The EconomistMarch 14th 2020 Technology Quarterly |Personalised medicine 132
1other parts of the body in which to model diseases and their cures.
Insitro is editing changes into stem cells—which can grow into
any other tissue—and tracking the tissues they grow into. By mea-
suring differences in the development of very well characterised
cells which differ in precisely known ways the company hopes to
build more accurate models of disease in living cells. All this work
is automated, and carried out on such a large scale that Dr Koller
anticipates collecting many petabytes of data before using mach-
ine learning to make sense of it. She hopes to create what Dr Scan-
nell complains biology lacks and what drug designers need: pre-
dictive models of how genetic changes drive functional changes.
There are also reasons to hope that the new “upstream” drugs—
asos, sirnas, perhaps even some gene therapies—might have ad-
vantages over today’s therapies when it comes to small-batch
manufacture. It may also prove possible to streamline much of the
testing that such drugs go through. Virus-based gene-therapy vec-
tors and antisense drugs are basically platforms from which to de-
liver little bits of sequence data. Within some constraints, a plat-
form already approved for carrying one message might be
fast-tracked through various safety tests when it carries another.
One more reason for optimism is that drugs developed around
a known molecule that marks out a disease—a molecular marker—
appear to be more successful in trials. The approval process for
cancer therapies aimed at the markers of specific mutations is of-
ten much shorter now than it used to be. Tagrisso (osimertinib), an
incredibly specialised drug, targets a mutation known to occur
only in patients already treated for lung cancer with an older drug.
Being able to specify the patients who stand to benefit with this de-
gree of accuracy allows trials to be smaller and quicker. Tagrisso
was approved less than two years and nine months after the first
dose was given to a patient.
With efforts to improve the validity of models of disease and
validate drug targets accurately gaining ground, Dr Scannell says
he is “sympathetic” to the proposal that, this time, scientific inno-
vation might improve productivity. Recent years have seen hints
that Eroom’s law is being bent, if not yet broken.
If pharmaceutical companies do not make good on the promise
of these new approaches then charities are likely to step in, as they
have with various asotreatments for inherited diseases. And they
will not be shackled to business models that see the purpose of
medicine as making drugs. The Gates Foundation and America’s
National Institutes of Health are investing $200m towards devel-
oping treatments based on rewriting genes that could be used to
tackle sickle-cell disease and hiv—treatments that have to meet
the proviso of being useful in poor-country clinics. Therapies in
which cells are taken out of the body, treated in some way and re-
turned might be the basis of a new sort of business, one based
around the ability to make small machines that treat individuals
by the bedside rather than factories which produce drugs in bulk.Run, rabbit, run
There is room in all this for individuals with vision; there is also
room for luck: Dr Nizar has both. Her problem lies in PTH1R, a hor-
mone receptor; her PTH1R gene makes a form of it which is
jammed in the “on” position. This means her cells are constantly
doing what they would normally do only if told to by the relevant
hormone. A few years ago she learned that a drug which might turn
the mutant receptor off (or at least down a bit) had already been
characterised—but had not seemed worth developing.
The rabbit, it is said, outruns the fox because the fox is merely
running for its dinner, while the rabbit is running for its life. Dr Ni-
zar’s incentives outstrip those of drug companies in a similar way.
By working with the fda, the nihand Massachusetts General Hos-
pital, Dr Nizar helped get a grant to make enough of the drug for
toxicology studies. She will take it herself, in the first human trial,
in about a year’s time. After that, if things go well, her children’s
pain may finally be eased. 7T
hreadworms aretrending, according to the app on Johannes
Schildt’s phone. The app was created by Kry, the Swedish digital
health-care firm Mr Schildt runs. It offers information on the sick-
nesses for which people are currently booking doctor’s appoint-
ments, as well as on things specifically important to its user—it
keeps Mr Schildt, who suffers from hay fever, up to date with the
pollen count. It lets him book an appointment with a family doctor
or a specialist, and indeed to have such an appointment by phone.
None of this sounds particularly stretching. But in health care, it
counts as radical.
According to the Organisation for Economic Co-operation and
Development (oecd), a club of richer nations, the world creates 2.5
exabytes of data a day—thousands of times what even the grandest
sequencing centre can produce in a month. Of those which get
stored, 30% pertain to health. The trove contains insights into the
health of populations and of individuals, the efficacy of drugs and
the efficiency of health-care systems, the failings of doctors and
the financial health of insurers. But oecd countries typically
spend less than 5% of their health budgets managing these data,
much less than is the norm in other areas. By failing to make the
most of their potential, these countries are wasting $600bn a
year—roughly the gdpof Sweden.
This underutilised resource has attracted the attention of a pa-
noply of private companies, from minnows like Kry to giants like
Amazon, Apple, Facebook and Google. Governments, hospitals
and insurers, they think, will pay for what they glean from it. So
will individuals—who will often pay for the privilege of supplyingThe coming of the datome
The way people live their lives can be mined, tooThe sum of all lives