prasavGraph: Android based Labour Monitoring
Shalini Singh
, Kolin Paul
, Geeta Yadav
and Sanjiva Prasad
Department of Computer Science and Engineering, IIT Delhi, New Delhi, India
ICMR, New Delhi, India
Partograph, Smart Phone, Labour, Maternal Health.
The simplified Partograph is a concise charting mechanism used by skilled birth attendants for recording and
communicating the important parameters and developments during the labour process. Compliance and cor-
rect plotting of partographs may be significantly improved with the use of technology. This paper reports the
development of an Android-based smart phone/tablet application to plot the partograph. To date such a com-
plete application is not available. We envisage that over a period of time, systematic plotting of partographs
will encourage providers to use it as a decision-making tool and thereby reduce morbidity and mortality both
of mother and foetus/newborn arising out of complications during labour and its sequelae.
The Millennial Development Goal 5 (MDG 5) tar-
geted a 75% reduction in maternal mortality and a
two-thirds reduction in child mortality. MDG 5 was
formulated following a 1987 WHO conference where
the Safe Motherhood Initiative was announced, of
which the adoption of the partograph was an integral
part (Hogan et al., 2010). An outcome of this Initia-
tive was a comprehensive manual on preventing ma-
ternal deaths (Royston and Armstrong, 1989). A re-
view in 2000 (Raleigh, 2000) cited areas for improve-
ment such as increased use of manual vacuum aspi-
ration for the management of incomplete abortion, as
well as the education of midwives in the use of the
partograph, and training traditional birth attendants.
Maternal mortality worldwide stood at about
343,000 in 2008. Studies showed that the main causes
of maternal death were anaemia, post-partum haem-
orrhage (PPH), sepsis, malaria, delay, and eclampsia.
“Delay” was cited as a primary cause relating to ob-
structed labour, i.e., the failure of fetal descent de-
spite healthy uterine contractions, which accounted
for 8% of maternal deaths worldwide (though of-
ten misreported as PPH or merely as “delay”). Ob-
structed labour is also the major cause of mater-
nal morbidities such as obstetric fistulae which are
thought to afflict an estimated 10-20 million women
worldwide. It can be prevented by better trained atten-
dants, timely monitoring and emergency intervention.
The simplified partograph is a concise chart-
ing mechanism used by skilled birth attendants
for recording and communicating the important pa-
rameters and developments during the labour pro-
cess (WHO, 2015). This single page sheet is prepared
for collecting and recording all important information
over a 12-hour period starting with the onset of labour
through to delivery. It acts as a decision-making aid in
that it indicates when labour is proceeding at a normal
rate and when doctors should prepare for carrying out
Compliance and correct plotting of the partograph
has been a major issue in its widespread use and in
research based on the data collected. The adoption
and accuracy of partographs may be significantly im-
proved with the use of technology. The ubiquity of
relatively inexpensive smart phones/tablets provides
an opportunity, with a well designed electronic Parto-
graph significantly increasing use and correct record-
ing of the labour processes. This paper reports the
design and development of an electronic version of
the simplified partograph. This mobile/smart phone
application prasavGraph (Figure 1) enhances the
simplified Partograph by recording events at the time
of admission and during the active phase of labour up
to 30 minutes after delivery.
The partograph may be considered to consist of
three sections, each of which has different inputs
the fetal record
the progress of labour record
the maternal record
Singh, S., Paul, K., Yadav, G. and Prasad, S.
prasavGraph: Android based Labour Monitoring.
DOI: 10.5220/0005820504580466
In Proceedings of the 9th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2016) - Volume 5: HEALTHINF, pages 458-466
ISBN: 978-989-758-170-0
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
Figure 1: prasavGraph.
The fetal record tracks fetal heart rate and the condi-
tion of amniotic membranes and colour of liquor. The
labour record tracks cervical dilation against time,
charting it with pre-printed alert and action lines. The
maternal record captures contractions, blood pres-
sure, pulse, urine output, temperature, and drugs ad-
ministered including those used to help uterine con-
tractions. This information is a concise summary of
the birth process and indicates when further interven-
tions are necessary.
