Latest / Cultural Wars | America Out Loud News / Pharmaceutical bribery and its impact on patient safety

Transcript

Pharmaceutical bribery and its impact on patient safety · Cultural Wars | America Out Loud News

0:06
Welcome to the unto Take a deep breath.
0:11
Take the higher roles.
0:13
It's what they always say.
0:16
They won't take it from me.
0:20
Don't have a dump yourself with slabing just
0:22
a dream.
0:24
You made your own so kid can scream.
0:27
The people who laugh with a never ending
0:29
force.
0:31
You never have the chance.
0:33
So watch it with this calm my friend,
0:36
cause this is well.
0:44
It's time and this is why.
0:54
Hello and welcome to this episode of the
0:55
Nurses Report here on the America Out Loud
0:57
Network.
0:59
I'm today's host, Nurse David Dwayne, joined by
1:02
nurses Nicole and Ashley.
1:04
Nicole, I'm gonna tap you in for a
1:05
quick introduction.
1:07
Yeah, thank you so much.
1:08
I'm glad to be here again.
1:09
Our usual Friday QA crew.
1:11
For those that don't know, I go by
1:12
Nurse Nicole.
1:13
I've been a registered nurse in the acute
1:16
care setting for roughly 20 years, predominantly taking
1:18
care of cardiac patients, all things related to
1:20
heart.
1:22
Yeah, currently work in a large chest pain
1:25
program and also do a lot on a
1:27
local and state level when it comes to
1:29
Florida politics about bodily sovereignty, the right to
1:31
choose, and medical freedom.
1:34
Awesome.
1:35
Ashley, who are you?
1:36
Uh, just some rando that popped in today,
1:38
I guess.
1:39
No, I'm just kidding.
1:40
Much like Nurse Nicole, I have been a
1:43
nurse for 20 years, started out in cardiac,
1:45
left the hospital, transitioned to functional medicine, do
1:48
a load of advocacy work on the state
1:50
level in regards to the right to refuse
1:53
vaccination.
1:56
And our primaries are upon us the next
1:58
time that we all chat.
2:00
We'll have had some major elections, and I'm
2:03
out there just trying to restore integrity in
2:06
politics and also have my book, resuscitated nursing,
2:09
given nurses a roadmap to take some of
2:12
that power back and revive our profession.
2:16
Restoring integrity in politics.
2:18
Why do you enjoy slamming your face into
2:20
the wall, Ashley?
2:22
Listen, it's not my face, it's my fists,
2:23
first of all.
2:25
There you go.
2:26
Politically speaking.
2:27
But I'm I'm actually pretty good at it.
2:30
And there's nothing quite like bullying the bully.
2:33
So we talk all the time about how
2:34
important it is to get involved at the
2:36
local level, and you two are exemplars at
2:38
that.
2:40
So well done.
2:41
I am really excited for this episode.
2:43
This is going to be different than anything
2:45
we've done before.
2:47
It's going to be a rehashing of some
2:49
of our prior work, but we are going
2:51
to do it through the lens of real-time
2:53
AI fact checks.
2:55
So when we prepare for episodes, there are
2:58
times where I mean it takes hours and
3:01
hours and hours to pull things together.
3:03
We're covering topics where information is difficult to
3:05
find.
3:07
Sometimes we go to source documents directly from
3:09
lawsuits.
3:11
I swear, I have learned more from lawsuit
3:14
documents about our healthcare system and the medications
3:17
that I've given to patients over the years
3:19
than I've learned from my nursing textbook.
3:22
So I hope that that sort of thing
3:24
is valuable to the audience, but we also
3:26
get pretty fired up sometimes.
3:29
And who knows, maybe we're prone to exaggeration,
3:32
maybe we are just right-wing conspiracy theorists.
3:35
So we're gonna see what different AIs think
3:38
about our prior work by copying and pasting
3:41
AI transcribed prior episodes of ours and just
3:44
see what it says.
3:47
The first AI I'm gonna use here is
3:49
ChatGPT.
3:52
And the first episode I'm going to use
3:54
is our episode about Paxel.
3:57
And I think instead of giving you a
4:00
summary of that episode, we're just gonna see
4:02
what Chat GPT says about it.
4:06
Yeah, it's really kind of interesting to look
4:08
at this from the AI perspective, and it's
4:10
gonna be really interesting.
4:12
I think what Dave is doing going through
4:14
multiple different AIs is kind of cool because
4:17
it really shows the difference and why you
4:20
can't just blanketly trust AI.
4:22
I do use ChatGPT and others for organizing
4:26
my thoughts, but in doing that, I never
4:29
just blanketly trust what it says.
4:33
I'm constantly fact-checking it.
4:35
I am constantly working through and correcting the
4:37
way that it presents my ideas.
4:40
Um, and so the idea of just relying
4:42
on it completely to do these things is
4:44
absolutely wild.
4:47
It's gonna be interesting to see what pops
4:48
up.
4:49
Yeah, I guess a quick disclaimer.
4:51
Um, I use AI to do research.
4:55
I use it to find links sometimes, but
4:56
I never ever trust it.
4:58
It is wrong so often that you have
5:01
to fact-check everything that AI tells you.
5:04
And uh some people haven't learned that lesson
5:06
yet.
5:07
In fact, something I saw in the news
5:10
recently was there was a law firm, this
5:13
incredibly prestigious law firm that charges clients $0,000
5:16
an hour, and they got busted using AI
5:19
in some of their cases, and the AI
5:22
was just making stuff up.
5:25
It was making up case law.
5:26
It was making up prior precedents.
5:28
It was just hallucinating things, and these lawyers
5:31
who were charging five thousand dollars an hour
5:34
were were just using AI and not fact-checking
5:36
it at all.
5:39
So uh when I use AI, I use
5:42
it to find links and I double check
5:44
everything because accuracy is very important to us.
5:47
So about this uh Paxel AI fact check.
5:51
I guess the first thing I'll notice is
5:53
that the AI transcriber misspelled Paxel.
5:56
So right off the bat, uh, we've gotten
5:58
a problem with that, but the chat GPT
6:00
caught that, corrected the spelling, and it's first
6:02
point uh number one.
6:05
Paxel had a big scandal about 20 years
6:07
ago involving kids.
6:09
It deemed that claim of ours mostly accurate
6:10
but simplified.
6:12
It mentions study 329.
6:15
It mentions where it was published just like
6:16
we did.
6:17
Internal documents later revealed that the published article
6:20
misrepresented efficacy and downplayed harms.
6:23
This is a real and well-documented controversy.
6:26
So I don't know why it's accusing us
6:28
of simplifying things because we spent an hour
6:30
talking about this.
6:32
Like, what do you want from us, chat
6:33
GPT?
6:34
Nicole, do you have any thoughts at this
6:35
point?
6:37
Well, I like how it's giving us some
6:39
corrections, but also right off the bat we
6:41
can see that there's some flaws to it.
6:44
And uh, you know, again, spending an hour
6:46
in in your episodes always have a lot
6:49
of great data and detailed information.
6:51
But again, we're pulling this from the notes,
6:53
so it's probably you know, simplifying that as
6:55
well.
6:56
So it, you know, just an observation.
6:59
I mean, we have these episodes where we
7:00
Episodes where we have a topic and we're
7:00
have a topic and we're like, well, I
7:02
like, well, I guess we're gonna need to
7:02
guess we're gonna need to do a two
7:03
do a two-part series, and then the two
7:04
part series, and then the two part series
7:05
part series turns into a three-part series.
7:05
turns into a three-part series.
7:07
This stuff does involve a lot of detailed
7:07
This stuff does involve a lot of detailed
7:10
work, and sometimes that's I don't know.
7:10
work, and sometimes that's I don't know.
7:12
I sometimes that's good for a podcast, but
7:12
I sometimes that's good for a podcast, but
7:14
sometimes you don't want people's eyes to just
7:14
sometimes you don't want people's eyes to just
7:16
glaze over.
7:16
glaze over.
