The accuracy of results and conclusions of studies financed by drug companies.

The accuracy of results and conclusions of studies financed by drug companies.

Some drug studies more likely to have favorable conclusions

* w w w .eurekalert.org/pub releases/2007-11/bmj-sds111607.php

Financial ties and concordance between results and conclusions in meta-
analyses: retrospective cohort

Previous work has shown that, when a drug study was funded by the
company that made that drug, the results might be biased in favour of
that drug because the methods or analyses were manipulated.

New research published on bmj . com today shows that, for blood pressure
drugs, studies are now much less likely to have biased results but
still tend to have overly positive conclusions favouring the company's
products.

The authors call on editors and peer reviewers to scrutinise the
conclusions of these studies to ensure that they contain an unbiased
interpretation of the results.

Meta-analyses represent the highest level of research evidence in the
hierarchy of study types. They pool data from multiple studies to
provide summary statistics on the effectiveness of a given treatment.
They have a great deal of influence on patient care and healthcare
policy and drug companies have started to reference meta-analyses in
their advertisements.

Previous studies have shown that randomised controlled trials with
financial ties to single drug companies are more likely to have
results and conclusions that favour the sponsor's products, and a
recent study suggests that the same holds true for meta-analyses.

So researchers in the US set out to determine whether financial ties
with single drug companies are associated with favourable results or
conclusions in meta-analyses on blood pressure lowering
(antihypertensive) therapies.

A total of 124 meta-analyses were included in the study, 49 (40%) of
which had single drug company financial ties. Differences in study
design and quality were measured.

Meta-analyses with single drug company financial ties were not
associated with favourable results but were significantly more likely
to have favourable conclusions, even when differences in study quality
were taken into account.

In fact, the data show that studies funded by a single drug company
have a 55% rate of favourable results that is transformed into a 92%
rate for favourable conclusions, representing a 37% gap. The gap
shrinks to 21% (57% to 79%) when two or more drug companies provide
support. Yet the gap vanishes entirely for studies done by non-profit
institutions alone or even in conjunction with drug companies.

These findings suggest a disconnect between the data that underlie the
results and the interpretation or "spin" of these data that
constitutes the conclusions, say the authors.

The findings also expose a failure of peer review, add the authors,
and should act as a wake-up call to editors and peer reviewers, as
well as to policy-makers, meta-analysts, and readers. All of these
groups should closely scrutinise the conclusions of meta-analyses to
ensure that they contain an unbiased interpretation of results, they
conclude.

The clear inference from this study is that impartial studies are more
reliable, say researchers in an accompanying editorial. However,
rather than imposing legal restrictions on drug company funding or
participation in these studies, they suggest that doctors should be
warned to be cautious in interpreting the conclusions of studies


Thanks Vince

Re: The accuracy of results and conclusions of studies financed by drug companies.

* w w w .bmj . com /cgi/content/full/bmj.39376.447211.BEv1

* w w w .bmj . com /cgi/content/extract/bmj.39381.655845.BEv1

Drug Company Funding Found to Bias Study Conclusions, Not Results

By Judith Groch, Senior Writer, MedPage Today
Reviewed by Zalman S. Agus, MD; Emeritus Professor
University of Pennsylvania School of Medicine.
November 16, 2007

STANFORD, Calif., Nov. 16 -- Hypertension drug studies with financial
ties to a single drug company were not more likely to present
favorable results, but they were significantly more likely to have
favorable conclusions, researchers here said. Action Points

There was a 37% gap between results and conclusions, suggesting a
disconnect between the data that underlie the results and the
interpretation or "spin" of these data that constitute the
conclusions, Veronica Yank, M.D., of Stanford University, and
colleagues wrote in BMJ Online First.

In contrast, they said, meta-analyses with financial ties to nonprofit
groups had excellent agreement between results and conclusions.

The findings came from a retrospective cohort study of 124 meta-
analyses published before December 2004 that evaluated the effects of
an antihypertensive drug compared with any comparator on clinical end
points in adults.

Financial ties in the study were categorized as being from one drug
company or from all other sources. Information on financial ties came
from disclosures in the meta-analysis itself; disclosures of industry
or other sponsorship in the journal supplement in which the study was
published; or disclosures of financial ties in previous research on
the drugs by the first author, going back three years.

The researchers found that 49 (40%) of the meta-analyses had financial
ties to a single drug company.

On logistic regression analysis, meta-analyses with better
methodological quality that evaluated studies' heterogeneity and did
sensitivity analyses were more likely to have favorable results (odds
ratio: 1.16, 95% confidence interval: 1.07 to 1.27). However, only
slightly more than half of the meta-analyses reported favorable
findings.

Studies that had financial ties to one drug company, even when
controlling for the quality of the meta-analysis, had the worst
concordance between results and conclusions, with 55% (27 of 49)
having favorable results, but 92% (45 of 49) having favorable
conclusions, representing a 37% gap, the investigators found.

