TELKOM
NIKA
, Vol.13, No
.2, June 20
15
, pp. 661 ~ 6
6
9
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i2.988
661
Re
cei
v
ed
No
vem
ber 1
2
, 2014; Re
vi
sed
F
ebruary 28,
2015; Accept
ed March 1
6
, 2015
Simple Screening for High-Risk Pregnancies in Rural
Areas Based on
an Expert System
Retno Supri
y
anti*
1
, Ahmad Fariz
1
, Te
dd
y
Septiana
1
, Eko Murdy
a
ntoro
1
,
Yogi Ramad
h
ani
1
, Haris B. Widodo
2
1
F
a
cult
y
of Sci
ence a
nd En
gi
neer
ing, Je
nde
ral Soe
d
irma
n Univers
i
t
y
2
F
a
cult
y
of Me
dical
and H
e
a
l
th Scienc
es, Jender
al So
erdim
an Un
iversit
y
Jl. HR. Boenj
a
m
in 70
8 Pur
w
o
k
erto, Phon
e: +62-2
81-6
3
5
2
9
2
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: retno_su
p
ri
yanti@
unso
ed.a
c
.id
A
b
st
r
a
ct
T
he hig
h
mate
rnal a
nd infa
nt
mortal
ity rates in
deve
l
op
in
g countri
es, especi
a
lly Ind
o
n
e
sia, ar
e
quite
al
ar
mi
ng.
T
here
ar
e
ma
ny factors t
hat
caus
e
hig
h
mortality
nu
mb
er
s; one
of t
h
e
m
is th
e
del
ay
i
n
han
dli
ng c
a
ses
of hi
gh-risk
pr
egn
anci
e
s. T
h
e
mai
n
pr
obl
e
m
fac
ed
by d
e
v
elo
p
in
g co
unt
ries is th
e l
a
ck
of
hea
lth faciliti
e
s
,
includ
ing
me
dical
equ
ip
me
nt and h
u
ma
n
resources. This rese
arch a
i
ms to d
e
vel
o
p a
simple system that can be
used t
o
screen high-risk
pregnancies. Th
is
system
is based on an
ex
pert
system
. The A
nalytic
al Hierar
chy Process (
A
HP) method
is
used
in making decis
ions about potentially
hig
h
-risk pr
egn
ancy p
a
tients.
Essentia
lly, the
system ca
n
b
e
use
d
by a
n
y
one, a
n
yw
here
,
to carry out e
a
rly
screen
ing of
hig
h
-risk pre
g
nancy p
a
tie
n
ts, so that
dela
ys in the treatment
of thes
e patie
nts can
be
resolv
ed, bec
a
u
se the sy
mpt
o
ms of hi
gh-r
i
s
k
pregn
an
cy a
r
e know
n from the begi
nn
ing.
Results in
dica
te
that this system
shows promis
e for further developm
e
nt.
Ke
y
w
ords
: M
a
terna
l
Morta
lit
y, High
Risk
P
r
egn
anci
e
s, D
e
ve
l
opi
ng Co
u
n
tries,
Exp
e
rt System,
A
nalyt
ical
Hierarc
hy Proc
ess
1. Introduc
tion
Duri
ng
pregn
ancy, it i
s
i
m
p
o
rtant to
mo
ni
to
r foetal
dev
elopme
n
t
con
t
inuou
sly, be
cause it
has
gre
a
t influen
ce o
n
the
health of b
o
th the mothe
r
and h
e
r u
nbo
rn child. Ge
n
e
rally, matern
al
and p
r
e
natal
mortality ca
n
be u
s
e
d
a
s
an indi
cato
r
of the nut
ritional an
d h
ealt
h
statu
s
of t
he
mother, the level of maternal he
alth servic
e
s
an
d the health en
vironme
n
t during pregn
an
cy.
Matern
al Mortality Rate (MMR) i
s
one of
the indica
to
rs to ascertai
n
the health status of wom
e
n.
Curre
n
tly in Indon
esi
a
, the MMR i
s
al
so on
e of th
e target
s tha
t
were
set in
the Millenni
um
Develo
pment
Goal
s; it i
s
i
n
clu
ded
in th
e fifth goal
of
improving m
a
ternal
he
alth
, in which o
n
e
of
the targ
ets to
be a
c
hieve
d
by 2015 i
s
t
o
red
u
ce the
matern
al mo
rtality risk by
three
-
qu
arte
rs.
The death
of a mother is very influential on t
he he
alth and live
s
of child
ren
left behind. If a
mother die
s
,
the children l
e
ft behind
a
r
e pote
n
tially three
to ten
times
more li
kely to die
wit
h
in
two yea
r
s th
a
n
those
who
still have p
a
rents [1]. To
d
a
y, at least 1
8
,000
wom
e
n
die eve
r
y year i
n
Indone
sia
as
a co
nsequ
en
ce of
pre
gna
ncy o
r
chil
dbi
rth. That me
a
n
s that
every
half an
hou
r, a
woma
n
di
es durin
g
p
r
eg
n
ancy or child
birth.
