SCIENCE sciencemag.org 17 JANUARY 2020 • VOL 367 ISSUE 6475 243
more than 2600 trials for which results are
overdue or were filed late, NIH has yet to
withhold a single grant as a result or post a
single violation notice on ClinicalTrials.gov.
No “wall of shame” exists.
“Public-facing websites run by the gov-
ernment should be accurate. That’s not ask-
ing much,” Senator Chuck Grassley (R–IA),
who advocated for the 2007 law, wrote in
an email after reviewing a summary of the
Science findings. “It’s a question of basic
management and agency competence. The
government has a duty to police its work
product, especially because the public trusts
.gov websites will be accurate and reliable.”
To physician Ben Goldacre, who directs
the Oxford program behind TrialsTracker,
“The lack of urgency is really troubling.”
IN A RECENT ARTICLE in The BMJ, Goldacre
and colleagues highlighted a long-running
MD Anderson trial as an example of what’s
at stake when clinical research results go
AWOL. Started in 1999, the trial tested a
specialized hormone therapy in patients
who had surgery for prostate cancer and
faced a high risk of recurrence. The treat-
ment might reduce that risk significantly,
but other trials suggested it had serious side
effects, including reduced bone density, sex-
ual dysfunction, and greater risk of diabetes
and cardiovascular events. When Science
collected the ClinicalTrials.gov data, the MD
Anderson trial’s results were overdue by
1 year and 8 months without explanation,
and no journal appears to have published
them. Doctors and prostate cancer patients
weighing the most appropriate treatment
have been left in the dark.
Stephen Hahn, who served as chief medi-
cal officer at MD Anderson until last month,
when he became the new FDA commis-
sioner, was unavailable for comment. An
MD Anderson spokesperson said the center
“believes in transparency,” and is making
“every effort to comply” with trial reporting
rules, but did not respond to a question on
the missing cancer trial data.
Mayo, Yale, the University of Minnesota,
Baylor, and Boston Children’s, which have
similarly poor reporting records, all said
via email that they, too, were committed to
fulfilling ClinicalTrials.gov requirements.
Mayo said it deployed dedicated staff to as-
sist researchers. Yale noted it had registered
hundreds of trials exempt from reporting
requirements and added a layer of review
to help ensure compliance. The University
of Minnesota recently created a monitor-
ing system and has made rapid progress
on reporting, according to a spokesperson.
Baylor said it planned to centralize trial
monitoring, as “a top priority.” Boston Chil-
dren’s said it was “committed to achieving
100% compliance.” It submitted data for one
long-overdue study after being contacted by
Science, but four others remain in apparent
violation as of the end of 2019.
The slow, apparently forgiving regula-
tory approach favored by NIH and FDA
will never force such organizations into
full compliance, some advocates of clini-
cal trial transparency say. “In an era when
every restaurant is obliged to publish onits front door the hygiene rating in their
kitchen, we’re seriously not saying whether
a trial that costs millions of dollars has bro-
ken the law and its obligation to ... patient
participants by failing to report its results?”
Goldacre asks. “That seems like extraordi-
nary special treatment for clinical trialists.”
Select institutions have made serious ef-
forts to comply. Twenty big pharma compa-
nies met all reporting requirements under
the 2017 rule and some major academic cen-
ters improved sharply compared with data
collected in 2017 (as detailed in the second
STAT investigation). Memorial Sloan Ket-
tering Cancer Center, Duke University, and
Johns Hopkins University—poor perform-
ers in 2017—complied with the law in nearly
all their registered trials covered by the new
rule. Johns Hopkins added staff to track and
assist on reporting and to identify “problem
records.” And it enlisted university execu-
tives to crack down on recalcitrant inves-
tigators, according to Anthony Keyes, a
clinical research manager there.
But such good performance shouldn’t be
an exception, Harvard’s Zarin says. “Fur-
ther public accountability of the trialists,
but also our government organizations, has
to happen. One possibility is that FDA and
NIH will be shamed into enforcing the law.
Another possibility is that sponsors will be
shamed into doing a better job. A third pos-
sibility is that ClinicalTrials.gov will never
fully achieve its vital aspirations.” jThis story was supported by the Science Fund
for Investigative Reporting.C
linicalTrials.gov relies on the organiza-
tions that fund or run clinical trials to
register studies and report accurate
results on time. Many miss the legal
reporting deadlines (see main story,
p. 240). But even those that comply often
do so with substandard effort. They face
no penalty: Sponsors are credited with
fulfilling their legal obligations on the date
they file the faulty data.
Deborah Zarin, who headed the
National Institutes of Health data site
from 2005 to 2018, recently reported
in JAMA Internal Medicine that typical
lapses in reported results included show-
ing data on more participants than were
enrolled and providing inconsistent units
of measurement—rendering the results
incoherent. ClinicalTrials.gov analysts
reject such results until they pass basicquality-control standards—a process than
can take many months. Science has found
that one of every seven trials credited with
reporting results shows no posted data as
a result of quality-control reviews. Results
for many other trials were posted only
after multiple rejections.
Similar quality issues often delay an
earlier step: registering a trial at the outset.
According to Zarin, ClinicalTrials.gov has
historically rejected half of all initial regis-
trations because of a basic flaw: unclear
primary outcome measures—what a trial
seeks to examine. Often, rejections stem
from “sloppiness or lack of rigor,” she says.
But she has heard of companies that want
to appear to comply with ClinicalTrials.gov’s
legal mandate but deliberately file shoddy
trial registrations—likely to be rejected
and therefore delay public posting—to
temporarily protect information they
regard as proprietary.
Jennifer Miller, an ethicist at YaleUniversity who created the Good Pharma
Scorecard to rank drug companies on
ethics and transparency, says some trial
sponsors go even further. In a recent
study in The BMJ, Miller and colleagues
examined 19 drugs approved by the U.S.
Food and Drug Administration (FDA) in
2015, all from large companies. In only
10 cases were all the trials provided
to FDA for its decisions registered on
ClinicalTrials.gov. When studies aren’t
registered, a failure to report results is
impossible to detect. “It’s easy to game
the system,” Miller says. “Just don’t regis-
ter your trial.”
“Minimal compliance with the law
should be a given,” she adds. But trial
sponsors should do more than just go
through the motions of registering their
studies and posting results, Miller says.
Evidence-based medicine demands com-
plete and careful reporting: “Ethics is not
minimal compliance with the law.” —C. P.Gaming the system
Published by AAAS