speed up and skin temperatures would lift –
up to three days before any symptoms were
physically felt. Snyder’s team is currently
working on an algorithm to help alert
wearers that they’re ailing.
Meanwhile, researchers at Carnegie Mellon
University in Pittsburgh are working on an
algorithm to alert hospital medics long in
advance of a patient hitting a critical ‘code
blue’ emergency. Using data gleaned from
133,000 patients, who had hospital treatment
between 2006 and 2011, their system can
currently correctly guess (two thirds of the
time) which patients will experience heart
attacks and respiratory arrests, sometimes four
hours before they happen. They’re currently
refining their algorithm, trying to decrease its
current false positive rate of 20 per cent.
Yet more impressive work is being carried
out by Bollen – a project that seeks to use
data from social media to be able to predict
when individuals might be at risk of suffering
a mental health crisis.
“I’m very excited about this,” says Bollen.
“Most of us have at least six or seven years of
history, either on Twitter or Facebook, that
provides an almost daily timeline of how we’re
feeling. And we’re looking to use that to
predict changes in your mental health, for
example, dropping into a major depression.”
Their theory is that there’s a common
pattern by which people’s mental health fails.
One signature of this pattern is a staggered
lowering of resilience against life’s routine
stresses. “Because resilience is lowered,
they’re not capable of bouncing back quite
as quickly from all sorts of perturbations. It
gets into a feedback loop that’s very difficult
to get out of. If an algorithm can spot that
pattern, it could potentially raise the alarm.”
Of course, this too comes with its
dangers. Do we want these deeply personal
observations about ourselves being made
by some tech company and held on remote
servers? Do we want employers possibly
gaining access to all this information aboutpersonal mental and physical health? Our
health insurers? The government? Could
hackers make it all public?
“We need very strong safeguards to protect
individuals,” concedes Bollen. “This will
require some serious legal changes.” But the
upsides could be revolutionary. “There’s no
doubt that within 10 or 15 years, instead of
seeing a doctor, you’ll have a computer
algorithm that can read the entire medical
literature in a matter of seconds and that also
knows more about your second-by-second
state of health than any doctor could possibly
gather from a visit to their office.”state
It’s not only healthcare professionals and
corporate bodies such as insurers and Google
that desire access to this great global brain.
Governments do too. We look to the state
to protect us, be it from crime, terrorism