As a
re
sult, every ye
ar 3
6
,000
children
un
der five
become orphans
. This
rate puts
Indones
i
a in firs
t plac
e in ASEA
N for high maternal mortality.
The
Hou
s
eh
old Health S
u
rvey 200
1 report
s
t
hat I
ndon
esi
a's
M
M
R
wa
s 39
6
per
100,0
0
0
live
births. Th
at numbe
r re
prese
n
ted an i
n
crea
se co
m
pare
d
with th
e 1995
surv
ey result
s, which
were 37
3 pe
r 100,000 live
births. T
he
MMR in Ind
o
nesi
a
is eve
n
worse th
an i
n
Vietnam. T
h
e
matern
al mo
rtality rate in that neigh
bo
uri
ng
co
untry
in 200
3 wa
s 95
per
100
,000 live birt
hs.
Among other ASEAN countries,
Malay
s
ia
reco
rded 30 per
100,
000 and
Singapore nine per
100,00
0 [2].
Lack of p
ubli
c
a
w
a
r
ene
ss about m
a
ternal he
alth
is the de
cidin
g
facto
r
in
mortality,
althoug
h ma
ny other fa
ctors
mu
st be
con
s
id
ere
d
in
addressin
g
this issu
e. Mo
st (60% to 8
0
%)
matern
al dea
ths are ca
use
d
by bleedin
g
during
chil
d
b
i
rth, obstructe
d labou
r, sep
s
is, hig
h
bloo
d
pre
s
sure in
p
r
egna
ncy
and
compli
catio
n
s from
un
safe
abortio
n
, a
s
shown in
Fig
u
re 1. T
he
grap
h
in Figure 1
shows the p
e
rcenta
ge
di
stri
bution of cau
s
e
s
of maternal mortality. Based
on the
s
e
data, there
a
r
e three mai
n
facto
r
s th
a
t
cau
s
e
mat
e
rnal mortalit
y:
bleeding, hyperten
s
io
n
or
pre
-
e
c
lamp
si
a durin
g preg
nan
cy, and in
fection [3].
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 661 – 66
9
662
In the m
edi
ca
l field, if a
pat
ient ha
s
a hi
g
h
-ri
sk
p
r
eg
na
ncy, the
r
e
are vario
u
s te
sts o
r
pro
c
e
d
u
r
es
in addition to
routine pren
atal scree
n
in
g test
s, depe
nding on the
circum
stan
ces. A healthcare
p
r
o
v
id
er
migh
t r
e
co
mmend
te
s
t
s
s
u
c
h
a
s
: (
a
)
Sp
ec
ia
lis
ed
o
r
tar
g
e
t
e
d
u
l
tr
as
oun
d
:
thi
s
type of
foetal ultra
s
o
und
- a
n
im
aging te
ch
ni
que th
at
use
s
hi
gh-f
r
equ
e
n
cy sou
nd waves
to prod
uce
image
s of a baby in the uterus
- target
s a su
sp
e
c
te
d probl
em, such a
s
abn
ormal develop
ment;
(b) Amnio
c
en
tesis:
duri
ng
this p
r
o
c
ed
ure, a
sa
mpl
e
of
the
fluid
th
at
su
rround
s and prote
c
ts a
baby du
ring
preg
nan
cy (a
mniotic fluid
)
is with
dr
a
w
n
from the ute
r
us; typically
done
after week
15 of
pregn
a
n
cy, am
nio
c
e
n
tesi
s
can
id
entify ce
rtai
n
gen
etic con
d
itions a
s
we
ll as n
eural t
ube
defect
s
- seri
ous abnorm
a
l
i
ties
of
the
brain or spinal
cord; (c)
Chorioni
c Villus S
a
mpling
(CVS):
durin
g thi
s
p
r
oce
dure, a
sample
of cell
s i
s
removed
from the
pla
c
enta; typi
cal
l
y done
between
wee
k
s 10 an
d 12 of pregn
ancy, CVS can identify
ce
rtain gen
etic
con
d
ition
s
; (d
) Co
rdo
c
ente
s
is
:
this test, also known as
percutaneous umbilical
blood sam
p
ling,
i
s
a highly specialised prenatal
test in
which
a foetal blo
o
d
sam
p
le i
s
re
moved
fro
m
the um
bilical
cord; typically done
after
we
ek
18 of pregn
a
n
cy, the test can ide
n
tify chrom
o
somal
con
d
ition
s
, blood disorders and infe
ctio
ns;
(e) Cervi
c
al length mea
s
u
r
eme
n
t
:
a
he
althca
re provider might u
s
e ultrasoun
d to measu
r
e the
length
of a
p
a
tient’s
ce
rvi
x
at
prenatal
app
ointme
nts to
dete
r
mi
ne
wheth
e
r she i
s
at ri
sk of
prete
r
m lab
o
u
r; (f)
Lab te
sts: a h
ealth
care p
r
ovid
er
might take
a
swab of vagi
nal secretion
s
to
che
c
k for foe
t
al fibrone
ctin
, a sub
s
tan
c
e that act
s
like a gl
ue b
e
twee
n the foe
t
al sa
c and t
h
e
lining of the
uteru
s
; the p
r
esen
ce of fo
etal fi
bron
ecti
n might be a
sign of p
r
et
erm lab
o
u
r
: (g)
Biophysi
cal p
r
ofile: this pre
natal test is u
s
ed to
che
ck
on a baby'
s
well-bei
ng. The
test combin
e
s
foetal heart rate monitori
n
g
(non
stress test) an
d foeta
l
ultraso
und [
4
].
