IEEE Spectrum July, 2014 - 15

smArtphone App
Keeps WAtch over
schizophrenic
pAtients
mads Perch/getty images

Monitoring behavioral patterns can predict a relapse
What if a schizophrenic patient could have the equivalent of a therapist
in a pocket, watching for symptoms of a relapse? That's the promise of
a smartphone-based system now being tested at a hospital in Glen Oaks,
N.Y. The app, called CrossCheck, uses a suite of sensors to create a profile
of a patient's healthy behavioral and social patterns and can then raise an alert when
the patient deviates from the norm.
For a schizophrenic patient, a relapse is both damaging and demoralizing, says Dror
Ben-Zeev, an assistant professor of psychiatry at Dartmouth College and the study's
principal investigator. The patient can end up in jail or in the hospital, which disrupts
therapies and normal routines.

The CrossCheck study will include 150
schizophrenic patients who have been discharged from Zucker Hillside Hospital, in
Glen Oaks, within the past year. Half of
those patients will receive smartphones
with the app, while the other half, the control group, will receive the standard clinical services that the hospital provides after
a patient is discharged. To see if the app
leads to better outcomes, "we'll be looking at the time lag between when they're
recruited and their first relapse, and also
at the number of relapses they have over
a year," Ben-Zeev explains.
Mobile health apps have the potential to
change the management of chronic illnesses
in general and may be particularly valuable
for mental illnesses, says John Kane, chairman of psychiatry at the hospital. Kane says
his patients come to the clinic about once a
month, and in between those visits he has
no way of knowing if they're sliding back
into psychosis. "When they're having the
relapse, they might not have the insight to
call us and say, 'I think I'm becoming ill
again,' " Kane says. "This technology lets us
observe the patient's condition in real time."
CrossCheck ties together a number of data
sets to create patient profiles. It uses GPS to
create a map of patients' typical locations,
and accelerometer data to determine when
patients are walking, running, or sedentary.
It uses the microphone to detect conversations that occur either over the phone or in
person, and it records the duration and frequency of conversations (it doesn't record
or analyze content). Finally, to discern sleep
patterns, CrossCheck looks for times when
the phone is stationary and not in use, and
when light and sound sensors determine
that the environment is dark and quiet. All
this behavioral monitoring occurs in the
background as patients go about their lives.
The patient's only active participation
is filling out a brief questionnaire once a
week, which asks about mood and symptoms. "When they indicate that things are
going poorly, that they're feeling bad or
not getting enough sleep, we mark that as
a near-relapse event," explains Ben-Zeev.
CrossCheck looks for behavioral markers

nEwS

SPECTRUM.IEEE.ORG

nORTh aMERICan

JUL 2014

http://SPECTRUM.IEEE.ORG

Table of Contents for the Digital Edition of IEEE Spectrum July, 2014