64 Science & technology The EconomistFebruary 29th 2020
21the University of Leuven are working on.
An ultrasound scanner is a device that em-
ploys high-frequency sound waves to
create images of parts of the inside of the
body. The idea, says Dr d’hooge, is that both
the transmitters which produce the ultra-
sonic pulses and the receivers which pick
up the returning echoes can be woven into
textiles used to make items of clothing,
such as vests. He is optimistic that it will
thus be possible to create garments capable
of conducting heart scans, and that these
might be washable.
All this would help cardiologists like Dr
Chico a lot. It is often hard for patients to
describe their symptoms fully, and hospi-
tal tests might not reveal a complete pic-
ture—especially as people tend to be under
stress when those examinations are car-
ried out. To start with, the data used to
model and update a digital heart-twin will
be recorded by the collection device and
uploaded therefrom at intervals. Eventual-
ly, though, it should be possible for them to
be transmitted directly to a medical centre,
just as data from a jet engine are transmit-
ted to an engineering base.
While wearable heart scanners are sev-
eral years away, some elements needed to
build digital heart-twins are close to de-
ployment. Rod Hose, a former aerospace
engineer who is now an expert in medical
modelling at the University of Sheffield,
led a recent project called EurValve, which
developed a system to help doctors treat
people with heart-valve disease. EurValve
gathered a variety of data about patients’
conditions from scans and other hospital
tests, and combined these with other infor-
mation acquired from those patients when
they were at home, via health-tracking
watches produced by Philips, a Dutch tech-
nology group. The EurValve system, which
the researchers hope will soon be put into
clinical practice, can model the severity of
disease and predict the outcome of heart-
valve-replacement surgery.
A digital twin of the whole heart will al-
low simulation of the treatment of a partic-
ular individual for many other conditions,
as well. That will give a clearer idea, in a
particular case, of the likely outcome of an
intervention. It might show, for instance,
what type of operation is best suited to a
patient’s condition, or if drugs and regular
check ups are more appropriate.
As more and more patients have their
heart twins analysed, machine learning, a
form of artificial intelligence that is good at
pattern recognition, will be used to study
the outputs. This should make the system
yet more accurate, and help with unusual
and rare cases that a cardiologist might not
have seen before. Just as pilots can relax
knowing that a digital twin is keeping an
eye on their engines, doctors will benefit
from a new depth of knowledge about how
their patients’ hearts are working. 7I
t goes withoutsaying that to be a ma-
rine you have to be tough, both physically
and mentally. But which is more impor-
tant? And, more specifically, which is the
bigger obstacle to successful training?
Working on behalf of America’s marine
corps, Leslie Saxon, of the Keck School of
Medicine of the University of Southern
California, has been trying to find out.
The elite of the marine corps is a group
called Force Reconnaissance. These troops
are employed in special operations, both
“green” (in which having had to engage the
enemy is deemed a failure) and “black”
(where such engagement is the whole
point). Initial training to join the force,
open only to those already marines or naval
doctors, lasts 25 days. Among other things
it requires volunteers to tread water for
nearly an hour, to run eight miles (12km)
while carrying more than 50lb (about 23kg)
of equipment, and to swim 100 yards (90
metres) with their hands and feet bound.
Only half of those who volunteer for this
training complete it. Of those who do not,
roughly half are failed by the judges for
posing a safety risk or for having a medical
problem that stops them completing the
course. The other half, though, drop out of
their own volition.
That high drop-out rate is both expen-
sive and vexing for Force Reconnaissance’s
recruiters. They therefore turned to Dr Sax-on to find out what is happening, so that
they can take steps to reduce the losses.
To gather the relevant data she picked
121 trainees and provided each of them with
two devices: an iPhone and an Apple Watch
(a wrist band that both tells the time like a
conventional watch and watches what the
wearer gets up to). She loaded the phones
with an app that asked participants a range
of demographic and psychological ques-
tions at the start of their training. From the
answers to these she generated, for each
volunteer, scores for the five main perso-
nality traits recognised by psychologists—
openness, conscientiousness, extrover-
sion, agreeableness and neuroticism—and
also for ego resilience (ability to control an-
ger and to control impulses when
stressed), positive affect (a person’s ten-
dency to experience positive emotions
when facing challenges), satisfaction with
life and level of psychopathy. The watch,
meanwhile, monitored the number of
steps its wearer took, and kept track of both
heart rate and calorie expenditure.
Once volunteers began training they re-
ceived further, daily questionnaires on
their phones. They were asked to rate their
pain, both mental and physical, on a scale
of one to five. They were asked if they
thought of quitting and if they thought
their instructors wanted them to graduate.
They were also asked about their sleep,
their hydration, their nutrition and their
own confidence that they would graduate.
Dr Saxon’s sample proved pretty repre-
sentative in their rates of completion of the
course. As she reports in the Journal of Med-
ical Internet Research, 56% were successful,
23% dropped out of their own volition and
21% were removed for a mixture of medi-
cal, safety and performance reasons. Ana-
lysing the data for those who dropped out,How to work out beforehand who will
pass advanced military trainingArmed forcesWar on attrition
Just going in for a quick dip