Figure1. Ca
u
s
e
s
of Matern
al Mortality (source: He
alth
Depa
rtment, 2011
)
As reg
a
rds t
he test procedures
discu
s
sed,
these
are
certai
nly not a probl
em if a
preg
nant
wo
man lives in
area
s
whe
r
e full health
servi
c
e
s
are provid
ed, inclu
d
ing h
u
m
an
resou
r
ces
an
d health
equi
pment. Howe
ver, they will
be a
pro
b
lem
for p
r
eg
nant
wome
n
who li
ve
in rural areas with vari
ous
limitations on existing
heal
th facilities. T
h
is
research
aims to
devel
op
a scre
enin
g
system fo
r e
a
rly dete
c
tion
of the ex
istence of a si
mple hig
h
-risk pregn
an
cy in a
preg
nant
wo
man. In thi
s
rese
arch
we
chose to
devel
op a
screeni
n
g
sy
stem b
a
sed o
n
an
exp
e
rt
sy
st
em,
be
ca
use
su
ch a
sy
st
em h
o
lds
cert
ai
n
adva
n
tage
s. An expert sy
stem
is a com
put
er
system that i
s
equal to the ca
pabilities of an expert in deci
s
ion
m
a
king. The word ‘equal’ has a
sen
s
e
that t
he exp
e
rt
sy
stem i
s
expe
cted to
work
in all ca
se
s
as well as
an expert. The
advantag
es
of an expert
system
are: (a) expe
rt advice i
s
available all th
e time; (b) the
kno
w
le
dge
of
expe
rt staff
can
be
captu
r
ed
to
some
extent befo
r
e
they move
o
n
; (c) it can
be
us
ed as
a training aid to inc
r
eas
e
the
expertise
of staff; (d) it ma
k
e
s
rational dec
is
ions
without
emotional
ov
erhe
ad
s; (e) i
t
doe
s
not g
e
t
tired
or ov
e
r
worke
d
; (f
) it
is a
n
efficient
way
of g
e
tting
answers,
as i
t
does
not inv
o
lve
additional hel
p
staff; and
(g) a na
tural
language interface would
make th
e ex
pert
system
more
huma
n
friendly [5
]. As re
gards t
he adva
n
tag
e
s of a
n
exp
e
rt
system,
th
ere
have bee
n some re
sea
r
ch
p
r
oje
c
ts
on
appli
c
atio
ns
in medi
cal
di
agno
si
s. Hasan
[6] carried
ou
t rese
arch
on
the diag
no
sis of hum
an d
i
sea
s
e
s
u
s
in
g a fuzzy ex
pert sy
stem.
His
resea
r
ch pro
j
ect focu
se
s on the research an
d developme
n
t of a web-b
a
s
ed cli
n
ical tool
Bleeding
Ecla
m
s
ia
In
fection
Ab
ort
i
on
Obst
ructe
d
lab
o
r
Em
bol
i
obst
pue
r
p
u
r
eum
ot
he
rs
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Simple Sc
reening for High-Ris
k
Pregnanc
ies
in Rural
Areas
Bas
e
d on .... (Retno Supriy
anti)
663
desi
gne
d to improve the
quality of th
e exchan
ge
of health informatio
n betwee
n
health
care
profe
ssi
onal
s and patients. Prasadl [7] carrie
d out re
sea
r
ch on an
appro
a
ch to developing
an
expert sy
ste
m
in medi
cal
diagno
si
s u
s
ing a m
a
chi
ne lea
r
ning
algorith
m
. He con
s
id
ered
the
dise
ase a
s
th
ma for
diagn
osi
s
. Mitra [8]
develop
ed a
fuzzy M
L
P
model
-ba
s
ed
expert
syste
m
for
medical diag
nosi
s
. It is used a
s
a co
nne
cti
oni
st expert system
for diagno
si
ng hepato
b
ili
ary
diso
rde
r
s. It
can h
andl
e u
n
ce
rtainty an
d/or imp
r
eci
s
i
on in the inp
u
t as well a
s
the output. Shah
[9] develope
d
an exp
e
rt
system fo
r dia
g
nosi
s
of
sk
in
dise
ase u
s
in
g an A
r
tificial
Neu
r
al
Net
w
ork
(ANN). Schat
z [10] provid
ed an ov
e
r
vi
ew of the
rol
e
of intellige
n
t tools in m
odern he
alth
care
system
s a
n
d
reviewed
cu
rre
nt ch
allen
ges i
n
the
fie
l
d. Iantovics [
11]
develo
p
e
d
a novel l
a
rge-
scale hybrid
medical
di
ag
nosi
s
system
call
ed
L
M
DS
.