The Android-based application prasavGraph em-
ulates the simplified paper Partograph and also pro-
vides additional features. The digitization of the
record keeping process during labour can help build
applications for decision support to reduce morbid-
ity and mortality both of mother and foetus/newborn
arising out of complications of labour and its seque-
lae. To the best of the authors’ knowledge, this is
the first such application that is based on the Android
platform, and which has been developed using a de-
lay tolerant 3-tier architecture. The Partopen is the
closest electronic aid that is available for recording
labour. However, the Partopen is not as comprehen-
sive as this effort (Underwood, 2013), although its
design goals were to work with the paper partograph
without changing the workflow processes.
The broad objectives of the prasavGraph applica-
tion are
1. Reducing data entry errors, achieved by stream-
lining and simplifying the data recording process.
2. Compliance with mandated/suggested procedures
during childbirth, ensured by making the applica-
tion run in real time with the option of recording
data “offline” too. However all such entries are
logged and colour coded.
3. Providing a Delay-tolerant Infrastructure, to en-
able the application to work in low network con-
nectivity areas.
4. Analytics, enabled by the availability of data in
digital form. A server database is integrated to
manage records of various devices in the same fa-
cility for better record sharing and visualization.
5. To act as a training aid.
In the next section, we describe the high level design
of the prasavGraph application.
ICT has been seen as an enabler for better provision
of healthcare. In particular, with the widespread use
of mobile phones, mHealth technologies are seen as
a means for low-cost and effective delivery of mater-
nal health care, in particular by providing better train-
ing to caregivers, better access to specialist support
and remote diagnosis, increased information about re-
productive health, and greater awareness of diseases
and their prevention and treatment. The deployment
of ICT does not automatically lead to improved out-
comes. This was seen in the case of the ClaimMo-
bile project (Densmore, 2012), where the smartphone
application did not replace the existing paper-based
flows, but suffered from additional limitations of un-
reliable connectivity and power. The e-IMCI project
in Tanzania achieved success in replacing bulky pa-
per protocols with PDA-based IMCI workflow mod-
els (DeRenzi et al., 2008). Open Data Kit (ODK),
an extensible set of phone based tools, supports the
collection of data using smartphones, thereby reduc-
ing errors in data and providing timely feedback and
decision support. While health-related ODK-based
applications have reported a degree of success, ma-
jor limiting factors include small screen size and con-
cerns related to the security of data (Anokwa et al.,
2012). CommCare is an open-source phone-based
platform that has been used by community health
workers in Tanzania for the care of pregnant women.
Since CHWs could now be monitored, this interven-
tion unfortunately resulted in a significant decrease
in the number of CHW submissions (Svoronos et al.,
). The MoTeCh project in Ghana (MoTeCH, 2011)
also faced resistance in its use from CHWs. Both
these experiences suggest that a greater participatory
role involving users is essential for successful adop-
tion. The Swasthya Slate (Swasthya, 2015) is a tablet-
based platform interfaced with a collection of med-
ical/diagnostic sensors that has been developed for
CHWs and medical professionals in rural India. The
most relevant work in addressing the labour process
is the PartoPen (Underwood et al., 2013), where a
pen-like device is used on specially printed parto-
graph forms, while also digitally recording the inputs
and providing timely audio reminders and feedback
prasavGraph: Android based Labour Monitoring
on a tiny screen on the device. The PartoPen ap-
proach is not to supplant the existing paper-based pro-
tocols, but to automatically extract information from
the forms and supplement the process with feedback
and reminders. The PartoPen has been used in teach-
ing and clinical studies in Kenya, showing the de-
vice’s robustness but also the limitations of such a
hardware-software solution in that it is oriented to-
wards successful filling of the partograph rather than
on the quality of care (Underwood et al., 2014).
As mentioned above, a key motivation for this ac-
tivity is to enable the faithful recording of data dur-
ing labour. Despite government and WHO initia-
tives, compliance in this aspect is low across the
world (Ollerhead and Barriers, 2014). The paper-
based approach is often prone to errors during data
entry, both in relation to the time and also in the fi-
delity of the data being entered. The fact that in many
settings the medical staff are overworked often causes
data to be entered retrospectively. Most importantly,
the data has to be manually (and often painstakingly)
digitized for any sort of data analysis.
Android-based smart phones, which are ubiqui-
tous today, are very good data acquisition devices.
