take a look at key objections of the letter
First this....
POW!
Lets take a look at key objections of the letter
First this....
POW!
Then this...
BAM!
BAM!
My feed has blown up over the last day with
Persistent Reservations against Reforming the First 2 Years of Medical School Curriculum aka bad arguments
I find it funny because 6 years ago, I thought I had heard it all and wrote this
drive.google.com/file/d/1mLepDy…
[Well, Here's 2 cents]
Persistent Reservations against Reforming the First 2 Years of Medical School Curriculum aka bad arguments
I find it funny because 6 years ago, I thought I had heard it all and wrote this
drive.google.com/file/d/1mLepDy…
[Well, Here's 2 cents]
Let's be very clear about this topic: there are lots of stupid & inefficient parts to medical training, (especially scut work & bad work hours, rotating w disinterested attgs, bad teachers, etc.) I can't tackle it all !
I want to talk about the content of yr1-2 of med school
I want to talk about the content of yr1-2 of med school
Most schools spend 2 years learning a litany of basic science factoids so students can take a stupid test called USMLE step 1, and use that score to decide who goes into a coveted residency (plastic surgery and derm) and who goes into a less coveted one (family practice, peds)