The
L
M
DS
sy
st
em co
mp
rise
s phy
si
cia
n
s,
medical exp
e
r
t syste
m
ag
ents
and
me
dical I
C
MA
a
gents. M
edi
cal ICMA a
g
e
n
ts rep
r
e
s
ent
a
novel cl
ass
of agent
s
with the ICMA
architectu
re.
The di
agn
osi
s
system
ca
n solve
difficult
medical dia
g
nosi
s
p
r
obl
e
m
s who
s
e
sol
u
tions m
u
st
b
e
discove
r
ed
coo
peratively by the memb
ers
of the syste
m
. Zahra
n
i [1
2] applied
an
expert sy
ste
m
for proph
e
t
ic medi
cine t
o
bre
a
st
can
c
er
diagnosi
s
and treatment. Haiji [13]
developed an account of a rule-b
ased expert system (RB
ES)
for ne
urol
ogi
cal di
so
rde
r
s, i.e., Alzheim
e
r’s, Pa
rkin
son’s, Huntin
g
t
on's di
sea
s
e
,
cerebral pal
sy,
meningiti
s, e
p
ilep
s
y, multiple
scl
ero
s
i
s
, stroke, cl
ust
e
r
head
ache,
migraine
an
d me
ningitis
for
child
ren. Mo
re than 10 types of ne
uro
l
ogica
l disea
s
e can be di
agno
se
d and
treated by his
system. Fue
r
bach
[14] de
scrib
e
d
a
syst
ematic metho
d
for examini
ng p
ubli
c
kno
w
led
ge fo
und
in
health
care t
e
xtbooks a
n
d
practi
ce g
u
ideline
s
su
rroun
ding th
e
con
c
e
p
t of
oral fe
edin
g
in
prem
ature i
n
fants in a n
e
o
natal inten
s
ive ca
re
unit. It inclu
d
e
s
the developm
ent
of an instrum
ent
for extractin
g
data from tho
s
e source
s to
standa
rdi
s
e
definition
s
of terminol
ogie
s
.
On the other hand, our p
r
eviou
s
re
se
arch
co
ncent
rates o
n
developin
g
low-cost and
easy-to
-u
se t
e
ch
nolo
g
y to sup
p
o
r
t diag
nosi
s
in
the
medical field
based
on i
m
age
pro
c
e
s
sin
g
techni
que
s [1
5] [16] [17] [18]
[19]. Other research
ers
also u
s
e
im
a
ge processin
g
techni
que
s
for
developin
g
compute
r
ai
de
d dia
gno
si
s [
21] [22].
Ho
wever, the
we
akn
e
sse
s
i
n
t
he u
s
e
of i
m
age
pro
c
e
ssi
ng te
chni
que
s that
we have
pre
v
iously
devel
oped
are th
at both ha
rd
wa
re an
d softwa
r
e
specifications used are
sometimes
not
compatible for rura
l areas. In
this
p
aper, we
will di
scuss
the develo
p
m
ent of a
si
mple
scree
n
i
ng sy
stem
fo
r hig
h
-risk p
r
egna
nci
e
s
ba
sed
on
an e
x
pert
system th
at is very e
a
sy t
o
implem
ent
in ru
ra
l a
r
ea
s. We em
pha
sise the
use o
f
the Analytical
Hierarcy Pro
c
ess (AHP
) m
e
t
hod for dev
elopin
g
our
system.
2. Rese
arch
method
Basically, the
diagn
osi
s
of
a di
sea
s
e
pro
c
e
s
s will
be
d
i
vided into t
w
o condition
s,
namely,
clini
c
-b
ased
and
comm
un
ity-base
d
. In
the cli
n
ic-ba
s
ed
co
ndition
, the diag
no
sis i
s
ba
sed
on
laboratory tests and exa
m
ination
s
are cond
ucte
d
by medical p
e
rsonn
el dire
ctly, while in the
comm
unity-b
ase
d
co
nditio
n
, the diagno
sis i
s
ba
s
ed
on the sympt
o
ms that ca
n
be felt by th
e
patients the
m
selve
s
. In this re
se
arch
we develop
bo
th of the con
d
itions fo
r screenin
g
high
-risk
preg
nan
cie
s
. The first is d
edicated for
use by health
worke
r
s in formal he
alth servi
c
e
s
unit
s
in
orde
r to sho
r
t
en the time of the initial examinat
ion. Th
e se
con
d
con
d
ition is dedi
cated for grou
ps
of ordin
a
ry pe
ople wi
dely a
v
ailable in de
veloping
co
u
n
tries
su
ch a
s
Indon
esi
a
in ord
e
r to p
r
o
v
ide
first aid in initial scree
n
ing
of high-risk pregna
nci
e
s in
rural are
a
s.
