Using Mobile Phones in a Real-time Biosurveillance Program:
Lessons from the frontlines in Sri Lanka and India
Gordon A. Gow
Associate Professor
University of Alberta
CANADA
gordon.gow@ualberta.ca
Nuwan Waidyanatha
Senior Researcher, LIRNEasia
12 Balcombe Place, Colombo 08
SRI LANKA
nuwan@lirneasia.net
Paper submitted to 2010 IEEE International Symposium on Technology and Society (ISTAS ’10),
Wollongong, New South Wales, Australia. [Submitted Dec. 16, 2009]
Abstract
The Real-Time Biosurveillance Program (RTBP)
is a multi-partner research initiative to study the
potential for new Information and
Communication Technologies (ICTs) to improve
early detection and notification of disease
outbreaks in Sri Lanka and India. A key
component of this project involves frontline data
reporting using mobile phones to overcome
problems of Internet-access in remote locations.
This paper provides a brief overview of
applications available for mobile phone-based
data reporting, describes a formative evaluation
framework used in the study, and discusses
initial findings related to technology acceptance
among frontline health workers operating in a
developing country.
1. Introduction
The Real-Time Biosurveillance Program
(RTBP) is a multi-partner research initiative to
study the potential for new Information and
Communication Technologies (ICTs) to improve
early detection and notification of infectious
disease outbreaks in Sri Lanka and India.
Under the current systems in Sri Lanka and
India, patient data from regional and community
health centres is gathered using paper-based
forms and procedures. These forms are then sent
to regional health officials where data analysis is
carried out by qualified staff to identify potential
disease outbreaks. Notifications are then issued
from the regional health administrations to local
authorities, again using paper-based reporting
methods. Under the present system it can take
up to 30 days for information to move through
these various steps, leading to delays in both
outbreak detection and notification.
Leading experts in the field of biosurveillance
and health informatics have argued that
improvements in disease detection and
notification can be achieved by introducing more
efficient means of gathering, analyzing, and
reporting on data from multiple locations [1].
The introduction of new information and
communication technologies (ICTs) is regarded
as a central means to achieve these efficiency
gains. The primary research objective of the
Real-time Biosurveillance Program (RTBP) is to
examine these claims more closely by producing
evidence to indicate in what ways and to what
extent the introduction of new ICTs might
achieve efficiency gains when integrated with
existing disease surveillance and detection
systems.
The RTBP project is a pilot study that
involves digitizing current paper-based
procedures using advanced ICT components. A
key step in this process is introducing an
efficient and cost-effective means of digitizing
patient data at front line health centres. Once in
digital form, patient data can be transmitted
immediately to a central server where rapid
analysis can take place using statistical data
mining software developed by the Auton Lab at
Carnegie Mellon University [2]. Results of this
analysis are then made available on an ongoing
basis to regional and local health officials as
electronic notifications accessible through a
variety of means, including mobile phones.
2. Digitizing patient data with mobile
phones
The first step in a real-time biosurveillance
system is to collect and report patient data in
order to conduct epidemiological analysis.
Mobile phones present an opportunity to both
digitize and transmit patient data using existing
commercial wireless data service in a relatively
cost-effective manner.
The use of commercial wireless services to
support health care initiatives in developing
countries is gaining recognition within an
emerging field known as “mHealth” [3]. In fact,
several notable open source software platforms
have been introduced in recent years to capitalize
on potential of the mobile phone as cost-effective
means for collecting and reporting data from the
field, including Episurveyor and FrontlineSMS.
In comparing across platforms, the technical
design of these applications generally involves
three basic considerations:
• Access to the application: (native or
download)
• Data entry method: device keypad, camera,
voice, touchscreen
• Data transmission: voice, SMS, packet access
(i.e., Internet) using GPRS or 3G wireless
link
The m-HealthSurvey is a customized
application designed to operate on a mobile
phone and was developed by indian Institute of
technology Madras’s Rural Technology and
Business Incubator (RTBI) as a means of
collecting real-time patient disease, syndrome,
and demographic data for rapid detection of
disease outbreaks.
