Information Technology Reference
In-Depth Information
behavior to be observed even when no daily alerts
were generated (see Glascock and Kutzik, 2006).
go to the individual's residence to determine the
problem and fix ELMS, the system was essentially
out of action for close to a month. When someone
finally went to the residence he found that all
sensors and even the base station were infested
with cockroach feces. Likewise, after ELMS
abruptly stopped working in a second residence,
over three weeks elapsed before someone made
a special trip to the residence to determine the
reason—a cat had chewed through the base sta-
tion's power cord—and repair the problem. After
these and several other examples, a maintenance
protocol was developed that was included in all
future installations.
Lastly, we discovered that caregivers have their
own preferences as to how they want to receive
the information generated. We assumed that the
methods that we developed in the laboratory and
were vetted by our focus groups, would be appro-
priate for overworked visiting nurses who would
visit eight to ten clients a day. Once again, we
were wrong. Our protocol called for email alerts
to be sent to the nurses and that after receiving
the alerts the nurses would check the client's web
page for more detailed information. Our assump-
tion was that after reviewing all the information,
the nurse would then determine what action was
necessary. This assumption proved naïve. Each
nurse developed his/her own emic protocol and
wanted the information presented in a way that
conformed to this construct. One nurse did not
want alerts sent for a decline in meal activity for
a single day. In his opinion, this told him noth-
ing and just wasted his time. He only wanted an
alert for meals if there were three consecutive
days of abnormally low meal activity. One nurse
never checked the client's web page and instead
developed a protocol that required him to make a
phone call to each of his clients who had any red
alerts while a different care provider found the
web pages to be of great use and requested even
more detailed information on certain activities,
e.g., sleep patterns. This has led us to build an
ability to customize the way that information is
Stage 4: Visiting Nurse Care Model
After the field test, we were ready to test ELMS in
an environment in which care would be impacted
by the information that the system generated. Thus,
over an 18 month period, we installed ELMS in
the homes of 34 chronically ill individuals living
in their own residences in and around Philadelphia,
PA. Given the complexity of this environment,
both geographically—people lived miles apart,
and structurally—visiting nurses had to access and
respond to the information as part of their normal
work day, the noise and the resulting challenges
appeared, at times, insurmountable. However,
we persevered and learned many lessons, the
two most illustrative of which were: the need to
develop a procedure for regular maintenance of
the hardware; and that the information had to be
customized for each individual care provider.
The maintenance issue, in hindsight, appears
obvious, but except for replacing batteries in the
sensors every 12-18 months—the sensors would
even tell us when the batteries were low—we did
not anticipate the need for a procedure for main-
taining the hardware. We had not had problems
with the hardware in either the laboratory or the
more limited real world environments. We were
wrong as a combination of poverty, serious health
problems and unexpected living conditions con-
spired to drive home the point that if something
can go wrong, it will and therefore, a plan must
be in place to correct it. A trained individual must
be ready to respond when sensors or base stations
stop working; if not, the information generated by
ELMS will degrade to the point of uselessness.
Two brief examples illustrate the unexpected type
of problems/noise we encountered. In the residence
of one client, the sensors failed one by one over
a two week period, eventually generating invalid
information on all activities monitored. Since there
was no one “charged” with the responsibility to
Search WWH ::
Custom Search