7:17
So, you know, one of the jobs that
7:17
So you know, one of the jobs that
7:19
we have hosting these episodes is to not
7:19
we have hosting these episodes is to not
7:21
have it be just incredibly detailed and boring,
7:21
have it be just incredibly detailed and boring,
7:23
because we want people to actually listen to
7:23
because we want people to actually listen to
7:25
us.
7:25
us.
7:27
So maybe that's a fair criticism by Chat
7:27
So maybe that's a fair criticism by.
7:29
GPT here.
7:31
I don't know that simplified is necessarily a
7:33
negative connotation.
7:36
Um it could just be a good idea.
7:38
Maybe I'm taking that personally, yeah.
7:40
Yeah, I I think that it could just
7:42
be like, okay, this is drilled down.
7:44
It would be interesting to take and ask
7:46
it if you were to do a podcast
7:48
on this.
7:50
How would you expand expound upon this topic
7:53
and just see where it would go?
7:56
Would it lead it down the path of
7:58
really digging into the scandal, or would it
8:01
try to shut it down?
8:03
Yeah, interesting.
8:04
Next claim that we made, they found it
8:06
didn't help at all and increased suicidality six
8:08
to eightfold.
8:11
The ChatGP verdict was exaggerated and partly misleading.
8:15
It says, what actually happened in the original
8:17
trial, study 329, Paxel did not show statistically
8:20
significant benefit on its primary outcome measures compared
8:23
to placebo.
8:25
So I I guess it's saying fact check
8:27
true on that when we said it didn't
8:28
help at all.
8:30
It says that rubber analysis later, researchers concluded
8:33
that there was no clinically meaningful efficacy.
8:37
Okay, that's exactly what we said, and there
8:39
were higher rates of harms, including suicidal ideation
8:41
and behavior.
8:43
However, the six to eightfold increase claim is
8:46
not an established consensus figure in the reanalysis,
8:48
suicidal behavior was higher in the Paxel group,
8:51
but the absolute numbers were small, as is
8:54
typical in these trials, and the relative risk
8:56
varies depending on how events are categorized.
8:59
Across pediatric SSRI trials and more broadly, pooled
9:02
FDA analysis in the early 2000s found about
9:05
a doubling of suicidal ideation behavior compared to
9:08
placebo, not six to eightfold.
9:12
So it's saying, yes, we were correct that
9:15
this did not benefit in terms of depression,
9:18
but we're exaggerating the amount of suicidality.
9:21
It's kind of funny because it's saying we
9:24
should trust them about their claim about safety,
9:28
even though we've just caught them red-handed lying
9:31
about efficacy.
9:34
So that's kind of weird.
9:36
And I I still stand by that six
9:39
to eightfold claim because I mean, you can
9:41
you can look at the rates of the
9:44
prescribing of antidepressants to youth and the rates
9:47
of youth suicide, and they go up together.
9:50
Like those lines are parallel to each other.
9:53
Like more and more and more kids are
9:55
taking antidepressants, and more and more kids are
9:57
committing suicide.
9:59
And again, whether or not it's double, like
10:02
to double their risk of suicide for a
10:05
product that offers no benefit.
10:08
What do you guys think?
10:10
Well, you look at that last line.
10:11
The six to eightfold claim is likely overstated
10:14
or cherry picked from a subset of events.
10:17
That is very interesting.
10:18
They're pointing out magic math or claiming that
10:21
we have some magic math going on here.
10:24
What would be interesting is to pull those
10:26
details from the study and look at it
10:28
to see, you know, where is it, and
10:30
then question AI on that aspect of it
10:33
too, to see if it would continue with
10:35
rhetoric or to see if it would double
10:37
back.
10:40
Yeah, the problem with that study, one of
10:42
the problems was that there just weren't enough
10:45
suicide attempts, thankfully.
10:47
But we also know that in other studies
10:50
when they have uh antidepressant trial that has
10:53
a suicide, they just remove that person from
10:55
the study and they don't even count them.
10:58
They are that ghoulish, and you just can't
11:00
trust them at all.
11:02
So it's like we're making these we're digging
11:04
into numbers that we know are fraudulent, and
11:07
it's like, what's what's the point of that?
11:09
We know that they gain their data.
11:11
So if if you want to know more
11:13
about this one, we we really dig into
11:15
that in that episode about Paxel.
11:18
Nicole, any thoughts here?
11:19
Well, I just you know, uh going back
11:21
to the magic math.
11:22
Is there because I it's a little small
11:24
on my end, I can't really see it.
11:26
But is it is it pointing anything out
11:28
about the numbers in general?
11:30
Uh how we always talk about the data
11:32
not always being accurate to the sample size
11:35
that they use.
11:37
Is it actually admitting that they that maybe
11:39
the process that they use or anything about
11:40
that?
11:42
I was just curious if it fact check
11:45
that, since we tend to always have your
11:47
episodes highlight the importance of the sample sizes
11:50
and like you said, how they're actually picking
11:53
and and the end result, what it looks
11:55
like in these cases where we start off
11:58
with a larger pool of people, and then
12:01
as things move along, like you said, instead
12:03
of them being very transparent and using all
12:06
of those individuals, it starts to kind of
12:09
like pluck them off one by one to
12:12
better suit their study.
12:14
I didn't know if it was catching any
12:16
of that in that chat, GPT.
12:18
Well, so we did not dive into the
12:20
study numbers in detail in our episode, and
12:23
so the fact check isn't diving into those
12:25
study numbers either.
12:28
But they it does bring up magic math.
12:30
That's always uh Ashley's big soapbox is the
12:32
magic math.
12:34
Uh our claim that ChatGPT fact-checked is they
12:36
did magic math and published a completely fraudulent
12:38
article.
12:40
ChatGPT says that is substantially grounded, but rhetoric
12:44
goes beyond legal findings, and it says that
12:48
no court officially found this study fraudulent.
12:52
They found that they lied, they found them
12:55
liable for billions of dollars in damage, but
12:58
that because the court didn't itself officially declare
13:01
this trial fraudulent, it was strong rhetoric by
13:04
us to say that this was a fraudulent
13:06
article.
13:10
I feel like this is just kind of
13:12
a pedantic criticism of us.
13:14
Like, yes, it is true that Glaxo and
13:16
this article were not found fraudulent in a
13:18
court of law, but it was established that
13:21
they were lying, so I don't know what
13:23
you want from us here, Chat GPT.
13:25
I think that's beautiful in all honesty, because
13:28
it's legitimately saying we're not wrong, but the
13:30
courts suck.
13:32
Like, if I'm reading that, that's what I'm
13:34
reading.
13:35
Like, we're not wrong.
13:36
The courts failed to nail it.
13:38
Um I'd call that a win.
13:41
Yeah, like we always say, like, it's not
13:43
just that our regulators are completely broken and
13:45
corrupt.
13:47
Also, our judicial system is completely broken and
13:48
corrupt.
13:50
And ChatGPT is basically agreeing with us on
13:52
that point.
13:54
It says serious misreporting and ghostwriting did occur,
13:56
but the transcript overstates the legal status of
13:59
proven fraud multiple times in court.
14:00
Fraud multiple times in court.
14:02
So yeah, like like we just said.
14:02
So, yeah, like like we just said.
14:04
Next one here, Paxel was given to millions
14:04
Next one here, Paxel was given to millions
14:06
of kids and generated over a billion dollars
14:06
of kids and generated over a billion dollars
14:08
from the adolescent market.
14:08
from the adolescent market.
14:10
It says this revenue claim is plausible, pediatric
14:10
It says this revenue claim is plausible, pediatric
14:13
volume claim unclear.
14:13
volume claim unclear.
14:15
It doesn't know exactly the number number of
14:15
It doesn't know exactly the number number of
14:17
dollars and cents or exact number of kids
14:17
dollars and cents or exact number of kids
14:19
that this was given to, and we acknowledge
14:19
that this was given to, and we acknowledge
14:21
that ourselves because that data just isn't out
14:21
that ourselves because that data just isn't out
14:23
there.