While the research team is aware of other
initiatives using mobile phones specifically for
disease surveillance, such as e-MICI [4], the
decision to develop the mHealthSurvey was
based on the need to assure interoperability with
other components of the biosurveillance system
and to install and operate it on any low cost Javaenabled mobile phone. The application is now
being pilot tested with frontline healthcare
workers in both India and Sri Lanka as part of
the Real-time Biosurveillance Project.
Digitizing patient data is done on the mobile
phone using an installed java-based application.
The application is a Java 2 Micro Edition
(J2ME) Midlet that incorporates MIDP2.0
(Mobile Information Device Profile) and
CLDC1.1 (Connected Limited Device
Configuration) JSR (Java Specification Request)
components. This allows the Midlet to be ported
on to any, manufacturer independent, mobile
phone that supports the same (MIDP2.0 and
CLDC1.1) or higher versions of the JSR that
transparently interact with the mobile device
display/controls and connectivity, respectively.
The midlet has been successfully tested on a
number of devices, including a Nokia 3110c,
Amoi A636, Gionee v6600, Gionee v6900,
Motorola , Sony Ericsson s302c.
2.2 Application access and configuration
Users must first download the application to
the mobile phone by entering a pre-defined URL
into the phone’s Wireless Application Protocol
(WAP) browser, which then contacts a server on
the Internet using a packet access service such as
GPRS (General Packet Radio Service). Once
contact is established the user sends a request to
the server and retrieves a Java Archive (JAR)
file, which is then downloaded and installed on
the phone.
After installing the application the first step is
executing the download list function which
retrieves the lookup values from a database that
includes pre-defined lists of disease names, sign,
symptoms, age-groups, gender names, casestatus, location types, and health worker types
(Figure 1 (a)). This is usually a onetime step but
users are encouraged to execute this function
periodically, especially, update the list of
disease, signs, and symptoms on their mobile
phones.
The next step is a profile registration process
(Figure 1(b)) to generate a universal unique
identifier for the user from the database to the mHealthSurvey application. This identifier is then
used to tag all records received from this
particular user profile. Each mobile phone can
carry more than one health worker profile so that
several healthcare workers at a health facility can
share a single mobile phone.
Next the user must identify the village they
work in, which is facilitated through the location
menu. The health worker selects their
jurisdiction type; i.e. the administrative
geographic area name; followed by entering the
2.1 The m-HealthSurvey application
The mHealthSurvey application developed for
RTBP involves three basic steps: (1)
downloading and customizing the application for
the frontline healthcare worker by creating a
simple user profile; (2) digitizing patient data;
(3) transmitting that patient data using General
Packet Radio Service (GPRS), which is a
wireless data service deployed as a standard
feature in many mobile phones. Patient data is
transmitted by GPRS over the mobile operator’s
network to an Internet gateway, where it then
goes to a central database for storage and
analysis.
2
name of that area (Figure 1(c)). For example,
location type = “PHI” and the location name =
“Kuliyapitiya” would retrieve all the villages
belonging to the Kuliyapitiya PHI area.
2.3 Data entry method
After installation and configuration, the
application presents the frontline healthcare
worker with a series of fields and menus that are
completed using the standard keypad on the
phone. Patient data is captured with several
attributes included on the health survey form:
case date and time; health worker id; location
name; patient first name; patient last name;
notes; gender; age-group; disease; symptoms;
signs; and case-status (Figure 1(d) and 2(e)).
Date and time are automatically set at the
moment a patient record is transmitted,
anticipating that health workers are sending realtime data. However, the application does gives
users the option of overriding this feature and
inserting customized date and time information
should there be a delay in submitting the record.
Health worker Id, location name, age-group,
gender, and status are pre-defined drop-down
lists that are initially set through the download
list, profile registration, and location defining
functions.
The m-HealthSurvey application includes
several design features intended to simplify data
entry. For example, a pre-defined menu of
disease types is incorporated into the application
so users need only begin to enter the first few
characters of the diagnosis in the appropriate
field, which will bring up a list of possible
corresponding diseases (Figure 1(e)).
Figure 1. m-HealthSurvey Screens (a) main
menu (b) profile (c) location (d) patient record
page 1 (e) patient record page 2
2.4 Data transmission
The m-HealthSurvey, client application,
communicates data, with the database, through
Hyper Text Transfer Protocol (HTTP) Post,
Request, and Get functions embedded in a server
side PHP hypertext pre-processor application
that follows a REST (Representation State
Transfer) like architecture. Each record is
typically about 2kb in size, and in both Sri Lanka
and India the transmission cost per record is less
than $0.01(USD), with a more specific estimate
placing it at $0.0002 USD per record.