The portability of applications across different hand-
sets and tablets is also very good. The design of the
electronic Partograph has been targeted at such de-
vices. The smart phone is used as the primary data
acquisition device. The application residing on it es-
sentially uses form-based entry with client-side data
validation on each data entry field. Colour-coded text
boxes and out-of-bounds warnings are built into each
such field. The application is built assuming an op-
portunistic communication framework (Pelusi et al.,
2006). It works as a standalone device but whenever
it is in the vicinity of a laptop in the premises, it repli-
cates the data that have been collected so far.
Figure 2: Architecture Diagram.
As shown in Figure 2, the tablet and laptop are
connected via WiFi or USB. The data is maintained
in a standard format (openMRS), and can be oppor-
tunistically uploaded to the central server (cloud) for
analysis. The design architecture resembles a Client-
Server model. The laptop containing the database and
web-server is used for backup and background pro-
cessing of the data. The laptop also provides the fa-
cility for making paper copies of the partograph.
The different modules of the Android application
Static User Interface Module: This module is used
for creating various user interfaces of the Android
Application. Static UIs in Android are created in
Network Module: This module is responsible for
interacting with the network either as a service or
otherwise. APIs used herein are:
Network Android API
Network Service API
Web Connection API
Database Manager Module: This module is use-
ful for carrying out database-related functions, in-
cluding backup and logging. APIs used herein
Database API
Database Connection API
Database Functionalities API
Notification Module: Notifications play an impor-
tant role in this Android application whereby it
notifies users (using an alarm / colour changes)
about the validity/correctness of the input data etc.
The partograph records the progress of labour.
The key parameters that are tracked are cervical di-
lation, and the vital parameters of the foetus and the
mother. The onset of active labour is indicated when
the cervical dilation is measured to be 4cm and con-
tinues up to 10cm, after which delivery occurs. Med-
ically, progress is normal if the rate of dilation is
1cm per hour. Partograph zones are colour coded, as
shown in Figure 3, to indicate different monitoring
modes for the birth attendant. Any time the cervix
Figure 3: Partograph Zones.
dilation is in the red zone, action has to be taken. Pro-
vision for alerting the attending physician is (option-
ally) present in the system. The yellow zone indicates
HEALTHINF 2016 - 9th International Conference on Health Informatics
that labour is sluggish and alerts the attending staff for
closer monitoring and/or augmentation of labour.
In the second and third stages a complete record
of the medication as well as other physiological con-
ditions of the mother is also maintained. On deliv-
ery, the details of the child(ren) need to be entered.
This feature, coupled with opportunistic communica-
tion with the cloud/hospital, can be used effectively
to automatically record births with municipal authori-
ties. Popups and/or beeps are built into the application
to serve as reminders to the attendant for recording the
data. The use of the web server helps in making these
data available on the central monitoring station.
A significant part of the design activity is concen-
trated on the User Interface (UI). The UI is essentially
composed of multiple forms which have been care-
fully designed to minimize the number of keystrokes.
Multiple versions of the application have been val-
idated by professionals, students, trainee doctors to
evaluate and report on the ease of using the applica-
tion. Figure 4 diagrams the application flow.
Figure 4: Flow Diagram of the Android application.
The application starts with the Home Screen,
which has options to either add a new patient and to
go to a pre-existing partograph, given a patient id or
a patient name. Clicking on the add new patient but-
ton takes the user to a screen which records the pa-
tient’s name and other attributes. The next form, in
Figure 5, allows the user to input the patient’s medi-
cal history and examination details at the time of ad-
mission. Some of the fields are mandatory while oth-
ers may be left blank. This is an important feature
because the state in which the patient is admitted is
not always completely known, e.g., all the test results
may not be known at that time. Many of the field
values (e.g., date and time of entry) are either set au-
tomatically from the machine settings or set to default
values. The right side of Figure 5 shows some of the
Figure 5: Form Based Entry.
physiological parameters of the patient that need to
be entered at admission. All these fields are checked
for validity and bounds, with alerts popping up on in-
correct or out-of-range values. The attending medical
staff has the option to override the flagged exceptional
values (e.g., a very high systolic BP value) but any
such value is (colour) coded and stored.
After the user inputs these data, the main parto-
graph screen is displayed. The cervical dilation value
is noted and the partograph “starts” recording in real
time only if the value is greater than or equal to 4 cm.
Once the details at admission have been entered, the
partograph is presented as shown in Figure 6. All
subsequent entries are made in this page. Note that
the top of the partograph continues to display admis-
sion time entries.
