77
science at the University of California, Davis.
He lives on contracting jobs, doing data sci-
ence and researching machine learning and
programming languages. Like many freelanc-
ers, he doesn’t have employer-sponsored in-
surance. But unlike most with the disease, he
knew enough about science to start thinking
there might be a better way—and so, in 2015,
he launched the Open Insulin Project.
“The current system was built to exploit
people with diseases,” Di Franco says. “Histo-
rians will look back and say, ‘How could they
have done such a terrible job?’ ”
Diabetes is causeD when the pancreas can
no longer make enough or any insulin, a hor-
mone that regulates blood sugar, or when a
person’s body builds up a resistance to the hor-
mone. Insulin, which helps the body use sugar
for energy and lowers its levels in the blood, was
first used to treat diabetes in the early 1920s.
Physician Frederick Banting and medical stu-
dent Charles Best were working with dogs, in-
ducing diabetes by removing their pancreases
and then trying to figure out a cure. The two
eventually extracted a substance from cow pan-
creases; purified it with the help of biochemist
James Collip; and proved it worked by injecting
it themselves and noting that they got dizzy, a
sign of low blood sugar. By 1922, doctors were
using insulin from cow pancreases to treat dia-
betes patients.
People with severe diabetes need insulin
injections to stay alive. Without it, your blood
turns acidic, your body dehydrates, your vision
blurs, you get weaker and start to vomit. Over
days, you slowly—and painfully—die.
This fact, coupled with the inefficiencies of
the American health system, as well as a manip-
ulable patent framework, has enabled pharma-
ceutical companies to steadily increase the price
of the lifesaving drug, even as it’s become easier
and less costly to produce. Generally speaking,
drugs are cheap to make. The costs are mostly to
pay for the research and development required
to discover them. For example, one 2016 study
that looked at 106 recently approved drugs from
10 different companies found that the average
R&D cost for each was $2.78 billion, compared
with only about $19 million per drug in costs of
actual clinical trials.
Much of the industrialized world has some
form of single-payer health insurance and
strict price controls on drugs, usually deter-
mined by a board of doctors and experts. In
the U.S., the pull of the free market was sup-
posed to keep prices down, but instead has
led to a complex system of profit-driven cor-
porations, from manufacturers to insurance
companies, who add cost at every juncture.
It wasn’t meant to be this way, especially
not with insulin. Banting, who shared a 1923
Nobel Prize for his work on insulin, demanded
his name not be put on the patent, believing
profiteering off a medicine was unethical. His
co-discoverers agreed, transferring their pat-
ents to the University of Toronto for $1 each.
The pharmaceutical corporation Eli Lilly
and Company of Indianapolis offered to help
the university develop the medication, and the
school eventually agreed to license the tech-
nology. Eli Lilly contracted with slaughter-
houses to receive pig pancreases by railroad
car in order to squeeze out the insulin. It was
crude, but effective—and cheap. Ads from the
1960s show vials of insulin available for 84¢ in
the U.S., just $7.36 in today’s dollars. And then
came a real breakthrough.
In 1982, Eli Lilly introduced insulin made by
genetically modified E. coli bacteria. The new
insulin was less likely to cause allergies than the
animal version, and it could be grown in vats.
Novo Nordisk started making its own bioengi-
neered insulin in 1991, and it looked like the
drug was about to get really affordable, thanks
to the competitive marketplace. Instead, prices
went up. A congressional report written in 2018
found the list price of competing insulin formu-
lations “appeared to rise in tandem,” doubling
from 2012 to 2018. According to the report,
that was most likely due to limited market com-
petition, and to the fact that “each part of the
insulin delivery chain is controlled by a small
number of entities.” The marketplace never be-
came competitive.
In theory, the U.S. patent system, which
gives manufacturers sole rights to a drug for-
mulation for 20 years, should eventually enable
other drug producers to bring cheaper versions
of the same medication to the market. But as
Reed and the co-chair of the Congressional Dia-
betes Caucus, Diana DeGette (D., Colo.), note,
companies skirt this by “evergreening” their
drugs—tweaking drug formulas slightly, often
making incremental improvements, to renew
the patent and prevent generics from ever en-
tering the market. Lantus, a long-acting insulin
patented by Sanofi in 1994, was due to enter
the public domain in 2015, but instead the com-
pany filed 74 patents for newer versions of the
drug, which delayed that until 2031. Novo Nor-
disk has done something similar with one type
of insulin by upgrading the mechanics of its
injection pen. These insulins are touted as im-
provements, although there is evidence these
are typically minimal.
Joanna
Shields
AI to read every
science paper
Every year, more than
2 million peer-reviewed
research papers are
published—far too many for
any individual scientist to
digest. Machines, however,
don’t share this human
limitation. BenevolentAI
has created algorithms
that scour research papers,
clinical trial results and
other sources of biomedical
information in search of
previously overlooked
relationships between
genes, drugs and disease.
BenevolentAI CEO Joanna
Shields was an executive
at companies such as
Google and Facebook, and
then the U.K.’s Minister
for Internet Safety and
Security, before joining
BenevolentAI. A frequent
critic of the tech industry’s
lapses in protecting young
people from exploitation
and abuse online, Shields
sees BenevolentAI as an
opportunity to harness
technology’s power for
good. “All of us have family
members, friends who are
diagnosed with diseases
that have no treatment,”
she says. “Unless we
apply the scaling and the
principles of the technology
revolution to drug discovery
and development, we’re
not going to see a change
in that outcome anytime
soon.” ÑCorinne Purtill