2.5 Workflow
A typical procedure will involve transcribing
patient data from a handwritten record (or “chit”)
into the phone to create a digital record. While
this approach may be relatively simple to
implement from an application design
standpoint, it presents several drawbacks in
terms of efficiency and accuracy when dealing
with frontline health care workers. For example,
entering data using the telephone keypad is
cumbersome and time consuming. Frontline
healthcare workers may find it disruptive to
current workflow or find that the transcription
process introduces unacceptable error rates when
healthcare workers are entering data from
handwritten forms.
One concern is that
healthcare workers may be reluctant to adopt this
technology if they perceive it to be disruptive to
3
current practices. Adoption and use of the
mobile phone for entering patient data at the
frontline is essential evaluating the real-time
biosurveillance system as a whole and it is
therefore essential that potential sources of
resistance be addressed by the research team.
For these reasons, it is important at an early
stage in the project to understand potential
problems associated with digitizing patient
records using mobile phones. Having clear
understanding of these challenges, can allow the
research team to modify the design of the
application or introduce other methods (e.g.,
training, procedural changes) to improve the
timeliness and accuracy of the digitization
process. The biosurveillance system relies on a
sound database of patient records being provided
in real or near real-time.
this stage in the project is to better understand to
what extent frontline health care workers are
willing (and able) to use the mHealth Survey
application to enter patient data in real-time, and
whether there are certain impediments or barriers
to this initial data collection that might be
addressed with future modifications to planning,
training, and design.
Measures of acceptability have been
developed perhaps most notably in the
Technology Acceptance Model (TAM), which
introduced two theoretical constructs perceived
ease of use and perceived usefulness, to study
factors influencing technology acceptance of
ICTs by individuals within organizations. TAM
has been influential in drawing attention to both
social and technical factors in technology
acceptance, with research findings suggesting
that on, the one hand, “mandatory, compliancebased approaches to introducing new systems
appear to be less effective over time than the use
of social influence to target positive changes in
perceived usefulness”. On the other hand,
perceived ease of use through system design is
also a critical factor for increasing user
acceptance, especially if it leads to
improvements in workflow that can empirically
demonstrated to show users the comparative
effectiveness of a new system [6, p. 199].
In addition to the TAM literature, a sectorspecific approach to acceptability has also been
developed by the US Centers for Disease Control
(CDC) and focuses on measures involving
various participation rates among health
professionals [7]. Wagner writing specifically
for the biosurveillance field, adopts the CDCbased approach, and defines acceptability as “the
willingness of people and organizations to
participate in a surveillance system or to use it”
(p. 512).
In evaluating the RTBP, the soco-technical
factors identified in the TAM literature have
been blended with Wagner’s general definition
to generate an evaluative framework based on
the fundamental principle that willingness to use
the system is, among other factors, indicative
how frontline health care workers perceive the
new technology in terms of ease of use and
usefulness in relation to current practices. An
assessment of these perceptions can be derived
from an intersubjective account based on
observations and survey data but can also be
obtained from various objective indicators
related to usage of the mHealthSurvey by
frontline healthcare workers involved in the
project.
3. Formative Evaluation Framework
Wagner [5] stresses the importance of field
testing, noting that “many important
characteristics of biosurveillance systems can
only be determined once they are deployed, at
least partially, in the field” (p. 507). Field
testing can be intended as formative or
summative in approach, with formative studies
aimed at producing insight as to the effectiveness
of a component or sub-system and identifying
potential improvements to that component. An
evaluation framework for field testing of
biosurveillance systems will examine the
attributes of the system or system components,
which can be divided into three general
categories:
• Human resources (e.g., frontline healthcare
workers, epidemiologists)
• Communications networks (e.g., wireless
links, Internet connections)
• Computing resources (e.g., mobile devices,
data mining and analytics software)
Whereas communications networks and
computing resources will be evaluated based on
attributes associated with technical performance,
evaluation of the human element of a
biosurveillance system must consider the
relationship between both social and technical
attributes.