14:23
there.
14:26
So none of us know.
14:26
So none of us know.
14:27
Next one, thousands of kids committed suicide because
14:27
Next one, thousands of kids committed suicide.
14:29
of Paxel.
14:32
It says not supported by established evidence.
14:34
There is a very strong causal claim.
14:36
The evidence does not support that level of
14:38
certainty.
14:40
And I do believe we couched this and
14:43
said that if you extrapolate some of these
14:45
numbers, we can kind of do some napkin
14:48
math that it was probably thousands of kids
14:50
committed suicide just because of how many were
14:53
given Paxel and then using that six to
14:55
eightfold increase in suicidality.
14:58
That's that's where we extrapolated that from.
15:01
But even if you just use what it
15:03
acknowledges as double the risk of suicidality, like,
15:06
okay, we're still talking about, oh, it might
15:09
only be hundreds of kids that Glaxo killed
15:11
for no benefit.
15:14
So the fact check is there's no reliable
15:17
evidence demonstrating that thousands of specific suicides were
15:19
directly caused by Paxel.
15:21
That is speculation, not an established epidemiological finding.
15:25
We couched it as speculation, so I feel
15:27
like this fact check is again agreeing with
15:28
us.
15:30
Any thoughts from you guys?
15:32
I feel like that goes back and supports
15:34
all the things that we've talked about with
15:37
vaccines and the way that the Institute of
15:39
Medicine just kind of says, oh, there's no
15:42
connection, there's no evidence of connection, but it's
15:44
because the studies really haven't been done.
15:47
They really haven't looked at it in something
15:48
like this.
15:50
It's it would be very difficult to say
15:51
yes.
15:53
Absolutely.
15:54
The Paxel is what pushed them over the
15:55
edge.
15:57
And so it's so disheartening.
15:59
I mean, well, it's just like vaccines.
16:01
They use the fact that they have terrible
16:03
data to give themselves plausible deniability.
16:06
And it's that same playbook over and over.
16:10
Next fact, the article was never retracted.
16:12
It says that is correct.
16:14
There's have been calls for attraction, but it
16:16
is not retracted.
16:18
Next fact is the journal is now being
16:19
sued for keeping it up.
16:21
It says fact check true.
16:23
Next one, all antidepressants increase suicidality.
16:27
And it says overgeneralized.
16:29
Uh the summary is increased risk signal exists
16:32
in youth, but all antidepressants increase suicidality is
16:35
an overgeneralization.
16:38
Hmm.
16:39
I don't remember this specific part of our
16:40
conversation.
16:42
I mean, we've talked about how they added
16:45
the black box warning for antidepressants that it
16:47
increases suicidality, they increase suicidality for youth.
16:50
So I think we did cover that.
16:53
Acknowledge that.
16:55
Maybe not specific numbers, but well, I mean,
16:57
maybe it's, you know, we never said that
17:00
every single child who's taken an antidepressant is
17:03
going to have an increase in suicidality, but
17:06
it is true that every single antidepressant has
17:08
that black box warning.
17:12
So I'm gonna fact check the fact checker
17:13
here.
17:15
Which could increase the likelihood of that.
17:18
Right.
17:18
Right.
17:19
Yes.
17:19
So that's that's what we said.
17:21
I I know that because we wouldn't have
17:22
said the other.
17:23
We know better.
17:25
Right.
17:26
Next one here.
17:27
There was never evidence of a chemical imbalance
17:29
or serotonin deficiency.
17:31
ChatGPT says partly correct, but oversimplified.
17:35
The simplistic serotonin deficiency narrative lacks strong evidence,
17:37
but it does not logically follow that antidepressants
17:40
are all risk, no benefit.
17:43
So that sounds like something I might have
17:46
said that antidepressants are all risk, no benefits.
17:49
And you know, I try and throw out
17:51
the disclaimer that I have worked with patients
17:54
who would swear that an antidepressant saved their
17:57
life.
17:59
And if that is someone's experience, that is
18:02
perfectly valid, and I'm happy for that person.
18:05
I just don't think people are given informed
18:07
consent about them, and I think they're over-prescribed
18:09
95% of the time.
18:12
That's just me throwing a number out there
18:14
based off of my experience in this field.
18:17
Ashley.
18:18
Oh, placebos, right?
18:19
30%.
18:20
That's fine.
18:21
40%.
18:22
By by the biggest meta-analysis to date, 40%
18:25
of the efficacy comes from the placebo effect,
18:28
and it's about 13% in the most gamed
18:31
rigged and biased trials to ever exist.
18:35
Those aren't our numbers.
18:36
Those aren't the numbers of antidepressant skeptics.
18:39
Those are the numbers from the manufacturers themselves.
18:42
When you pool together every single antidepressant study
18:45
you can get your hands on from the
18:48
manufacturers, their data is terrible.
18:51
It is terrible.
18:53
You don't need to like cite some guy's
18:54
blog.
18:56
You can just go to their data if
18:57
you can get your hands on it because
18:59
so much of the bad studies aren't even
19:00
released to the public.
19:02
So or the study designs are that terrible
19:04
to show positive effect, they're picking it up
19:06
on placebo.
19:08
I mean, because uh the last time I
19:11
looked and read a study about generalized placebo
19:13
effect was like 30, 33%, I thought.
19:16
Um so when you're looking at these studies
19:18
for antidepressants, is it 30%?
19:22
Is that different to up to 40% because
19:24
of the the design flaw of the studies
19:26
to show a positive outcome?
19:29
Or is the power of the mind and
19:31
thinking that you're getting help for your mental
19:32
health?
19:35
Yeah.
19:35
Just that powerful.
19:36
I mean, very interesting to think about.
19:39
So at some point last year, we looked
19:41
at up to date, which is a clinician
19:44
resource tool used millions of times every day
19:47
by doctors around the world where they look
19:50
up topics for patients that they are treating,
19:53
and we looked at their information on antidepressants.
19:56
They looked at all the data.
19:59
Um, they looked at the biggest umbrella reviews
20:01
and meta-analyses of antidepressants.
20:04
The most recent one, if I remember correctly,
20:06
was by Cipriani et al.
20:09
from 2018, and it looked at 52 different
20:12
antidepressant studies, and their numbers were the pooled
20:16
efficacy was about 53%, and 40% of that
20:19
was placebo effect.
20:23
So again, that's not that's not some guy's
20:25
blog.
20:27
That's from the big pharma companies themselves after
20:30
they were often compelled to release data publicly,
20:33
which was then organized and scrutinized by independent
20:36
actors.
20:39
So uh SSRI use is causing school shootings
20:41
or loss of empathy.
20:44
It says no credible causal evidence supports this.
20:49
Many people take SSRIs without violent behavior.
20:52
Okay, we're we're back to the black box
20:54
warnings, and we're back to just the general
20:55
warnings.
20:57
If you look up any antidepressant, any SSRI,
21:00
Antidepressant, any SSRI, and you scroll down to
21:01
and you scroll down to important side effects,
21:03
important side effects.
21:04
violence is listed, aggressive behavior is listed for
21:07
Violence is listed.
21:07
all of them.
21:08
Aggressive behavior is listed for all of them.
21:11
So does that mean that every single school
21:11
So does that mean that every single school
21:13
shooter was on an antidepressant?
21:13
shooter was on an antidepressant?
21:16
No.
21:16
No, but it is a known side effect
21:16
But it is a known side effect for
21:17
for all of them.
21:18
all of them.
21:19
So okay, chat GPT.
21:19
So okay, chat GPT.
21:21
I you know, we never claimed that every
21:21
I, you know, we never claimed that every
21:23
single school shooter is on an antidepressant, but
21:23
single school shooter is on an antidepressant, but
21:26
is it something that should be looked into?
21:26
is it something that should be looked into?
21:28
Obviously.
21:28
Obviously, obviously.
21:29
Obviously.
21:30
Nicole, thoughts on this one?