In our scree
n
i
ng syste
m
, we u
s
e the A
nalytical
Hi
erarchy Pro
c
e
s
s (AHP
) met
hod to dete
r
mine
an altern
ative conditio
n
in preg
nant pati
ents.
2.1. Analy
t
ic
al Hierarchy
Process
(AHP)
This m
e
thod
is a frame
w
ork fo
r eff
e
ctiv
e de
ci
si
on ma
king
o
n
co
mplex i
s
sue
s
by
simplifying a
n
d
accel
e
ratin
g
the p
r
o
c
e
s
s of deci
s
io
n
makin
g
to resolve the p
r
obl
em into its
pa
rts,
arrangi
ng
parts o
r
va
riabl
e
s
in
a
hie
r
a
r
chy of n
u
me
rical value
ba
se
d on
the
subj
ective ju
dgm
ent
of the import
ance of ea
ch variable
a
nd synthe
si
se the variou
s co
nsi
d
e
r
ati
ons to d
e
termine
whi
c
h va
riabl
es h
a
ve the
h
i
ghe
st pri
o
rity and
act to
affect the
outco
me of the
situ
ation. The A
H
P
method help
s
to solve co
mplex
proble
m
s
by stru
ct
uring
a hi
era
r
chy
of criteri
a
and
interested
partie
s
, with intere
sting
re
sults
an
d co
nsid
eratio
ns for
d
e
velopin
g
weight
s
o
r
pri
o
ritie
s
.
T
h
is
method
also
combi
n
e
s
the
stre
ngth
of feeling
and
lo
gic
con
c
e
r
n
e
d
on va
riou
s
issue
s
, and
then
synthe
sises
a diverse ra
nge of co
nsi
deratio
ns
to
be matched
with the re
su
lts we intuitively
estimate a
s
p
r
esented in th
e con
s
id
erati
ons that have
been mad
e
[3] [20].
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 661 – 66
9
664
2.1.1. Decom
position
De
comp
ositio
n
is solving
the
problem
by
divi
ding it
into a n
u
m
ber
of more
spe
c
ific
variable
s
. In t
he AHP, the
decompo
sitio
n
process
i
s
defined
as th
e prepa
ration
of the hie
r
a
r
chy
of criteri
a
, su
b-criteri
a
and
alternatives
rela
ted to th
e probl
em to
be solved [
3
] as sh
own
in
Figure 2.
Figure 2. Hierarchical structure whi
c
h i
s
a tr
ansl
a
tion
probl
em (So
u
rce: Sup
r
iyan
ti R, 2013)
By using t
he
comp
arative judgme
n
t pri
n
ciple,
we
det
ermin
e
p
r
iorit
i
es that
will b
e
u
s
ed
as the score f
o
r each sym
p
tom that will be used in
our
screening
system as shown in Table 1.
Table 1. AHP
pairwi
s
e
com
pari
s
on of a
s
se
ssm
ent sca
l
es (So
u
rce: Supriyanti, 20
13)
2.1.2. Priority
s
y
nthesis
Priority synth
e
si
s is do
ne
by
multiplying the local prio
rities a
n
d
the prioritie
s
of the
relevant
crite
r
ia at that l
e
vel and
addi
ng this to
ea
ch el
eme
n
t in the level th
at influen
ced
th
e
c
r
iteria [3].
Figure 3. System De
sign (S
ource: Supri
y
anti R, 2013
)
Priorit
y
scale
Definition
Explanation
1
Equally
impo
rtan
t
Both activities ha
ve the same contribution to the obj
ect
3
Some
w
hat mo
re
important
Experience an
d
assessment sho
w
s that t
he activity of the
rather m
o
re impo
rtant than t
he oth
e
rs.
5 Quite
important
Experience an
d
assessment sho
w
s that
one activit
y
is more
important than t
h
e others
7 Ver
y
Imp
o
rtant
Activity
compare
d
w
i
th acti
vit
y
ha
s dominance over the
other.
9
Ver
y
ver
y
impor
t
ant
Activity
that one i
s
really
impo
rtant
and influential than other
activit
i
es.
2,4,6,8
The midpoint bet
w
e
en t
w
o a
d
jacent values decisio
n.
Contra
r
y
When the activity of "I" has a high
er value of activit
y
"j" the
n
"j" has the oppos
ite value w
hen c
o
mpared
w
i
th t
h
e
"I".
Ratio
Value / ratio is obtained
directl
y
fr
om measuremen
ts.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Simple Sc
reening for High-Ris
k
Pregnanc
ies
in Rural
Areas
Bas
e
d on .... (Retno Supriy
anti)
665
2.2. Sy
stem
design
Acco
rdi
ng to
the a
bove
explan
ation
,
we devel
oped
ou
r scre
enin
g
system by
impleme
n
ting
the Analytical Hierarchy
Proce
s
s.