Among these considerations,
acceptability is a defining attribute for the human
element in a biosurveillance system because it
will exert an influence on usage practices among
frontline health care workers with “downstream”
consequences for data analysis and disease
detection components in the system. As such,
the key goal of undertaking a formative study at
4
Wagner identifies a total of 21 objective
attributes relevant to the evaluation of
biosurveillance systems. For the purpose of the
formative evaluation study, the research team
selected five attributes based on Wagner’s
framework, structuring them into a relational set
around the key attribute of acceptability:
• acceptability of system or components
• reliability
• data quality
• time latency
• portability
For the purpose of evaluating the RTBP, four
attributes related to participation have been
identified as indirect indicators of acceptability.
Reliability is defined as the likelihood that a
trained frontline healthcare worker will be
capable of using the mHealth Survey to correctly
enter patient data application. This was
measured by performance testing conducted by
the research team as part of a certification
exercise.
Data quality is defined as the accuracy and
completeness of patient information as entered
by the frontline health care worker using the
mHealthSurvey. This was measured against
basedline data obtained from official (paper)
records that continue to be generated in both
countries.
Time latency is defined as the delay between
the moment a frontline health care worker
receives patient’s case data and the time that data
is uploaded to the RTBP analytics software.
Given the focus of this project as a real-time
biosurveillance system, latency is a significant
concern. Latency is measured by comparing
timestamps.
Finally, portability is defined as the likely
amount of effort that will be required to
introduce the system into a new location. This
was measured as disruptive impact on existing
procedures and practices involving frontline
healthcare workers as well as other medical staff
indirectly involved in patient records. In terms
of acceptability, the measure associated with the
portability attribute may be indicative of
potential resistance to the introduction of
mHealthSurvey into current arrangements.
Where it is perceived as highly disruptive,
frontline and other health care workers may be
less accepting of making changes to
accommodate the mobile phone as a means of
digitizing patient records.
A group of frontline health workers, 29 in
India and 16 in Sri Lanka, were supplied with a
mobile phone and the mHealthSurvey
application in order to digitize patient records
and enter data into the real-time biosurveillance
system. Data in India come from four Primary
Health Centers and 25 Health Sub centers
located in Tamil Nadu province. In Sri Lanka
data are collected from a total of seventeen
hospitals and clinics from various regions of the
country. Health workers in India began
submitting case data in June 2009. Due to delays
in receiving government clearance the Sri
Lankan participants began in submitting data in
September 2009.
In Sri Lanka, the dataset was extracted from
treatment chits, which are essentially pieces of
paper given to each patient upon examination by
a senior medical officer at the hospital or clinic.
Each chit carries the name, gender, and age of
the patient as well as the diagnosed disease,
symptoms, signs, and the treatment. Research
assistants belonging to a community-based
primary Health Care services providing
organization, Sarvodaya Suwadana Centers
visited each hospital or clinic to obtain the chits
and then digitize the health records by entering
the data using the mHealthSurvey application.
The role of the research assistants in this case
were intended mimic the introduction of a new
resource person to the clinic or hospital that
might accompany the introduction of a real-time
biosurveillance system.
The dataset in India was extracted from the
OPD registry, where the senior medical officer
also documents each patient’s identification
number, gender, age, and diagnosis on a chit
similar to that used in Sri Lanka. However, in
this case almost all of the frontline healthcare
workers (26 of the 29) submitting data were
already working in the health centre. The senior
medical officers associated with the health
centres or hospitals in either country were not
involved in digitizing the health records.
4.1 Real-time vs other-time submission
patterns
There was an initial training period of about 4
weeks (W01-W04, Figure 2) during which time
the senior medical officers needed to be
reminded constantly to record the diagnosis on
the chits. This had not been general practice
previously and posed an initial challenge for the
research team. For those in Sri Lanka Figure 2
shows a gradual increase of data influx rate of
4. Initial Findings
5
200 per week after the initial training period.
Figure 3 shows a learning curve gradient of 100
records per week increase for those in India.
These two figures reflect growing familiarity
with the application among frontline healthcare
workers, suggesting that acceptability levels as
indicated in real-time submission patterns may
increase over time in conjunction with
perceptions around ease of use.
In terms of real-time data submission counts,
these varied across countries and appear to be
linked to the ability of the frontline health care
workers to manage this imposed new
responsibility with other existing duties. In the
case of Sri Lanka, the research assistants were
tasked with this role specifically and appear to
have been able to more effectively manage realtime submission of records as noted in Figure 2.