This screen has three different input modes: 1
and 2
stages and the 3
stage with delivery de-
tails. The 1
stage requires recording cervical dilation
values periodically and the cervical dilation value vs
time graph gets updated automatically. During 1
stages, the recording of data such as foetal heart
rate, duration and frequency of uterine contractions,
pulse, BP etc. is done at predefined intervals (30 or
prasavGraph: Android based Labour Monitoring
Figure 6: Cervical Dilation.
15 minutes). The tabular display gets automatically
time-stamped and hence only the actual time at which
data are entered gets noted. The first column of left
half of the figure shows a partial list of the parameters
that are monitored at regular intervals of 30/15 min-
utes depending on the stage of labour. The values are
displayed in the table below the graph aligned with
the time axis of the cervical dilation graph (CDG). As
mentioned earlier, this application is used for mon-
itoring all the 3 stages of labour. In Figure 7, we
show one of the forms that needs to be filled in the
stage. The right side presents the data regarding
the neonate(s) post delivery.
In case the data input from the 2
stage in the
CDG shows some abnormality, the application noti-
fies the user by an alarm. Similarly during a 3
Figure 7: Third Stage.
input procedure, if any parameter is found to be ab-
normal, the user is immediately notified about it. The
application also suggests possible precautionary mea-
sures that can be undertaken. Delivery details of the
patient such as number of babies born, date and time
of birth, sex etc. are fed in the third part of the par-
tograph. The data collected are periodically synced
with the mysqlite database residing on the tablet and
the laptop.
To increase the reliability of the software and to
avert data loss due to malfunctioning of the device,
two approaches have been developed. First, the data
are replicated on an external SD card on the tablet.
Second, data are replicated over a network to the
server after a fixed time or done manually by the user.
In case of device failure, the SD Card is removed from
the tablet and inserted in another tablet and all the Par-
tograph data captured by previous device are restored.
Also, data captured at the server end can be copied to
an SD card and inserted in another tablet to restart
operations in case of device failure. Thus, the Par-
tograph can be seamlessly migrated from one tablet
to another (Figure 8). The use of this application
Figure 8: Restoring Partograph data from one device to an-
in a field setting required making the application ro-
bust, and the tablet safe from accidental mishandling.
Accordingly, the tablet where the application runs is
kept in a casing that has been made using Rapid Pro-
totyping. As shown in Figure 9, the device can be
screwed to the nurses’ trolley conveniently to prevent
it from being misplaced/mishandled in a field setting.
At the control station, doctors can view the parto-
graph(s) of the currently admitted patients arranged
in a grid pattern according to current cervix dilation
of the patient(s). On clicking any partograph, com-
plete details of patient are displayed. The criticality of
various patients can be assessed remotely and simul-
taneously priorities to attend to cases can be decided
HEALTHINF 2016 - 9th International Conference on Health Informatics
Figure 10: Information Flow from attending technician to Clinician.
Figure 9: Rapid Prototype Model of the device holder.
as illustrated in Figure 10. This application paves
the way for timely delivery of medical services in re-
mote/rural areas. A paper record of the partograph
can be obtained at any time from a browser, as a web-
server running on the laptop serves the page. As
shown in Figure 11, the webpage has been designed
to print on an A4 size printer and shows all the data
captured during labour.
The partograph has been tested initially by a resi-
dent doctor at Safdarjung Hospital, New Delhi. This
phase helped us identify crucial gaps in the imple-
mentation with respect to actual field use. A version
of this application underwent field testing in the Dis-
trict Hospital, Panchkula. Currently the application
is being used by student nurses and resident doctors
at two tertiary care centers in Chandigarh and New
Delhi. Additional trials are planned at primary, sec-
ondary and tertiary facilities including 23 district hos-
pitals across 13 states of India to understand scala-
bility and other system issues. Data collected from
these studies will be analyzed for reporting in a med-
ical journal.
In this section, we report some of the findings that
prasavGraph as a Data Collection and Reporting ap-
plication can provide. We would like to inform the
reader that, in this paper, we do not attempt any anal-
ysis or even comment on the data that have been col-
lected only the potential of such an infrastructure for
officials planning healthcare activities and for people
working in health care analytics is illustrated.