The first sta
g
e
of the AHP is Structuring;
stru
cturi
ng th
e flow
of de
ci
sion
ma
king i
s
ba
se
d
o
n
t
w
o m
a
in
com
pone
nts; the f
i
rst
com
pone
nt is
the purp
o
se of the AHP and the vari
able
s
us
ed, while the
se
con
d
com
p
o
nent com
p
ri
ses
alternative
s
that can
be ta
ken to fulfil th
e purpo
se
of
the AHP. In phase st
ructu
r
i
ng, the purpo
se
of the A
H
P
will be
dete
r
mi
ned, a
s
well
as
wh
at va
ri
a
b
les an
d
sub
-
variable
s
are
use
d
a
nd
wh
at
alternative
s
a
r
e available.
Structu
r
ing th
e AHP
proce
ss i
s
the prep
aration of the
AHP framework
con
s
i
s
ting of
the main aim, variable
s
use
d
as
con
s
ide
r
ation a
n
d
the alterna
t
ives that ca
n be
taken
to m
e
et the g
oal
s. The
next p
hase of th
e
AHP is Asse
ssment, i.e.
stage
sco
r
in
g o
r
weig
hting of the variabl
es,
sub
-
vari
able
s
and al
tern
atives. Figure 3 sho
w
s our d
e
s
ign
system.
2.3. Compar
ison s
y
stem
using Super
D
ecision V2.
2
In ord
e
r to
evaluate the
accu
ra
cy of
our
screeni
ng system, we
th
en con
ducte
d
a
comp
ari
s
o
n
a
gain
s
t othe
r software
an
d a
l
so u
s
in
g
real
field data. S
uperde
cisi
on
is softwa
r
e th
at
impleme
n
ts t
he Analytic
Network Pro
c
e
s
s (A
NP)
and An
alytic Hie
r
a
r
chy Process
(AHP
) for
deci
s
io
n maki
ng.
2.4. Compari
s
on s
y
stem using real fi
eld data
In ord
e
r to
o
b
tain a
c
curate pe
rform
a
n
c
e, in ad
dition
to co
mpa
r
in
g the
re
sults of the
system
which
we
ma
de
wit
h
the
re
sult
s
obtaine
d
by SuperDe
ci
sio
n
V2.2, we al
so co
mpa
r
ed
the
results with result
s
of dia
gno
stic syste
m
s
in
real
condition
s th
a
t
a do
ctor o
r
midwife
wo
uld
operate und
er. In orde
r
to obtain co
mparative
da
ta on the re
al con
d
ition
s
, we dist
ribut
ed
que
stionn
aire
s to
pregna
n
t
wom
en p
a
tients, a
nd
asked
them
to
reply to
the
same
qu
estio
n
s
usin
g o
u
r
system, and th
e
n
a
s
ked th
eir do
ctor
or
mi
dwife to
dia
g
nose the
sam
e
patie
nts. T
hen
we compa
r
e
d
the diagno
sti
c
re
sult
s to ev
aluate the p
e
rform
a
n
c
e o
f
our system.
3. Results a
nd analy
s
is
As already
descri
bed i
n
sub
-
sectio
n
2
above, we built the system ba
sed
on two
con
d
ition
s
,
n
a
mely clinic-based and
community-
b
a
s
ed. The
det
ailed clini
c
-b
ase
d
system
ha
s
alrea
d
y been
discu
s
sed in [21], in which we a
c
hieve
d
perfo
rman
ce
of 80%. In this pap
er, we wil
l
discu
s
s the
developm
ent
of ou
r
previous res
earch
by devel
opi
ng
comm
unit
y
-based
sy
stems.
One of th
e a
d
vantage
s in
the u
s
e of a
comm
unity
-b
ase
d
sy
stem
is that the
de
velopment of
a
kno
w
le
dge b
a
se i
s
not as
compl
e
x as the clini
c
-ba
s
e
d
system
s [3] as sh
own in Table 2.
Table 2. Kno
w
led
ge Base of Symptoms have Pursue
Code
S
y
mptom
A
Positive pregnancy
t
e
st
B
Is age less than 18
y
e
a
r
s
C
Is age more tha
n
40
y
e
a
r
s
D
Does feel a fe
w
e
r
E
Does feel nause
ous
F
Does feel dizzy
G
Whether abd
ominal pain
H
Whether blood p
r
essure decrease
d
I
Whether blood p
r
essure increased
J
w
h
ether
rapid he
artbeat an
d short
ness of breath
K
Whether have di
abetic histor
y
L
Whether have h
y
pertension histor
y
Acco
rdi
ng to
our p
r
eviou
s
resea
r
ch,
we m
ade
five predi
ction
s
of po
ssible
cau
s
e
s
of
matern
al mortality, namely Eclampsi
a Hyperte
ns
i
o
n
,
Pre-ecl
a
mp
sia Hype
rten
sion, Intrap
artum
Infection, Post Partum Bleedin
g
, Other Disea
s
e an
d
Normal [3]. But in real condition
s, if it is
applied to the ordi
nary community
it will be
quite
confusing. T
herefore i
n
this research, we
summ
ari
s
ed
our p
r
edi
ction
s
into two con
d
iti
ons o
n
ly: normal a
nd hig
h
-ri
sk pre
gna
ncy.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 661 – 66
9
666
Referrin
g to Table 2,
we co
uld e
s
ta
blish a
relati
onship bet
ween di
sea
s
e
s
du
ring
preg
nan
cy an
d the sympto
ms that acco
mpany them, as seen in Ta
ble 3.