21:32
No, I think I think it's just kind
21:34
of proving our point, but it's just, you
21:35
know, doing what ChatGP chat GPT does best,
21:37
playing the other side.
21:41
So it gives us a summary here at
21:43
the bottom.
21:44
It says overall assessment, what's accurate?
21:46
Study 329, misrepresented results.
21:49
Ghostwriting and conflicts of interest occurred.
21:51
Paxel showed no significant primary efficacy in that
21:53
trial.
21:55
Suicidality risk in youth is a real and
21:57
FDA recognized concern.
22:00
The article has not been retracted.
22:01
Pharmaceutical misconduct settlements occurred.
22:04
So it's saying we got all that right.
22:06
What's exaggerated or unsupported?
22:09
The six to eightfold suicidality.
22:11
Okay, we've talked about that.
22:12
Thousands of suicides.
22:13
That was speculation.
22:15
All antidepressants are all risk, no benefit.
22:18
If I did say that in that episode,
22:20
I was probably fired up, and I will
22:22
stand by the chat GPT correction here.
22:25
I should not have said that.
22:26
I I do have to acknowledge that some
22:28
people will tell you that they have benefited
22:31
incredibly from taking an antidepressant, and that's perfectly
22:33
valid.
22:35
Uh assertions linking SSRIs to school shootings.
22:38
Again, I stand by our claim.
22:40
You can read it on every single page
22:42
package insert that violence is a potential side
22:44
effect of SSRIs.
22:47
Bottom line, the podcast is built around a
22:50
real and serious case of research misrepresentation.
22:53
One of the most cited examples of pharmaceutical
22:55
influence in psychiatry.
22:57
However, the episode mixes legitimate historical scandal, documented
23:01
regulatory failures, accurate concerns about suicidality in youth
23:05
with inflated statistics, causal claims not supported by
23:08
population data, sweeping generalizations about antidepressants, rhetoric that
23:12
goes well beyond established evidence.
23:16
I like how it says, like, well, you
23:17
didn't get your facts wrong, but your rhetoric
23:19
is pretty strong.
23:22
Okay, chat GPT.
23:23
Any more thoughts from you guys, or should
23:25
we jump to the next one here?
23:27
We gotta take a break here pretty soon,
23:29
so I think you should go keep going.
23:30
But it's like, yeah, your rhetoric is pretty
23:31
strong too, chat.
23:33
Keep keep going.
23:35
All right.
23:36
Uh the next one I want to take
23:38
a look at is kind of along the
23:41
same lines because it's the same company.
23:43
This is Glaxo again, and we did an
23:46
episode about Glaxo where they had this incredible
23:50
bribery scandal in China.
23:53
So for this one, I'm gonna go to
23:56
uh Google Gemini and ask it about that.
23:58
While you get that cued up, why don't
23:59
we go ahead and take a real quick
24:00
break?
24:02
We are the nurses report here on America
24:03
Outloud.news.
24:05
You can hear us every Monday through Friday
24:07
at 10 a.m.
24:09
Eastern time.
24:09
And then if you miss us, make sure
24:11
you catch us on the replay in your
24:12
favorite podcasting app.
24:14
We'll be right back after this quick break.
24:16
Your heart is crucial.
24:17
Are you supporting it as you age?
24:18
Try the wellness company's healthy heart formula with
24:21
taurine, deribose, and coenzyme Q10.
24:24
Designed by Dr.
24:24
Peter McCullough, Healthy Heart helps boost heart function
24:26
and support cardiac strength.
24:29
Get 20% off your first order plus free
24:31
shipping at TWC dot health forward slash out
24:33
loud with code out loud.
24:36
Struggling with your health, energy, and constantly feeling
24:37
hungry.
24:39
Want to save money on the rising cost
24:42
of groceries and dramatically improve your health?
24:44
Go to ChemicalfreeBody.com forward slash out loud today,
24:47
get nature's super multivitamin, doctor formulated green 85
24:50
juice formula, empower your immune system, cut your
24:52
grocery bill, and save 20% on your first
24:55
order.
24:58
Oral hygiene hasn't changed in 50 years, but
25:01
our diet and the way we eat has.
25:03
Creating an environment in your mouth for bacteria
25:06
to wreck havoc on your teeth and gums.
25:08
For better oral health, get spray dental defense,
25:11
an oral care line designed to combat acid-creating
25:13
bacteria.
25:16
The toothpaste, mouthwash, mints, and gum all contain
25:19
xylitol, and natural ingredients shown to dramatically improve
25:21
oral health.
25:24
Sprite can be found online and at all
25:26
fine natural retailers.
25:30
Well, it's high time Americans refuse to be
25:32
haunted in the human chest.
25:35
Under the name of diversity, politicians divide the
25:38
people in their quest for mom.
25:41
Well, diversity of opinion and expression is as
25:44
American as Apple Pond.
25:50
Now is our time.
25:53
America Outlaw.news.
25:56
Liberty and justice for all.
26:02
Hello and welcome back to this episode of
26:04
the Nurses Report here on the America Outloud
26:06
Network.
26:08
We are talking about AI, and we're using
26:11
AI to fact check some of our prior
26:14
episodes where we've talked about different scandals.
26:17
My background is as a psych nurse, so
26:20
a lot of my episodes end up being
26:22
about psychmeds, and we're having a pretty interesting
26:25
time seeing what AI thinks about our opinions
26:28
as we dive into some of the lawsuits
26:31
and scandals and the evidence base for antidepressants.
26:34
And that's where we're going to pick up.
26:37
So we just covered a scandal involving Glaxo's
26:39
drug Paxel, and now we're going to fact
26:42
check using a different AI and a different
26:44
episode.
26:47
And make sure that we let everybody know
26:50
that this is pretty cool because the different
26:52
types of AI can give us different outcomes
26:55
that uh we may not have seen from
26:58
a previous AI platform.
27:00
So, you know, at this point, we use
27:02
Chat GPT for the first one, and Dave's
27:04
got on his screen right now.
27:07
So this is another platform.
27:09
Gemini says the podcast transcript you provided from
27:11
the Nurses Report on the America Out Loud
27:13
News Network discusses a very real and significant
27:15
historical event.
27:18
The 2013 Glaxo bribery scandal in China.
27:22
While the podcast includes the host's strong personal
27:24
opinions, there we go.
27:25
I guess I guess we're pretty strongly opinionated.
27:28
Uh, and rhetorical flourishes.
27:30
I like that.
27:31
Rhetorical flourishes.
27:32
I think they're talking about you, Ashley, referring
27:34
to executives as pharma ghouls.
27:37
Well, I get that was that was probably
27:38
me.
27:39
Or criminals, I stand by that claim.
27:41
The core historical facts regarding the investigation, the
27:44
sex tape, and the private investigators are largely
27:46
accurate based on reporting from the New York
27:49
Times, Reuters, and the Financial Times.
27:52
Here is a fact check of the specific
27:53
claims made in the transcript.
27:56
Hey, I think we're doing pretty well so
27:57
far.
27:58
It says that we're very opinionated and spicy,
28:00
Opinionated and spicy, but largely factually accurate, which
28:01
but largely factually accurate, which is I think
28:03
is I think what we're aiming for.
28:03
what we're aiming for.
28:06
Ashley, is that what you aim for?
28:06
Ashley, is that what you aim for?
28:08
A hundred and ten percent.
28:08
A hundred and ten percent.
28:11
I think that's just the way God made
28:11
I think that's just the way God made
28:12
me.
28:12
me.
28:13
A little spicy, a little sassy, but factually
28:13
A little spicy, a little sassy, but factually
28:15
accurate.
28:15
accurate.
28:18
All right, fact check number one the bribery
28:18
Alright, fact check number one, the bribery allegations
28:19
allegations and the playbook.
28:19
and the playbook.
28:21
The claim.
28:21
The claim GSK China used travel agencies and
28:22
GSK China used travel agencies and third party
28:24
third party vendors to funnel nearly 500 million.