The preliminary data collected on prasavGraph
from hospitals give an insight into trends in cases of
pregnancy and its potential complications. The in-
terface of prasavGraph is user friendly and feeding
data in various columns of the app are kept simple
so that medical staff attending patients can send re-
ports without any glitches. Data obtained on prasav-
Graph from various hospitals were centrally recorded
and analyzed to study patterns among patients. The
medical/pathological evaluation of the patients blood
sample takes some time and the medical condition of
the patient may not allow the medical staff to fill all
the details on the prasavGraph. In this situation, the
staff has to rely only on the readily available reports.
Recording cervical dilation is a reliable option for the
prasavGraph: Android based Labour Monitoring
Figure 11: Partograph.
medical staff in determining the stage of labour, espe-
cially in cases with very little reaction time.
Figure 12 displays some of the admission time
parameters of patients admitted in this particular hos-
pital during the period of survey. Green and red por-
tions of the bar represent normal and “out of nor-
mal” ranges of medical/pathological parameters re-
Figure 12: Data Captured at Admission time.
spectively. The data were analyzed for about 200 en-
tries available on prasavGraph. The haemoglobin bar
shows that 46.55% of patients had lower than normal
blood haemoglobin value. The number of patients
who were detected HIV positive can also be seen.
10.61% patients had abnormal urine sugar levels and
25.95% patients had abnormal urine albumin. The
patients history of pregnancies also helps the doctor
to prudently scrutinize the case. In prasavGraph we
take values of gravida (number of times a patient got
pregnant), para (number of times a patient give birth
to a premature baby of gestation age greater than six
months) and number of live child(ren). Clearly these
data can identify potential areas where intervention is
Figure 14 shows some of the medical parameters
whose values can potentially fluctuate during active
labour. The figure shows the number of patients with
HEALTHINF 2016 - 9th International Conference on Health Informatics
Figure 13: Previous Pregnancy Data.
Figure 14: Parameters during Active Labor.
normal and “abnormal” values. Light and dark por-
tions of the bar represent static and ”fluctuating” med-
ical parameters respectively. For example, the FHR
bar shows the foetal heart rate varied significantly in
56 patients out of 196. The Pulse Rate bar represents
that 30 patients out of 197 had a fluctuating pulse rate.
Systolic blood pressure varied in 70 patients out of
196. The body temperature was found to change in
20 patients out of 190 compared with admission time
The standard partograph is shown by the orange
line in Figure 15 which assumes that the rate of in-
crease in cervical dilation is 1cm/hour. The data ob-
tained from prasavGraph shows deviations from this.
It can be inferred that the observed rate of dilation
was faster than this for dilation values upto 9 cm af-
ter which it progressed at a slower pace to reach the
maximum level of 10 cm. In cases of prolonged deliv-
ery, the duration of full cervical dilation i.e. 10 cms
was reached even in the 11
and 12
hours. Table
1 displays the number of patients with each cervical
dilation value at the time of admission to the hospi-
tal. Analyzing data collected over 200 patients, it was
Figure 15: prasavGraph Summary of Patients.
Table 1: Cervical Dilation at Admission Time.
Cervix Dilation (in cm) No. Of Patients
4 86
5 32
6 19
7 15
8 22
9 3
10 23
Figure 16: Measurements at Each Hour.
found that 88.5 % patients were admitted in the first
stage of labour, i.e., patients whose cervical dilation
is below 10cm. 11.5 % patients were admitted in the
second stage of labour i.e., cervical dilation is 10 cm.
This is essentially the percentage of patients admitted
to hospital at advanced stages of delivery/just after de-
livery. In 43.5 % cases, the data were properly filled
on the device from the start of active labour i.e., cer-
vical dilation 4 cm for the patient. Figure 16 shows
the number of points of data recorded from the start
of active labour, i.e., from hour 0 up to hour 12. In
11% cases, prasavGraph entered the red zone. A to-
tal of 1030 readings for each of the parameters (BP,
Fetal Heart rate, Temperature) were recorded during
patient examination. So, on an average, a patient was
examined 5 times during active labour. This illus-
trates some of the ways that statistical properties of
the data collected by the application can be analyzed
for developing decision support systems.
This paper reports the first complete implementation
of an Android application which enhances the tradi-
tional partograph to record the developments in the
entire labour-to-childbirth process. The application
has been built in a delay-tolerant framework, which
enables it to be used in network deficient or poor net-
work conditions.