Table 3. Rel
a
tionshi
ps b
e
twee
n Symptoms and P
r
eg
nan
cy Disea
s
es
CODE
PREG
NANCY
DI
SEASES
High-Risk
Pregn
anc
y
Normal
A
V
V
S B
V
Y C
V
M D
V
V
P E
V
V
T F
V
V
O G
V
V
M H
V
V
S I
V
S J
V
K
V
V
L
V
V
After a com
pari
s
on of p
r
iority sympt
o
ms
to sym
p
toms was
obtaine
d, su
bse
que
nt
pairwise
com
pari
s
on
s
we
re cond
ucte
d
to d
e
term
in
e alte
rnative
pri
o
rity di
se
ase
to
symp
tom
crite
r
ia. Table
4 describ
es t
h
is pri
o
rity.
Table 4. Co
m
pari
s
on b
e
tween Sympto
m and Symptoms Prio
rity
1
1
1
0
0
1 1 0
1
1
1
1
A B
C
D
E
F
G
H
I
J
K
L
Pv
1
A
1.00
1.00
0.33
0.00
0.00
0.20
0.33 0.00 0.33
0.33 0.33
0.20
0.03
1
B
1.00
1.00
0.33
0.00
0.00
0.20
0.33 0.00 0.33
0.33 0.33
0.20
0.03
1
C
3.00
3.00
1.00
0.00
0.00
0.33
1.00 0.00 1.00
0.33 1.00
0.33
0.08
0
D
0.00
0.00
0.00
0.00
0.00
0.00
0.00 0.00 0.00
0.00 0.00
0.00
0.00
0
E
0.00
0.00
0.00
0.00
0.00
0.00
0.00 0.00 0.00
0.00 0.00
0.00
0.00
1
F
5.00
5.00
3.00
0.00
0.00
1.00
3.00 0.00 3.00
3.00 3.00
1.00
0.24
1
G
3.00
3.00
1.00
0.00
0.00
0.33
1.00 0.00 1.00
1.00 1.00
0.33
0.09
0
H
0.00
0.00
0.00
0.00
0.00
0.00
0.00 0.00 0.00
0.00 0.00
0.00
0.00
1
I
3.00
3.00
1.00
0.00
0.00
0.33
1.00 0.00 1.00
1.00 1.00
0.33
0.09
1
J
3.00
3.00
3.00
0.00
0.00
0.33
1.00 0.00 1.00
1.00 1.00
0.33
0.11
1
K
3.00
3.00
1.00
0.00
0.00
0.33
1.00 0.00 1.00
1.00 1.00
1.00
0.11
1
L
5.00
5.00
3.00
0.00
0.00
1.00
3.00 0.00 3.00
3.00 1.00
1.00
0.22
9
Jml
27.00
27.00
13.67
0.00
0.00
4.07
11.67
0.00 11.67
11.00
9.67
4.73
0.00
Principle Eigen V
a
lue
9.31
Consistency
In
de
x
0.04
Acco
rdi
ng to Table 4, the first ro
w and t
he fi
rst col
u
m
n
are pri
o
rity multiplier sy
mptoms
with other
symptoms. If the cell in both
the first ro
w and the first
colum
n
is bla
n
k, this indi
cates
that the cell i
n
inactive, while if the cell i
s
filled
in, this indicates
that t
he cell i
s
a
c
tive. Accordi
ng
to Table
4, it
app
ears th
e
r
e a
r
e
several symp
to
ms that have
a
domin
ant p
r
iority; these
are
symptom
s
F and L, wh
ose
PV values are 0.24 and 0.
22.
In orde
r to d
e
velop a l
o
w-co
st an
d ea
sy
-to-use screenin
g
sy
st
e
m
, we u
s
ed
Microsoft
Excel to buil
d
our
syste
m
. The adv
antage
s of
usin
g Micro
s
oft Excel are that almost all
comp
uters, b
o
th de
skto
p
s
and lapt
o
p
s,
have the Mi
crosoft Excel a
pplication. Th
is appli
c
atio
n
is
also
ea
sily operate
d
by e
v
eryone. Th
e
s
e a
d
v
antag
es i
s
eno
ugh
for buildi
ng
a low-cost a
n
d
easy-to
-u
se
scre
enin
g
sy
stem. Figure 4
sho
w
s an
ex
ample of the
displ
a
y
of patient data entry,
while Fig
u
re
5 sho
w
s an
example of the disp
lay of symptom da
ta entry. Figure 6 sho
w
s an
example of the display of diagno
si
s re
su
lts. In order to evaluate
the performan
ce of our syste
m
,
we al
so comp
are the dia
g
n
o
si
s re
sults
with analyse
s
u
s
ing Sup
e
rDe
c
isi
on. First we develop
ed
a
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Simple Sc
reening for High-Ris
k
Pregnanc
ies
in Rural
Areas
Bas
e
d on .... (Retno Supriy
anti)
667
hiera
r
chy st
ructure u
s
in
g
SuperDe
ci
sio
n
as sh
o
w
n
i
n
Figu
re 7.