28:25
vendors to funnel nearly 500 million.
28:27
The tramp skip says 100 million per.
28:27
The trampscript says 100 million per year in
28:30
bribes to doctors and officials.
28:32
Fact check, true.
28:34
Awesome.
28:35
Then we talk about the fine, and it
28:36
says the fine, true.
28:38
Uh I'm not gonna read more details.
28:41
If you want to hear the details, pull
28:43
up the episode, but I I like that
28:45
so far uh we're doing pretty well.
28:47
Number two, the private investigators, Humphrey and Shu.
28:51
The claim.
28:52
GSK hired a husband wife private investigator team
28:54
to investigate a whistleblower, and they were subsequently
28:57
in prison.
28:59
Fact check true.
29:01
Awesome.
29:01
All right, next one the sex tape scandal.
29:03
The claim a sex tape involving GSK is
29:05
China head.
29:06
Mark Riley was sent to GSK executives.
29:09
Fact check true.
29:10
Awesome.
29:11
I feel like we're getting off easy on
29:12
this one, Nicole.
29:14
Are you kind of surprised that it's not
29:15
even like pushing back?
29:17
Not a one, not a one pushback, and
29:19
it's kind of funny because I'm tempted to
29:20
see if you could put this in chat
29:22
DPT if it would give us something different,
29:24
since the other one critiqued us so much
29:26
more.
29:28
It did, yeah.
29:29
Well, you know what?
29:30
Let's do that.
29:30
Let's finish this one and then see.
29:32
Yeah, let's let's see what some other ones
29:33
say about this.
29:35
Next one, safety concerns, Lamictal and Paxel.
29:38
The claim GSK bribed regulators to hide side
29:41
effects of Lamygdal, such as Stevens Johnson syndrome
29:43
and illegally marketed Paxel to children.
29:46
Fact check.
29:47
Mixed.
29:48
Context required.
29:49
Okay, so it's not giving us a fact
29:50
check true on this.
29:52
Let's find out why.
29:53
It says it is true that GSK reached
29:56
a three billion settlement with the U.S.
29:58
Justice Department for various infractions, including the illegal
30:02
marketing of Paxel to children and adolescents.
30:05
Okay.
30:06
Lamictal, while Lamictal can cause Stevens Johnson syndrome.
30:10
The specific claim that GSK used red packages
30:13
of cash in China specifically to hide deaths
30:15
from Lamicdal is a narrative derived from the
30:18
broader bribery charges.
30:20
The 2014 conviction focused on bribery to increase
30:23
sales rather than a specific cover up of
30:25
a single drugs side effects.
30:28
I feel like it's putting words in our
30:30
mouth.
30:32
We said that they were specifically trying to
30:35
cover up that side effect and that side
30:37
effect can be fatal, but we didn't say
30:40
anything more specific than that.
30:43
So I feel like it's I don't know.
30:45
I I stand by what we said.
30:47
Ashley, do you have thoughts on this one?
30:49
I based on context I'm reading.
30:52
I feel like it thinks the red packages
30:55
of cash is where the issue is, right?
30:58
Because is red packages of cash in the
31:02
testimony or like the red envelopes that they're
31:05
giving, or is it in the final verdict?
31:09
And so I wonder if that's like their
31:11
sticking point, and that's what they're getting uptight
31:12
about.
31:14
But I I feel like it should be
31:15
true.
31:16
Like this feels pretty accurate.
31:18
Yeah, they weren't disputing that these red packages
31:20
of cash existed.
31:22
Like that's something that was established in the
31:25
lawsuit that that Glaxo was doing.
31:28
Um that's just what they were referred to
31:30
as, and some of the people that they
31:32
were giving them to were people who had
31:34
this side effect, but not necessarily people who
31:36
were, you know, or sort of family of
31:38
people who were killed by this side effect.
31:41
So I don't know.
31:41
It's a weird thing for it to quibble
31:43
about, so I won't quibble with it anymore,
31:45
other than to say I stand by uh
31:47
what we what we said.
31:49
Next one, whistleblower Vivian She.
31:52
The claim.
31:53
GSK suspected a former employee named Vivian She
31:54
was the whistleblower and an asset of the
31:56
CCP.
31:59
Fact check unverified.
32:00
We were very clear about that in this
32:02
episode that everyone thought it was her, but
32:04
there was no conclusive proof that it was
32:05
her.
32:07
So again, I you know, if it says
32:10
fact check unverified, and yeah, exactly.
32:13
That's exactly what we said.
32:14
So the summary table here, it's got claims,
32:17
it's it made a table for us.
32:19
So, first of all, let's just say that's
32:22
kind of cool, Gemini, that you made a
32:24
table of all the uh fact checks that
32:26
you conducted on our work here, and it
32:29
has it all correct except for that little
32:31
nuance about the red packages that that we
32:34
already addressed.
32:36
So the verdict, while the podcast uses a
32:39
highly spicy and provocative tone on that reco.
32:43
That's awesome.
32:44
Uh Ashley.
32:46
Ashley, yeah, exactly.
32:49
Yeah.
32:50
The timeline and the primary scandalous elements of
32:53
the GSK China bribery case align with the
32:57
documented outcomes and investigative journalism from the mid-2010s.
33:00
Hey, good job, guys.
33:02
Let's go to Chat GPT and just put
33:04
the exact same thing in there.
33:07
That's what I'm thinking.
33:09
Yeah, the copy and paste should be great
33:12
because you're feeding it the same exact information.
33:15
Same exact information, and see what kind of
33:17
stuff it comes back.
33:19
And I really do think that it's it
33:22
was quibbling over the use of red cash
33:24
packages.
33:27
I really think that that's the problem it
33:28
had.
33:29
Oh, good old chat GPT.
33:32
Oh, chat GPT.
33:34
All right.
33:35
Uh the GSK China bribery scandal.
33:38
Core facts accurate.
33:40
Hmm.
33:41
It's liking this one apparently a lot more
33:44
than our Paxel episode.
33:47
But uh partly accurate, overstated.
33:51
GSK had a 100 million per year bribery
33:53
budget.
33:55
Chinese prosecutors allege bribes totaling around 3 billion
33:58
one or 480 million dollars over several years,
34:01
not per year.
34:04
Uh oh, okay.
34:05
So I think we did the math on
34:07
that and just estimated it was a hundred
34:08
million a year.
34:10
It's like, what it's not like that discredits
34:11
us at all.
34:13
It's like it's saying it wasn't a hundred
34:15
million a year, it was four hundred and
34:17
eighty million they were spending on bribery over
34:19
several years.
34:21
Okay, chat GBT, you got me.
34:23
Gosh, that makes it so much better that
34:26
they were spending that much on bribery.
34:28
I I don't know if I want to
34:30
dive into that fact check any more than
34:31
that.
34:33
I just think that's kind of silly, but
34:35
uh claim GSK rehired alleged whistleblower Vivian Chi.
34:38
There is limited reliable reporting.
34:40
So again, we're back to this person.
34:43
Um, and yeah, we said that the fact
34:45
that she was a CCP asset was completely
34:47
speculative.
34:50
We acknowledge that.
34:51
We also said that everybody involved believes it
34:53
was her, so again, that's not our speculation.
34:56
That's that's their speculation.
34:58
Next part, Paxel scandal, accurate elements in 2012.
35:00
Paxl scandal, accurate elements in 2012, GSK agreed
35:02
to pay three billion.
35:03
GSK agreed to pay three billion.
35:05
So again, it's saying like, well, your facts
35:05
So again, it's saying like, well, your facts
35:07
are mostly right here.
35:07
are mostly right here.
35:09
Uh, it does have an overstatement.
35:09
Uh it does have an overstatement.
35:10
It says a transcript implies Paxel caused hundreds,
35:10
It says a transcript implies Paxel caused hundreds,
35:13
if not thousands of youth suicides.
35:13
if not thousands of youth suicides.