The gap narrowed to 21% (57% to 78%) when two or more drug companies
provided support. It vanished entirely for studies done by nonprofit
institutions alone or even with some drug company funding.

The researchers also collected data on characteristics of meta-
analyses that might be associated with favorable results or
conclusions. Studies of better quality were associated with favorable
results, they said.

Although financial ties to one drug company were not overall
associated with favorable results, such ties constituted the only
characteristic significantly associated with favorable conclusions
(odds ratio: 4.09, 95% CI: 1.30 to 12.83).

For example, when controlling for other characteristics of meta-
analyses, studies that had financial ties to one drug company remained
five times more likely to report favorable conclusions (OR: 5.11, 95%
CI: 1.54 to 16.92).

Because investigators used conservative assumptions in defining
financial ties, the odds ratio for the main finding is likely to be an
underestimate of the true relation between financial ties to one drug
company and favorable conclusions, the researchers said.

They noted that the study had a potential for confounding, but said
they were able to make adjustments.

Another methodological limitation is that only one researcher, Dr.
Yank, reviewed the meta-analyses both for inclusion in the study and
for data and quality assessment. Also, the reviewer was not blinded to
important characteristics of the studies, including financial ties,
the investigators noted.

The generalizability of this study is limited by its restriction to
one clinical topic. However, Dr. Yank said, these findings have
relevance to the real world, as the marketing of antihypertensive
drugs constitutes a multibillion-dollar-a-year industry and these
drugs are some of the most prescribed in the world.

These findings also expose a failure of peer review, the investigators
said. Both editors and peer reviewers must have read manuscript
versions of those meta-analyses containing discordant results and
conclusions, yet they did not prevent publication of biased
conclusions.

Editors and peer reviewers, as well as policymakers, meta-analysts,
and readers should closely scrutinize the conclusions of meta-analyses
to ensure that they are supported by the data, Dr. Yank and her
colleagues concluded.

In an accompanying editorial Richard Epstein, L.L.B., of the
University of Chicago, said the key findings of this study are robust
and will likely draw the ire of the many critics of the drug
industry.

In terms of explaining the results, Epstein said, the study was too
small to identify the characteristics beyond financial ties that might
be involved. For example: How much direct control does the drug
company exercise over the study? Do its own doctors participate?

These questions matter, he said, because any bias in a meta-analysis
is unlikely to disappear in ordinary clinical trials where company
experts commonly team with outside experts.

We face a dilemma, he said: Do we want fewer studies of presumably
better quality, or do we want more studies with possibly biased
quality? "I would opt for the last option. Nothing in the work of Dr.
Yank's team suggests that the raw data were defective," he wrote.

As long as the disagreements lie in the interpretation of data and not
its collection, the solution is not state regulation, he said. Rather,
doctors should be warned to be cautious in interpreting the
conclusions of studies.

The study was funded in part by the Eugene Garfield Foundation.

The study authors reported no financial conflicts. Dr. Yank received
support from a dean's quarterly research grant (University of
California at San Francisco) and from the internal medicine residency
program (University of Washington, Seattle).

The editorialist, a law professor, wrote that he has worked as a
consultant to drug companies for many years over a wide range of
issues involving liability, parallel importation, and FDA matters. He
has not worked for a drug company on conflicts of interest questions
of the sort discussed here. The Institute for Policy Information,
which has ties to the drug industry, provided financial support for a
book he recently wrote.

Re: The accuracy of results and conclusions of studies financed by drug companies.

From * w w w .bmj . com /cgi/content/full/bmj.39376.447211.BEv1

"Conclusions
That we found poor concordance between results and conclusions in some
meta-analyses of antihypertensive drugs suggests that meta-analyses,
as with other study types, are open to the influence of systematic
bias, in this case by having financial ties to one drug company. Our
study also exposes a failure of peer review. Both editors and peer
reviewers must have read manuscript versions of those meta-analyses
containing discordant results and conclusions, yet they did not
prevent publication of biased conclusions. Editors and peer reviewers,
as well as policymakers, meta-analysts, and readers should closely
scrutinise the conclusions of meta-analyses to ensure that they are
supported by the data.




What is already known on this topic

The results and conclusions of randomised controlled trials with
financial ties to one drug company are more likely to favour the
sponsor's products

A study that compared Cochrane meta-analyses to industry supported
meta-analyses in print journals suggests that the same holds true for
meta-analyses

What this study adds


Meta-analyses with financial ties to one drug company are no more
likely than others to have results that favour the company's drugs but
are more likely to have favourable conclusions

Editors and peer reviewers failed to prevent publication of biased
conclusions in meta-analyses "

The original post noted that "Previous work has shown that, when a
drug study was funded by the
company that made that drug, the results might be biased in favour of
that drug because the methods or analyses were manipulated. "

Thanks Vince