1206 3 DECEMBER 2021 • VOL 374 ISSUE 6572
PHOTO: TRISTAN SPINSKI/THE NEW YORK TIMES/REDUXscience.org SCIENCEINSIGHTS | POLICY FORUM
ity to determine precisely what evidence will
be adequate for approval and to review the
details of that evidence helps ensure both
quality and confidence beyond what would
be possible through market forces alone ( 10 ).
However, once FDA approves a drug,
companies begin to profit, and patient de-
mand builds, company incentives to pur-
sue further rigorous study drop off precipi-
tously—future trials might fail to confirm
benefit and companies already have what
they want. If followed, FDA’s guidance rec-
ommending that confirmatory
trials be underway at the time
of accelerated approval could
help address this problem ( 11 ),
but the agency often makes
exceptions. FDA also does not
necessarily require the param-
eters of confirmatory trials to
be agreed upon as a condition of
accelerated approval ( 6 ). Both
types of laxity stem from FDA’s
reticence to further hold up
promising products while post-
market study details are worked
out ( 6 ), but they give companies
leverage to “bargain down” after
approval. Company incentives
are further weakened by knowl-
edge that even with problematic
study designs and results, FDA
may allow accelerated approv-
als to stand ( 6 , 8 ).
On this explanation, early
approval pathways do not in-
herently sacrifice rapid, con-
firmatory evidence—FDA and
Congress merely have to insist upon it.
An appropriate policy solution would en-
tail making early approvals time-limited
with automatic withdrawal for failure to
confirm benefit by a set deadline ( 3 ). The
EMA is on the right track with its require-
ment for periodic renewal of conditional
approvals, although that is watered down
by not limiting the number of permissible
renewals. A true expiration date, such as
that reflected in a new provisional approval
authority proposed in a bill dubbed the
Promising Pathway Act (PPA, S.1644) cur-
rently pending in Congress would help ad-
dress challenges to FDA’s willingness to act
on evidence once produced by reducing its
discretion. However, for the threat of with-
drawal to be meaningful, required confir-
matory evidence would also have to be truly
confirmatory, generated through trials with
appropriate randomization, blinding, and
controls, measuring meaningful health out-
comes. Unfortunately, even if enacted, the
PPA would not meet this standard, instead
proposing reliance on “real world evidence”
(RWE) and registries.
In recent years, RWE, including obser-
vational clinical data produced outside
trials, has increasingly been valorized as a
potential solution to some of the tension
between access and evidence. So far, how-
ever, reality has not lived up to the hype,
leading ICER to conclude that, at least at
present, RWE likely will not suffice “in
smaller patient populations with greater
uncertainty surrounding the studied end-
points” ( 3 ). RWE can likely help comple-
ment rigorous confirmatory trials, but fordiseases with a highly variable course of
progression or drugs anticipated to have
only modest effect, concurrent, random-
ized control groups are critical to deter-
mining whether observed differences can
be causally attributed to the intervention.It’s the patients
Companies also may fail to produce rigor-
ous postapproval trials because not enough
patients are willing to participate in them.
Once a product receives marketing ap-
proval, many patients facing serious disease
are likely to want to try it, especially when
they lack other options. If their clinicians
will prescribe and insurers will pay, patients
may have insufficient reason to enroll in tri-
als in which they may be randomized to
something else ( 2 , 3 ). This is why some ini-
tially wondered if further randomized trials
of Aduhelm might be feasible only outside
the US, despite major open questions.
Although typical ethical concerns about
withholding proven products from trial
participants can be avoided when benefit
is truly uncertain, exploitation concernsremain at the prospect of allowing US pa-
tients to reap the potential benefits of early
access while foisting the burdens of con-
firmatory trials onto patients elsewhere,
especially for high-cost drugs. Patients and
contexts abroad may also differ in mean-
ingful ways from those in the US, poten-
tially inhibiting generalizability of inter-
national trial data. For example, there may
be differences in disease stage at diagnosis,
standard care, and environmental factors,
not all of which can be fully addressed
through enrollment criteria.
The alternative to international
trial sites is to gather confirma-
tory evidence in the US using
weaker designs more amenable to
recruitment, such as without ran-
domization or use of placebo, as
has been common for accelerated
approval drugs in the past. It may
also be possible to make rigorous
trials more attractive to patients,
for example, by minimizing the
proportion randomized to the
control group either at the out-
set or through response-adaptive
randomization or using crossover
designs that provide all partici-
pants the opportunity to receive
the intervention even if initially
randomized to the control group.
Access to the intervention for free
as part of a trial may help entice
patients who would otherwise be
unable to afford it, as could of-
fering other financial incentives.
However, it is not clear that these
approaches would be successful, especially
for patients facing fast-moving, terminal
conditions who may seek any means neces-
sary to access an intervention FDA has des-
ignated as potentially beneficial.
If patient willingness is the problem fac-
ing early approval pathways, no amount of
insistence on companies producing rigorous
confirmatory trials will help. In this case,
one policy response could be to require con-
firmation of benefit before allowing a drug
to be marketed—thereby rejecting early ap-
proval, and its upsides—while simultane-
ously supporting preapproval access though
broader inclusion in clinical trials and im-
proved use of FDA’s expanded access path-
way. Expanded access (“compassionate use”)
allows individuals or groups of patients with
serious or life-threatening conditions to ac-
cess unapproved drugs outside trials when
they lack alternatives and such access will
not interfere with a product’s testing and de-
velopment. As currently conceived, however,
expanded access serves far fewer patients
than marketing approval, owing to a vari-
ety of logistical hurdles, including clinicianA mother moves her son, who has Duchenne muscular dystrophy,
out of bed. Accelerated approval was used to bring a Duchenne
muscular dystrophy drug, Exoyndys 51, to market in 2016. Confirmatory
trials are ongoing, and insurers have balked at coverage.