35:15
There is no established number of confirmed suicides
35:15
There is no established number of confirmed suicides
35:18
directly attributed to Paxel.
35:18
directly attributed to Paxel.
35:20
All right.
35:20
All right, we already we already talked about
35:20
We already we already talked about that.
35:21
that.
35:22
This was speculation on our part.
35:22
This was speculation on our part.
35:25
We're trying to do some napkin math here
35:25
We're trying to do some napkin math here
35:26
because God knows our regulators haven't done this,
35:27
because God knows our regulators haven't done this,
35:28
and we're
35:29
and we also know our judicial system is
35:31
not holding them accountable.
35:33
It's it's like we talk about with all
35:36
these class action lawsuits that go to Delaware
35:39
to die because the way they have these
35:42
corporate-friendly courts set up in Delaware, they say,
35:45
like, well, you can't prove your child killed
35:48
themselves because of Paxel with 100% confidence, therefore
35:51
you can't sue Glaxo.
35:54
That's a topic we hit on a lot,
35:56
is how the judicial system is just rigged
35:58
in situations like this.
36:00
And so you don't have these convictions, then
36:03
the you know, the AIs are fed off
36:06
this information where they say, well, it wasn't
36:09
proven, there wasn't a judge that said it.
36:12
There wasn't the New York Times that said
36:14
this is what's happening.
36:15
So then we get uh then we get
36:17
a fact check false or an overstatement sort
36:18
of thing.
36:20
Nicole, what do you think?
36:21
I mean, it's again just taking a lot
36:23
of the information that we say, and it's
36:26
gonna skew it to benefit the the AI
36:29
portion of it, because again, it's gonna give
36:31
the idea that these people are doing exactly
36:34
what we say that they're doing, even though
36:36
there's a lot of lawsuits that haven't been
36:39
able to prove it entirely.
36:42
We know what's been going on, and so
36:44
it doesn't surprise me that it's saying that
36:47
we're overstating some of these things or try
36:50
to make it seem like it's something that
36:53
it's not when we know it is and
36:56
then uh the rest of the fact checks
36:59
that Chat GPT is giving us is about,
37:01
you know, we were trying to lay out
37:04
some patterns, so we were just saying how
37:07
this scandal with Galaxo is the exact same
37:10
playbook as some other scandals that we've looked
37:13
at, and and of course, Paxel came up,
37:16
Xantac came up, VyOX came up, um, McKinsey
37:18
and Purdue, Purdue, of course, with OxyContin, uh,
37:21
and then the Pfizer COVID trial with Ventavia.
37:24
I don't know how much we want to
37:26
um I'll scroll through them, and if people
37:29
want to kind of uh pause it on
37:31
any of these, they're they're welcome to, but
37:33
we've done entire episodes on on some of
37:35
these, and I think in our episode uh
37:38
that it's fact-checking, we just kind of mentioned
37:40
them in passing.
37:42
We didn't dive into the details on them,
37:44
and so I'm just gonna largely skip past
37:45
them.
37:47
Hold on, though.
37:48
Like for one second.
37:50
Number 10.
37:50
Nobody goes to prison.
37:52
Stop right there.
37:52
Yeah.
37:53
Partly true, partly overstated.
37:55
Corporate executives are really imprisoned in the U.S.
37:58
pharma settlements.
37:59
I would venture to say that it's probably
38:01
a never never.
38:04
Yeah, none that I'm aware of.
38:05
I, you know, we there was one doctor
38:08
involved in that uh ghostwriting of the Paxel
38:12
scandal that ended up losing his medical license.
38:15
It wasn't because of Paxel, it was because
38:17
they started to look at all his academic
38:20
work, and uh it was something like 68
38:22
out of 70 of his articles were completely
38:24
fraudulent, so that's why he lost his medical
38:26
license.
38:28
But yeah, he was never criminally charged.
38:31
No one has ever criminally charged.
38:32
Sorry to interrupt, go on.
38:34
Yeah, that's okay.
38:34
I I mean I'm just reading what ChatGPT
38:36
is saying.
38:37
But however, in China, GSK executives were uh
38:40
convicted and sentenced, though some were suspended.
38:44
So that's really interesting when you look in
38:46
contrast America versus China.
38:49
They're willing to put some punishment in, even
38:51
if it was suspended.
38:54
It's interesting to follow up.
38:57
They they threw executives in jail while the
38:59
trial was going on, and they ended up
39:01
uh sending them back to the UK.
39:04
They didn't put them in prison in China,
39:06
but they did put them in uh a
39:09
very secretive jail, actually, along with the private
39:12
investigators who were basically innocent people caught up
39:14
in stepping on the toes of the CCP.
39:17
So even though they did nothing wrong, they
39:19
were investigating this scandal and uncovering truths, they
39:22
went to jail too.
39:24
And and the uh husband of that husband-wife
39:27
duo has been very outspoken about the horrible
39:30
conditions that they were imprisoned in China.
39:34
And I would wish that upon every pharma
39:36
executive that we've talked about on this show.
39:39
I I think they all deserve prison time.
39:41
Yeah.
39:42
It's just it's really surprising to me that
39:45
there is never any accountability that that wouldn't
39:48
even be thought of in the United States,
39:52
and yet China is working to maybe protect
39:55
its people better than the United States is.
39:58
Like, can we just let that sink in
39:59
for a second?
40:01
I ain't buying it.
40:02
I ain't buying it.
40:03
Yeah, no, neither.
40:04
China requires that if you're committing fraud, you
40:07
have the government participate in the grift.
40:11
So it's not that they're protecting people, it's
40:13
that they are insisting that they are included.
40:16
Like that's that's that was the point of
40:18
the entire Oh, come on.
40:20
No, that's that's I know, I know that
40:22
was the lesson learned there.
40:25
I know, but man, it's just like, come
40:27
on, America.
40:29
Pharma business model is fraud.
40:31
That's something that uh I repeat after hearing
40:32
Dr.
40:34
Asim Malhotra say it, and I stand by
40:35
that.
40:37
ChatGPT says this is an opinion, not a
40:39
factual claim.
40:41
Debatable.
40:43
Uh sex workers and honey traps, Chinese authorities
40:46
alleged bribery that included travel gifts and possibly
40:49
sexual services, but detailed proof of widespread high-class
40:52
prostitute schemes is largely based on Chinese prosecution
40:55
claims and media reporting.
40:58
Not publicly documented court evidence in detail.
41:01
This is only interesting because it shows what
41:04
these AI chatbots prioritize.
41:07
Like if it's been established in court, they
41:10
give that a high amount of credibility.
41:12
And like we talked about just a little
41:15
while ago, one of the problems that we
41:17
see is so many of these pharma lawsuits
41:19
are conducted in secrecy under gag order and
41:22
settled under non-disclosure agreements, so that stuff doesn't
41:24
go public.
41:27
That was the whole scandal with the Zyprexa
41:29
papers that we've talked about, is that here
41:32
are these people involved in a lawsuit where
41:34
in court it's being established that EI Lilly
41:37
is knowingly harming people, that their medication causes
41:39
side effects that profoundly affect their quality of
41:42
life and shorten their life, and nobody in
41:44
the public knows about it because they're lying
41:47
about the side effects.
41:50
And that turned into a huge scandal, not
41:52
because it wasn't true, but because it wasn't
41:54
supposed to be released to the public, and
41:56
Eli Lilly went after the leakers hard in
41:58
court.
42:00
So yeah, it's it's interesting that um that
42:00
So, yeah, it's it's interesting that um that
42:03
these AI chatbots prioritize things that are established
42:03
these AI chatbots prioritize things that are established
42:06
in court when a lot of what we
42:06
in court when a lot of what we
42:09
talk about on our episodes is how grossly
42:09
talk about on our episodes is how grossly
42:13
the judicial system is completely failing us with
42:13
the judicial system is completely failing us with
42:16
a lot of these lawsuits, just through no
42:16
a lot of these lawsuits just through no
42:19
consequences or through a complete lack of transparency.
42:19
consequences or through a complete lack of transparency.
