IEEE Technology and Society Magazine - Spring 2013 - 42

Depending on the local regulatory environment on healthcare,
an e-health solution may attract
regulatory scrutiny or be subject
to legislative constraints. This
may come from the perspective
of health and safety, privacy, or
scope of services, and will have
to be addressed on a case-by-case
basis in accordance to the local
health law and practices over the
basic requirements.
In the long run, the upkeep of
the system technically will help
to bring the solution up-to-date
but will require additional maintenance cost. This will have to be
factored into the long-term development budget.

Technology Aspects and
Computational Intelligence
A number of technologies are
applicable for individual identification, and for storing and retrieving records. However, a remaining
challenge is system integration of
multiple technologies into a coherent solution tailored for use with
specific institutions. Computational
intelligence (CI) paradigms offer
some advantages in automating and
creating a human-like capability in
healthcare applications [8].
A generic biometrics system
captures, usually in real-time, the
characteristics of an individual [9],
then processes and stores a record
in a database [10]. Biometric techniques can be chosen from among
fingerprint systems and facial,
voice, and iris recognition. In general, a combination of biometrics
and human-readable alphanumeric
characters such as a name and a
number, barcodes, or rFID tags,
are needed. Electronic record systems for a remote area should be
simple [11] and have a user interface that adapts to the local culture
and language(s). A summary of
recent efforts to create such systems in developing countries is
found in [12]. This paper discusses
projects in Kenya, peru, Haiti,
Uganda, Malawi, and Brazil.
42

CI-based techniques are suggested for process mining in hospital information systems [13]. Mining
of electronic medical records to
examine physician decisions is suggested, to identify the impact of
physician decision on patient outcomes and hospital costs [14]. This
is of significant importance in environments where physicians are not
held accountable for their costs in a
healthcare environment, perhaps due
to regulatory limitations. In fact, the
adoption of record systems in developed countries indicates that change
management, policy, and strategy
issues are the primary problems,
rather than technology [15]. In addition, training and managing healthcare staffs to deal with technological
systems are significant issues [16].
CI-based biometric technologies are
reported for the representation and
recognition of incomplete biometric
data, discriminative feature extraction, biometric matching, and online
template updating [17]. Wireless
systems for patient-to-point-of-care
communications will need careful consideration of data processing
designs based on when, where and
how the information is obtained and
stored [18]. A tailor-made solution
for tracking information integrates
rFID, GpS and Wi-Fi [19], with the
need to assess information security
risk and adjustment of technologies
and procedures allowing access to
authorized personnel [20]. Intelligent optimization methods have an
important role in event mining of
network activities in rFID logistics applications [21], including the
optimization of rFID sensory capabilities [22].

A Simplified Example
In order to put the discussion of
the solution into perspective, it is
instructive to describe an example
healthcare system to illustrate the
overall principle quantitatively. In
our example, we make an assumption of three community healthcare
workers per central medical facility and that the central medical

facility has an addressable patient
base of 8000 inhabitants. In other
words, the ratio of central medical
facility to community healthcare
worker to addressable inhabitants
is 1:3:8000, respectively.
If we assume that 1% of the
addressable patient base is not
feeling well each day, there will
be an average of 80 patients using
the system per day. Dividing the
workload among the three community healthcare workers each
will have a case load of attending
to about 27 transactions each working day. For a 10 hour shift, each
patient will have a transaction time
of about 20 minutes, which may
include commuting time by the
community healthcare worker if it
is an outreach family visit. We may
assume that the time is reasonable
as most of the illness probably
would be routine.
Also in our example, we assume
each month the central medical
facility will handle 80 # 30 = 2400
transactions. Over the same period,
each community healthcare worker
will handle about 27 # 30 = 810
transactions. For each transaction
a patient is assumed to be charged
$1, so the total revenue for sharing
between the healthcare workers
and the central medical facility
amounts to $2400. If the healthcare worker takes 60% of the revenue, the monthly income for each
worker is 2400 # 0.6/3 = $480 or
$16 per day. The central medical
facility will have a monthly-revenue
of $960.
For locations where people
generally live on less than $2 a
day, charging $1 for seeing a community healthcare worker could
be a heavy burden. If the patient
fee is reduced, it will weaken the
income for both the community
healthcare workers and the central medical facility. Some kind
of local business approach has to
be implemented to make the case
sustainable.
One way is to overlay the system with a micro-health-insurance

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

SprING 2013

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - Spring 2013