Unlike the research assistants in Sri Lanka,
record counts from India suggest that frontline
health workers had difficulty coping with realtime submissions as a result of competing
demands on their time during peak patient
visitations. As such, many of them were
required to transcribe and submit patient records
after working hours, and in some cases on the
weekends. As a result, 86 percent of the data
arrived during other time with only 14 percent
real-time.
4.2 Corrupt and clean records
While the research assistants in the Sri Lanka
may have achieved higher levels of real-time
submission of records, they are also relatively
inexperienced health workers compared with
their Indian counterparts and face challenges in
terms of English language competency and
awareness of medical terminology. Insofar as
the mHealthSurvey predictive menu is
incomplete, the research assistants were
sometimes required to guess at incomplete
details provided on their chits when transcribing
to the mobile phone. This is one possibility for
the relatively high error rate in weekly record
counts from Sri Lanka as illustrated in Figure 4.
Error and Good record counts by Week
Sri Lanka
10000
Record Counts
8000
6000
4000
2000
0
W01 W02 W03 W04 W05 W06 W07 W08 W09 W10 W11 W12 W13
Week
Bad
Good and Bad record counts by Week
Real-Time and Other-Time Counts for each Week
India
India
800
800
Record Count
700
600
Record Counts
600
500
RT
OT
400
400
200
300
200
0
100
W01 W02 W03 W04 W05 W06 W07 W08 W09 W10 W11 W12 W13
0
Week Number
W01 W02 W03 W04 W05 W06 W07 W08 W09 W10 W11 W12 W13
week
Figure 2. Sri Lanka real-time (OR) and othertime (OT) submission counts by week.
Bad
1800
On the other hand, the Village Health Nurses
(VHNs) in India with 10 or more years of
experience may have been submitting relatively
few real-time records, but error rates were
significantly lower than those found in Sri
Lanka. Moreover, as Figure 5 shows, these
levels began to subside over time, as the VHNs
gained experience with application and when the
error rates were drawn to their attention by the
research team. In fact, by week 10 of the
1200
600
0
W01 W02 W03 W04 W05 W06 W07 W08 W09 W10 W11 W12 W13
Week No
OT
Good
Figure 5. India good and bad record counts by
week
Rea-Time vs Other-time plot by Week
Record Counts
Good
Figure 4. Sri Lanka good and bad record counts
by week
RT
Figure 3. India real-time (OR) and other-time
(OT) submission counts by week.
6
evaluation, the error rate had declined to almost
zero for the remainder of the evaluation period.
given above for record submissions clearly show
that the project is not achieving this target. In
the case of Sri Lanka, concerns about data
quality are also in need of attention.
4.3 Certification/usability exercise
In light of these concerns, the research team
undertook a series of observations and interviews
with participants about their knowledge on
dealing with issues in relation to lack of
diagnosis information in the mobile RMS, about
their habits concerning maintenance and upkeep
of the mobile phone and practices for seeking
technical support when confronting problems
with the device or application.
Part of the identified problem in Sri Lanka
was that the mHealthSurvey application created
a new step in existing procedure, asking the
senior medical officers at the hospitals to record
the disease and syndrome on a chit for the
research assistants. Having observed that
medical officers often forgot or ignored this
additional request, the research team recently
introduced an intermediary step involving a
stamp and ink pad which is intended to help
remind the medical officer and to facilitate the
process by creating a more standardized
procedure. The stamp produces a blank form
with the required fields that are to be transcribed
into the mHealthSurvey application. Because it
was introduced relatively late into the study
period, the effectiveness of this intervention is
not yet known.
The research team also learned that a number
of frontline healthcare workers in India had
under-reported patient data in an effort to avoid
unwanted attention and possibly extra duties
from senior health officials.
For accurate epidemiological analysis patient
data must be relatively free of errors (ideally,
less than 5 per cent of all records per batch).
While the India health workers tended to correct
many user-instigated errors the problem in Sri
Lankan remains a concern. Improved data
quality might be achieved through a more
comprehensive predictive menu of disease and
syndrome incorporated in the mHealthSurvey
application. Even when experienced health
workers are submitting patient records it was
found that local spelling or descriptive
conventions may differ for various symptoms.