42:22
Nicole, what do you think?
42:22
Nicole, what do you think?
42:24
Yeah, it reminds me of the Gardasil, one
42:24
Yeah, it reminds me of the Gardasil, one
42:26
of the Gardasil episodes we did where there
42:27
of the Gardasil episodes we did where there
42:28
was
42:29
was some sort of loophole, remind me.
42:32
But and all those, I think there was
42:35
like 200 cases that were going to be
42:37
heard on a national level, and they all
42:40
got dropped basically for this one um particular
42:43
outcome that happened with a specific case, and
42:46
you know, just to know how many people
42:48
had tried to come forward and how broken
42:51
the system is, and none of them received
42:54
any justice because all of those cases, due
42:56
to that one instance, uh dropped them all.
42:59
So, like you said, it's just a constant
43:02
uh wound and traumatic, you know, outcome for
43:05
these people who just keep being abused by
43:08
the system with no reprieve ever.
43:11
The claim that we made that the uh
43:14
that chat GPT has the biggest problem with
43:17
was regarding the OxyContin scandal, and that was
43:20
where I claimed that Purdue hired the McKinsey
43:23
Institute in order to help them with marketing
43:26
and promoting OxyCotton while some of those exact
43:29
people were working at the FDA, and it
43:32
says that that is unsupported.
43:35
And I know I've read that.
43:37
I know that that's established out there, so
43:39
I I'm gonna contend that one.
43:42
Either of you remember that one off the
43:43
top of your head.
43:45
Is it really the McKinsey, or is that
43:47
the McKinn um the Kinsey Institute out of
43:50
IU?
43:53
I thought it was the Kinsey Institute.
43:55
Let's just look at Google.
43:58
All right, we're gonna fact check the fact
43:59
checker.
44:01
Fact check the fact checker, because we might
44:03
be we might be on to something here.
44:05
All right, can one of you do that?
44:06
And I'm gonna read major patterns in the
44:07
transcript.
44:09
It says real scandals are mixed with rhetorical
44:13
escalation, documented legal cases are blended with speculation,
44:16
unproven allegations are something presented as fact, and
44:19
separate events are linked narratively without evidence.
44:23
Okay.
44:24
Separate events are linked narratively without evidence.
44:27
You know, we talk about all these different
44:29
pharma scandals, and a lot of the time
44:32
it is the exact same time frame with
44:34
many of the exact same players at the
44:36
company.
44:39
So when we talk about VIOX, that was
44:42
the exact same time frame of the Protocol
44:44
7 scandal with the MMR vaccine.
44:47
It was the exact same executives working there
44:49
at the time.
44:51
A lot of the time it's the same
44:53
people at the FDA, or it's, you know,
44:55
a lot of the people involved in these
44:57
scandals then go to the FDA, or they
44:58
start at the FDA and then they go
45:00
to the pharmaceutical companies.
45:02
So it's saying, like, oh, you're connecting these
45:05
things when there's no clear evidence they're connected.
45:08
Yes, there is.
45:10
I would argue all day that there is
45:12
a connection there.
45:14
Ashley.
45:15
Yeah, so this is really frustrating because I've
45:17
got the New York Times headline.
45:19
McKinsey settles for nearly 600 million over rule
45:22
and opioid crisis from 2021.
45:25
McKinsey advised Purdue Pharma on how to turbocharge
45:28
opioid sales going from 200 to 400 million.
45:31
That was February of 2019.
45:33
McKinsey apologizes for helping Purdue Pharma turbocharge profits
45:36
2020.
45:38
Um people v.
45:40
McKinsey file stamped complaint.
45:42
Purdue hired McKinsey to increase brand loyalty to
45:44
OxyCotton.
45:46
That's out of the State of California Department
45:47
of Justice.
45:49
So NPR, New York Times, New York Times,
45:51
and the California Department of Justice are all
45:54
saying that you're factual.
45:56
Well, so my claim went a little further
45:58
because I said that some of the people
46:01
working at McKinse were simultaneously working at the
46:03
FDA.
46:05
Can you ask it if that's true?
46:07
Okay, let me ask.
46:08
Okay.
46:09
Back to this one on ChatGPT here.
46:11
It says bottom line: the episode is built
46:13
around a real and significant GSK bribery scandal,
46:15
and that core story is grounded in documented
46:17
legal proceedings.
46:19
However, some financial numbers are exaggerated.
46:22
Some claims McKinsey slash FDA duo roles are
46:24
incorrect.
46:26
Ashley's about to get us the facts on
46:27
that.
46:28
COVID connections are speculative.
46:30
We never said they weren't speculative.
46:33
Uh several interpretations are presented as factual conclusions.
46:37
Go ahead, Ashley.
46:39
U.S.
46:39
food and drug McKinsey and company face significant
46:42
scrutiny and legal repercussions for deep undisclosed conflicts
46:44
of interest while consulting for the U.S.
46:47
Food and Drug Administration.
46:48
So they're consulting.
46:50
Um and major opioid manufacturers, notably Purdue Pharma,
46:53
simultaneously.
47:04
Which included failing to disclose these conflicts to
47:06
the FDA.
47:08
Okay.
47:09
So I said that they were simultaneously working
47:12
for the FDA, and technically they were not
47:14
direct FDA employees.
47:17
They were merely hired as consultants to the
47:19
FDA.
47:21
So gosh.
47:22
You know, if that's its biggest quibble, if
47:24
that that's hold on.
47:26
Okay.
47:27
Wait for just a second.
47:28
Simultaneous representation.
47:29
At least 22 McKinsey consultants worked on FDA
47:32
contracts while also working for opioid manufacturers, including
47:36
Purdue Pharma on related topics.
47:39
Some consultants worked for both at the exact
47:40
same time.
47:42
Okay, thank you.
47:43
Thank you.
47:44
All right.
47:46
Suck it, chat GPT.
47:47
We were right.
47:49
Why did it get that wrong?
47:50
Because you're pulling that from like oh.
47:53
We should make sure people really understand what
47:55
just happened.
47:56
Can you just recap that really quick?
47:58
So people that are following, trying to follow,
48:00
they can just hear what just happened.
48:03
Okay, so we're using ChatGPT to fact-check transcripts
48:06
of prior episodes of the nurses report.
48:10
And we just asked that to fact-check an
48:13
episode we did about this big GLAXO scandal
48:16
in China, and ChatGPT was pretty generous with
48:19
us.
48:22
It says, you know, you got most of
48:23
your claims correct.
48:25
The one that it had the biggest problem
48:28
with, the one that it says is unsupported
48:31
was that people from the McKinsey Institute were
48:33
working on supercharging sales of OxyContin for Purdue
48:36
pharmaceuticals at the exact same time they were
48:39
being employed at the FDA for jobs like
48:41
overseeing narcotics and opioids and that sort of
48:44
thing.
48:47
And ChatGPT said nope, that's unsupported.
48:50
This it was by far its biggest quibble
48:53
with our episode.
48:55
This is the fact that it said we
48:56
got wrong, and we we just looked it
48:58
up and no, we were totally right about
49:00
No, we were totally right about this.
49:00
this.
49:02
Totally right.
49:02
Totally right.
49:03
This is from committee release report uncovering significant
49:03
This is from committee release report uncovering significant
49:07
conflicts of interest at McKinsey and company related
49:07
conflicts of interest at McKinsey and company related
49:10
work for FDA and opioid manufacturers.
49:10
work for FDA and opioid manufacturers.
49:14
This is from the Democrats uh Oversight Democrats
49:14
This is from the Democrats Oversight Democrats House.gov.
49:16
House.gov.
49:19
And it directly quotes the Oversight's Committee investigation
49:19
And it directly quotes the Oversight's Committee investigation
49:22
is shining a spotlight on the unregulated and
49:22
is shining a spotlight on the unregulated and
49:24
secretive world of private consulting firms like McKinsey
49:25
secretive world of private consulting firms like McKinsey
49:27
that created conflicts of interest by working.