Additional features like transferring the birth
record to municipal authorities can also be incorpo-
prasavGraph: Android based Labour Monitoring
rated into this application. A crucial feature that will
improve the adoption of this application will be the
integration of sensors to help semi automate the tak-
ing of readings. Another important feature that will be
tested in the next trials will be to alert the supervising
gynaecologist via sms.
With more trials being conducted successfully,
the application can be adopted by the Government
of India for monitoring labour especially in periph-
eral centres, for training providers and in promoting
self learning etc. It is expected that the easy-to-use
mobile/smart partograph will promote the practice of
plotting partograph during labour by peripheral health
workers which is currently very low. Over a period of
time, plotting of partographs will encourage providers
to use it as a decision-making tool and thereby reduce
morbidity and mortality, both of mother and newborn
arising out of complications of labour and its seque-
Anokwa, Y., Ribeka, N., Parikh, T., Borriello, G., and Were,
M. C. (2012). Design of a phone-based clinical de-
cision support system for resource-limited settings.
In Proceedings of the Fifth International Conference
on Information and Communication Technologies and
Development, ICTD ’12, pages 13–24. ACM.
Densmore, M. (2012). Claim mobile: when to fail a tech-
nology. In Proceedings of the SIGCHI Conference on
Human Factors in Computing Systems, pages 1833–
1842. ACM.
DeRenzi, B., Lesh, N., Parikh, T., Sims, C., Maokla, W.,
Chemba, M., Hamisi, Y., S hellenberg, D., Mitchell,
M., and Borriello, G. (2008). E-imci: Improving pedi-
atric health care in low-income countries. In Proceed-
ings of the SIGCHI Conference on Human Factors in
Computing Systems, CHI ’08, pages 753–762. ACM.
Hogan, M. C., Foreman, K. J., Naghavi, M., Ahn, S. Y., and
Wang, M. (2010). Maternal mortality for 181 coun-
tries, 1980–2008: A systematic analysis of progress
towards millennium development goal 5. The Lancet.
MoTeCH (2011). Mobile technology for community health
in ghana (MoTeCH).
Ollerhead, E. and Barriers, O. D. (2014). Barriers to
and incentives for achieving partograph use in obstet-
ric practice in low- and middle-income countries: a
systematic review. BMC Pregnancy and Childbirth,
The work has been supported by a research grant given
by ICMR. The authors acknowledge the work done by Dr
Karishma and Dr Tavleen at the field sites supervised by
Dr Pratima Mittal, Dr Neelam Aggarwal and Dr AG Rad-
hika. Thanks are also due to reviewers of the project for
their feedback and encouragement.
Pelusi, L., Passarella, A., and Conti, M. (2006). Oppor-
tunistic networking: data forwarding in disconnected
mobile ad hoc networks. IEEE Communications Mag-
azine, 44(11):134–141.
Raleigh, V. S. (2000). Safe motherhood initiatives: Critical
issues. British Medical Journal.
Royston, E. and Armstrong, S. (1989). Preventing maternal
Svoronos, T., Mjungu, D., Dhadialla, P., Luk, R., Zue, C.,
Jackson, J., and Lesh, N. Commcare: Automated
quality improvement to strengthen community-based
Swasthya (2015).
Underwood, H., Ong’ech, J., Appley, M., Rosenblum,
S., Crawley, A., Sterling, S. R., and Bennett, J. K.
(2014). Partopens at the point of care - evaluating
digital pen-based maternal labor monitoring in kenya.
In HEALTHINF 2014 - Proceedings of the Interna-
tional Conference on Health Informatics, pages 90–
100. SciTePress.
Underwood, H., Sterling, S. R., and Bennett, J. K. (2013).
The design and implementation of the partopen mater-
nal health monitoring system. In Proceedings of the
3rd ACM Symposium on Computing for Development,
ACM DEV ’13, pages 8:1–8:10. ACM.
Underwood, H. M. (2013). The Partopen: Using Digital
Pen Technology to Improve Maternal Labor Monitor-
ing in the Developing World. PhD thesis, Boulder,
CO, USA. AAI3607375.
WHO (2015). World health organization maternal health
and safe motherhood programme. The Lancet,
HEALTHINF 2016 - 9th International Conference on Health Informatics