For example, one user might be asked to enter
“back pain” and another may enter “backache”.
This can be avoided if all various disease,
symptoms, and signs stored in the mobile phone
were linked to predictive menus. The project
had considered using the mobile phone
An m-HealthSurvey Certification Exercise
was carried out as part of a formative evaluation
of the m-Health Real-Time Biosurveillance
Program (RTBP) in an effort to assess the
usability and acceptability of the mHealthSurvey mobile application. The exercise
was conducted in summer 2009, with health
workers in Sivagangai District, Tamil Nadu,
India and in Kurunegala District, Sri Lanka.
The first step of the exercise required
participants to demonstrate their ability to
establish GPRS connectivity, download the JAR,
and configure the JAD for immediate use. The
second step had participants enter and submit six
health records in various combinations using the
mHealthSurvey on a mobile phone.
Observations from the exercise revealed a
disparity between the age groups of the health
workers using the m-HealthSurvey for RTBP
data submission – younger Sarvodaya Suwadana
Center health workers, between the age of 18 –
35 in Sri Lanka, were able to complete the
exercise easily in the allotted time by themselves
without any help. While the older Tamil Nadu
Health Department Village Health Nurses,
between the age of 30 – 50, but with 10 or more
years field experience, were unable to complete
the exercise in time and, except for one or two of
them, all others required guidance and
assistance. This, adoption and usability,
disparity of mobile phone applications between
older and younger generations is also evident in
findings from a study of mobile adoption at the
Bottom of the Pyramid (BOP), where
“youngsters are more likely than older to adopt
mobile phones beyond voice” [8]. Higher levels
of reliability in predicted use of the application
by younger frontline healthcare workers,
however, is tempered by concerns about data
quality as noted above.
4.4 Impact on existing procedures
Current records show that over 100 patients
visit the hospital and clinic each day in Sri
Lanka. The Public Health Center sees the same
number of patients; where as the Health Sub
Centers may see only 50 patients a week. Given
these numbers, the project has set a target of at
least 4000 records to arrive from India and the
same count from Sri Lanka each week. Figures
7
dictionary as a predictive text function but this
feature is still not standardized across all mobile
devices and would compromise the ability to
port the m-HealthSurvey J2ME across to
different java compliant mobile hand-helds.
In both Sri Lanka and India a major concern
for frontline health care workers appears to be
the overwhelming patient care rates; thus,
facilities in both countries often have to provide
care to over 100 patients within a span of 5
hours. In order to reduce disruption to workflow,
the research team has looked at the option of
bypassing the transcription process and having
the medical officers digitize and submit the
records directly. This measure might reduce
workload and improve real-time data entry as
well as data quality. However, the standard
mobile phone keypad has proven to be
cumbersome method for entering data and
remains a potential source of difficulty for
frontline healthcare workers.
A less disruptive approach may be to
introduce a digitization process that retains
current handwriting method but provides a more
efficient means of capturing and transmitting
patient records.
The benchmark for the
digitization process of each record is 15 seconds
or less—approximately 10 per cent of a patient
care time (150 patients in 5 hours would result in
spending 2 minutes with each patient and 12
seconds to enter a record), which is an
approximate equivalent to the hand writing
procedure currently in practice. In order to
maintain or improve current workflow against
this benchmark may looking beyond the standard
keypad for data entry and to introduce a touch
screen or other method for digitizing data using
the mobile phone’s camera for optical data
capture [9, 10].
and clean records (quality); results from a
certification and testing exercise conducted with
frontline healthcare workers (reliability); and
observations of and interviews with healthcare
workers to determine impact on data collection
procedures (portability).
Initial findings suggest that younger
healthcare workers are more likely to adopt and
use the mHealthSurvey application as compared
with older, more mature healthcare workers.
However, there appears to be a significant tradeoff in terms of data quality between these two
groups, with more mature health care workers
providing more accurate and comprehensive
records than their younger counterparts.
Initial findings also suggest that minimizing
impact on current workflow is a key
consideration with the introduction of a mobile
phone-based data entry system. However, the
additional human resources necessary to support
this system in the form of a data entry clerk or
assistant may offset any efficiency gains or
improvements in data quality that would
otherwise be achieved with real-time data
provision. As such, the research team will
continue to examine ways to maintain or
improve workflow through modifications to
interface design, including bypassing the
standard telephone keypad entry in favour of
other more experimental methods in optical data
capture.