49:27
that created conflicts of interest by working for
49:30
both federal government and regulated interest industries.
49:33
This is not just us saying this.
49:35
This is this website.
49:36
Today reports show that the same time the
49:38
FDA was relying on McKinsey's advice to ensure
49:40
drug safety and protect American lives.
49:43
The firm was also being paid by the
49:45
very companies fueling the deadly opioid epidemic to
49:48
help the help them avoid tougher regulation of
49:51
these dangerous drugs.
49:54
McKinsey staffed at least 22 consultants at both
49:57
the FDA and opioid manufacturers on related topics,
50:00
including at the same time.
50:04
And it goes on.
50:05
But how is that not a conflict of
50:06
interest?
50:07
How is that allowed?
50:08
I how do well we just talked about
50:10
why don't people go to prison?
50:12
Like we know that this led to many,
50:14
many tens of thousands of deaths.
50:17
I mean, we're at what something like a
50:19
hundred thousand deaths per year with the opioid
50:21
crisis at this point, give or take.
50:23
It's it's uh yeah, it's insane.
50:26
So I think I said like three hundred
50:28
million.
50:30
This says the American public will soon learn
50:32
even more about the McKinsey role in the
50:35
opioid epidemic, thanks to the hard work of
50:37
State Attorney's General and their landmark 573 million
50:40
dollar settlement.
50:43
This is just wild.
50:45
I mean, so it just goes to show
50:48
you we gave that disclaimer in the beginning.
50:50
Like when you're using AI, you absolutely need
50:52
to fact check it.
50:55
So we found a point of contention where
50:57
we said, I don't like what AI is
50:59
saying here.
51:01
I don't know, should we talk about confirmation
51:03
bias at all, or should we talk about
51:05
the risk that you have, you know?
51:07
AI, it doesn't want to tell you the
51:08
truth.
51:10
AI wants you to keep engaging.
51:13
So to some extent the AI is incentivized
51:15
to tell you sweet lies that it knows
51:17
you're going to enjoy to keep you using
51:20
the platform.
51:22
That's how these algorithms work.
51:24
So is that one of the risks here
51:25
that AI is going to have a profile
51:27
about you and it's not going to tell
51:29
you the truth.
51:30
It's going to tell you what you want
51:32
to hear to keep you engaged.
51:34
Well, for someone who doesn't use AI, I
51:36
don't use these platforms because it's just I'm
51:38
one of those weird people.
51:41
Um I'm told that it tends to learn
51:43
you, or that it has to learn you.
51:45
I've heard people make comments like, hey, the
51:47
more you use it, it kind of learns
51:49
kind of how you give it information and
51:50
what you feed it.
51:52
Is that not part of how the AI
51:54
responses then come back?
51:56
Would it be different from different person depending
51:59
on how you give the information?
52:02
I can punch in information, the same information.
52:05
Like if I don't like an answer or
52:07
the way that it presented, but I don't
52:09
really I don't do it for research.
52:11
I do it to help me organize my
52:14
thoughts and to like pull in different strategies,
52:16
kind of like a brainstorming session.
52:19
But it definitely does change based on the
52:21
way you put the information in and the
52:24
questions you ask, which could be like query-based.
52:26
I argue with ChatGPT all the freaking time
52:29
because I'm constantly I'm constantly asking about vaccines,
52:32
and it will go and it will sway
52:35
into these other realms and go off the
52:39
rails of what I want it to.
52:42
I'll correct it and it'll be like, that's
52:43
you're right.
52:45
You're right.
52:45
And even things like putting together some questions
52:48
for a Bible study or whatever, it will
52:51
take me completely in the opposite direction of
52:54
what I want it to do.
52:57
And I'm like, well, you're not even getting
52:58
that right.
52:59
And so I'll correct it.
53:00
It's like you're right.
53:02
And I've had it analyze my personality and
53:05
tell me what it knows about me.
53:08
We had a deep conversation one night.
53:10
I was I had insomnia, and I was
53:12
like, okay, chat.
53:13
If if you were to psychoanalyze me based
53:16
on the conversations that we've had, um, go
53:19
through, and it was scary accurate and hit
53:21
on some things that I was like, I
53:24
don't know that I really admit a lot
53:26
of this stuff, but apparently it's reading that
53:29
in my writing and the things, the information
53:32
that I'm giving it, because I'll have it
53:34
help me with different projects and whatever for
53:37
brainstorming.
53:40
So it's scary, and I'm sure all this
53:43
information is being used for crazy stuff.
53:47
And it's not private, like the information that
53:49
you plug into a chat bot and what
53:51
it learns about you, you don't have control
53:53
over that.
53:55
That is their property.
53:56
That's not your property, even though it can
53:58
get to know you that well.
54:01
These are wild times we're living in people.
54:04
So I plugged this into GROC, and I
54:06
think I'm signed into my GROC account and
54:08
I have it set to give me really
54:10
short answers.
54:13
So I pasted the exact same thing into
54:15
there, and it says transcript largely accurate on
54:18
core GSK China bribery scandal, and then it
54:20
kind of lists all the facts that we
54:23
laid out, and then it says that um
54:25
some of it is unsupported or speculative, such
54:28
as Lamycto Hush money, direct tie to the
54:30
COVID lab leak, and that uh again that
54:32
um hundred million dollars a year bribery budget,
54:35
which that's the same fact check that we
54:37
got from the other one too, where it
54:40
was like, no, it was 480 million over
54:42
several years.
54:45
Like, oh, okay.
54:46
But ask it how many years the several
54:47
years is.
54:49
Just for fun, just for giggles.
54:51
And then we we probably need to wrap
54:52
it up here.
54:54
I almost wanted to ask you to elaborate
54:56
what it what it like, what is it
54:58
like arguing with AI?
55:00
Like what?
55:01
When you guys said you're fighting, what does
55:02
that mean?
55:03
But that'll be for another episode.
55:05
Um, we should probably do a whole episode
55:08
about talking to AI about vaccines because you
55:11
know, it goes off the freaking rails about
55:14
uh about vaccines.
55:17
So one interesting thing about GROK here is
55:19
that it gives you the sources that it's
55:21
using as it thinks about things, and it
55:24
also takes a long time to think before
55:26
it starts to give you an answer, and
55:28
that's definitely different than ChatGPT or Google Gemini,
55:31
where it just like rattles off an answer
55:33
right away.
55:36
So I like that about Groc.
55:37
I like that it looks up a lot
55:39
of different sources, gives you the sources that
55:41
it's referencing before it gives you an answer
55:43
that it's thought about for a while.
55:45
So I asked it elaborate on the amount
55:48
spent per year, and it says about 489
55:51
million dollars in total bribes over about six
55:53
years.
55:56
So eighty to eighty-two million US dollars annually,
55:59
and I said it was a hundred million
56:00
Annually, and I said it was a hundred
56:01
annually.
56:02
million annually.
56:04
So, okay, I guess I was off by
56:04
So okay, I guess I was off by
56:07
eighteen million a year, but I stand by
56:07
18 million a year, but I stand by
56:11
the claim that having an eighty-two million dollar
56:11
the claim that having an eighty-two million dollar
56:14
per year bribery budget for one company in
56:14
per year bribery budget for one company in
56:17
one country is pretty extraordinary.
56:17
one country is pretty extraordinary.
56:21
And that is all the time we have
56:21
And that is all the time we have
56:22
for today on this episode of the nurses
56:22
for today on this episode of the Nurses
56:24
report.
56:24
Report.
56:26
I wish that we were getting paid that
56:26
I wish that we were getting paid that
56:27
uh the good guys had an eight.
56:28
uh the good guys had an $82 million
56:30
a year bribery budget.
56:32
I haven't made anything from doing these episodes
56:35
with you guys, but I have had a
56:37
great time and I hope that people learn
56:39
some things and enjoy listening to us scream
56:41
into the void.
56:43
Until next time, this has been nurses Nicole,
56:45
David, and Ashley.