A
new federal agency—approved last month by the
United States Congress—is already off to a rocky
start. The Advanced Research Projects Agency for
Health (ARPA-H), proposed by President Biden in
2021, aims to tackle the most intractable biomedi-
cal problems by funding innovative, high-risk,
high-reward research and swiftly turning discov-
eries into treatments and cures. But Congress gave the
agency a much smaller budget than sought by the admin-
istration—$1 billion over 3 years, a fraction of the $6.
billion requested. And as happens whenever there is new
money and a new federal agency, a political scrum has
erupted over who should control ARPA-H. It is now ex-
pected to answer to both the National Institutes of Health
(NIH) and the Department of
Health and Human Services (HHS).
If it is to deliver on its mission,
ARPA-H needs to be an autonomous
entity that approaches biomedical
research in a way never done be-
fore by the federal government. The
stakes are high: If ARPA-H fails to
produce new clinical advances rela-
tively quickly, it will erode trust in
US science. It’s time for clear think-
ing and action about what it will
take to make ARPA-H successful.
ARPA-H is modeled on the De-
fense Advanced Projects Agency
(DARPA), an independent agency
that reports to the US Department
of Defense and whose goal is to
fund breakthrough technologies
for national security. For more than 60 years, DARPA
has successfully accelerated innovation to application.
The question is whether ARPA-H can follow this model
because biomedical research is so different from the
work supported by DARPA. In biomedical research, lots
of things work in vitro, some things work in cells, fewer
work in mice, and only a precious few actually work
in humans. There is also a huge gap between basic re-
search and immediate application, known as the “valley
of death,” where innovations get stuck. It will be a ma-
jor feat for ARPA-H to bridge this chasm, particularly
because the ecosystem in which it has been administra-
tively placed—the NIH—is not one with a track record
for driving this kind of research directly.
The NIH has been wildly successful at funding basic
research, and although this has led to important clini-
cal advances, these have almost always been carried
forward by the private sector. The National Center forAdvancing Translational Science was launched at NIH
a decade ago to improve the bench-to-bedside process.
It has had some success in fostering application of re-
purposed drugs, but no obvious de novo medicines or
technologies have come from the division. Nevertheless,
Francis Collins, the former NIH director and current
White House science adviser, and other officials, includ-
ing current NIH interim director Lawrence Tabak, have
argued that ARPA-H should be embedded within NIH
and controlled by the NIH director. Their rationale is
twofold: The aims of ARPA-H fall squarely within the
NIH’s mission to improve human health, and ARPA-H
needs access to NIH biomedical expertise. Meanwhile,
critics have said that ARPA-H’s mission begins where
NIH’s ends. “Everyone I know
from NIH gets very excited about
science, which is an input, but
not the output, not the objective,
of an ARPA,” said former DARPA
director Arati Prabhakar. Prab-
hakar is correct—basic science is
exciting in a way that is absent
from the sometimes tedious and
scientifically conservative process
of shepherding drugs from the lab
to approval.
Xavier Becerra, the secre-
tary of HHS, decided to place
ARPA-H under the auspices of the
NIH, but the director would report
to Becerra, not the NIH director.
So the ARPA-H director will have
to answer to one boss but get ad-
ministrative support from another. And it is unlikely
that Becerra and NIH leadership will agree: Collins and
Tabak want to fully absorb ARPA-H, but Becerra told
Congress that NIH’s role was merely to be the back of-
fice, providing human resources, payroll, and informa-
tion technology. Indeed, Becerra has said that ARPA-H
will be in a different physical location from NIH, which
should help establish the new agency’s independence.
President Biden has touted ARPA-H as a way to tackle
Alzheimer’s disease, cancer, and diabetes. It’s hard to
name three more difficult problems in biomedicine. It
probably would have been better to take all of the new
money and put it into basic research, which has led to
new translational advances. But that opportunity has
passed and it’s time to stop arguing over who controls
ARPA-H and do what it takes to make it successful.–H. Holden ThorpWill ARPA-H work?
H. Holden Thorp
Editor-in-Chief,
Science journals.
[email protected];
@hholdenthorp10.1126/science.abq
PHOTO: CAMERON DAVIDSON
SCIENCE science.org 15 APRIL 2022 • VOL 376 ISSUE 6590 223EDITORIAL
“If ARPA-H fails
to produce
new clinical
advances relatively
quickly, it will
erode trust
in US science.”