6. Acknowledgements
The authors wish to acknowledge the funding
agencies: International Development Research
Center (IDRC) of Canada for the grant (105130)
in full support of the project titled – Evaluating a
Real-Time Biosurveillance Program and the US
Centre for Disease Control and Prevention (R01PH000028) for partially supporting the
researchers Maheshkumar Sabhnani and Artur
Drubrawski. The authors are grateful for the
enthusiastic participation and support of the
Provincial Director of Health Services office in
Wayamba (Northwestern Province), Sri Lanka
and the Deputy Director of Health Services
office in Sivagangai District, Tamil Nadu, India
inclusive of the Medical Officers and Health
Workers in those project areas.
5. Summary
The mobile phone offers an innovative and
potentially effective means of digitizing patient
records to support real-time biosurveillance
programs in developing countries. The RealTime Biosurveillance Project (RTBP) is piloting
the mHealthSurvey application among frontline
healthcare workers in India and Sri Lanka as part
of an end-to-end biosurveillence system.
The project team has recently concluded a
formative evaluation of the digitization process
using the evaluative framework based on four
indicators of usage and proficiency related to
Wagner’s acceptability attribute: real-time vs
other time submission patterns (latency); corrupt
7. References
[1] M. Wagner, “The Challenge of Biosurveillance:
Introduction,” in Handbook of
Biosurveillance, M. Wagner, A. Moore, and
8
R. Aryel, Eds. London: Elsevier Academic
Press, 2006, pp. 3-12.
[2] A. Dubrawski, M. Sabhnani, M. Knight, M.
Baysek, D. Neill, S. Ray, A. Michalska, and
N. Waidyanatha, “T-Cube Web Interface in
Support of Real-Time Bio-surveillance
Program,” presented at 3rd International
Conference on Information and
Communication Technologies and
Development ICTD 2009, Doha, Qatar,
2009.
[3] Vital Wave Consulting, “mHealth for
Development: The Opportunity of Mobile
Technology for Healthcare in the
Developing World.,” vol. 2009: UN
Foundation-Vodafone Foundation
Partnership, 2009.
[4] D. DeRenzi, T. P. N. Lesh, C. Sims, M. Mitchell,
M. Chemba, Y. Hamisi, D. Schellenberg,
and G. Borriello, “e-IMCI: improving
pediatric healthcare in low income
countries.,” presented at ACM CHI 2008
conference on Human Factors in Computing
Systems, Florence, Italy, 2008.
[5] M. Wagner, “Methods for Field Testing of
Biosurveillance Systems,” in Handbook of
Biosurveillance, M. Wagner, A. Moore, and
R. Aryel, Eds. London: Elsevier Academic
Press, 2006, pp. 507-515.
[6] V. Viswanath and F. D. Davis, “A Theoretical
Extension of the Technology Acceptance
Model: Four Longitudinal Field Studies,”
Management Science, vol. 46, pp. 186-204,
2000.
[7] Centers for Disease Control and Prevention,
“Updated Guidelines for Evaluating Public
Health Surveillance Systems:
Recommendations from the Guidelines
Working Group,” Morbidity and Mortality
Weekly Report. Recommendations and
Reports, vol. 50, pp. 1-35, 2001.
[8] H. d. Silva and D. Ratnadiwakara, “Bottom of
Pyramid in Emerging Asia,” vol. Dec. 10.
Chicago, Ilinois: Paper presented at Mobile
2.0: Beyond Voice? pre-conference
workshop at the International
Communication Association (ICA)
Conference, 2009.
[9] T. Parikh and P. Javid, “CAM: A Mobile
Interaction Framework for Digitizing Paper
Processes in the Developing World,”
presented at ACM Symposium on User
Interface Software and Technology (UIST),
Seattle, Washington, 2005.
[10]
A. Joshiy, M. Zhangx, R. Kadmawalay, K.
Dantuy, S. Poduriy, and G. S. Sukhatme,
“OCRdroid: A Framework to Digitize Text
Using Mobile Phones,” vol. 2009:
OCRdroid Project, no date.
9
View publication stats