T
hen we deve
l
oped a
p
r
iori
ty
desi
gn an
d filled in the
prio
rity values. A
ll pro
c
e
s
se
s used in
stru
ction
s
e
m
bedd
ed wit
h
in
SuperDe
ci
sio
n
Figure 4. Display of Patient Entry Data
Figure 5. Display of Symptoms Entry Data
Figure 6. Display of Diagn
osi
s
Re
sult
Acco
rdi
ng to Figure 6, our
system al
so p
r
ov
ide
s
advice on wh
at first aid shoul
d b
e
given
by the use
r
. For exampl
e, Hartini i
s
a preg
nan
cy
patient. She had sym
p
toms of hig
h
-risk
preg
nan
cy b
e
ca
use the system
sho
w
ed symptom
s
that identify
the directi
on of a high
-ri
sk
preg
nant pati
ent at the age
of 45 years.
Figure 7. Hierarchy Schem
e usin
g Supe
r De
cisi
on
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 661 – 66
9
668
A
cco
rdi
ng t
o
t
h
is
ca
se,
ou
r
sy
st
em
will
p
r
ov
ide a
n
initia
l treatme
nt by
re
com
m
en
di
ng h
e
r
to have
suffi
cient
nutritio
n
an
d
re
st to imp
r
ov
e
h
e
r
co
ndition.
In an
othe
r
ca
se, o
u
r sy
stem
recomme
nde
d that a
pati
ent take vitamins
as
ea
rl
y treatment.
Acco
rdi
ng to
Figure 7,
after
finishin
g
the
Hierarchy
S
c
heme usi
ng SuperDe
ci
si
o
n
, we
then
d
e
velope
d rating p
r
io
rities
as
descri
bed in
Figure 8.
Figure 8. Super De
ci
sion
Rating
After ente
r
ing
wei
ghting
va
lues in th
e p
r
ioriti
es table
as
de
scribe
d
in Fi
gure 8,
we
ca
n
run the p
r
og
ram and get th
e results a
s
d
e
scrib
ed in Fi
gure 9.
Figure 9. Re
sults of SuperDe
cisi
on
Acco
rdi
ng to
Figure 9, we
achi
eved a
p
e
rform
a
n
c
e
rate of 79%,
while i
n
ou
r
system we
got 80% pe
rf
orma
nce. Thi
s
mea
n
s th
at our
system
is quite
simil
a
r to the exi
s
ting
system
in
SuperDe
ci
sio
n
. The
r
efore
our sy
stem
is p
r
omi
s
in
g for i
m
ple
m
entation i
n
real
conditi
ons,
esp
e
ci
ally for ordin
a
ry pe
op
le.
4. Conclusio
n
Our sy
stem is low-cost a
nd easy for
every
one to use, even if they are not from the
medical field.
Ou
r
system
is
also
com
patible
fo
r u
s
e on
all type
s of
co
mpute
r
s be
cau
s
e
t
h
e
softwa
r
e
wa
s develope
d u
s
ing
only Excel, whe
r
e
we
can b
e
sure
that all com
puters have t
h
e
Excel tool, so
it is ea
sy to
acce
ss th
e
system.
Ho
we
ver, even tho
ugh it i
s
a
si
mple
system,
our
system
can
still be develo
ped to be a
better
sy
ste
m
and can e
v
en be devel
oped fo
r mo
bile
-
bas
ed d
e
v
i
ce
s.
Ou
r
sy
st
e
m
is
als
o
a
c
cur
a
t
e
.
Th
i
s
has ha
s b
e
e
n
proven
by the fact th
at ou
r
system
ha
s
q
u
ite simil
a
r p
e
rform
a
n
c
e t
o
the
exis
ting
syste
m
in
SuperDe
ci
sion.
Refe
rri
ng to
the
advantag
es o
f
our syste
m
, we con
c
lud
e
that our
sy
ste
m
is pro
m
isi
n
g for implem
e
n
tation in ru
ral
area
s a
s
a si
mple screeni
ng syste
m
for high
-ri
sk
pregna
ncy. It is expe
cted t
hat by using
this
system, we can improve the quality of life for pr
e
gna
n
t
women in th
e developin
g
cou
n
trie
s.
Ackn
o
w
l
e
dg
ment
T
his work is suppo
rted b
y
Directo
r
ate
G
eneral of High
er Edu
c
ation throu
g
h
Hibah
Ung
gulan Pe
rguru
an Tin
ggi
with cont
ract
numbe
r 121
4/UN23.10/P
N
/201
4.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Simple Sc
reening for High-Ris
k
Pregnanc
ies
in Rural
Areas
Bas
e
d on .... (Retno Supriy
anti)
669
Referen
ces
[1]
Dinas Kesehat
an Ja
w
a
tengah.
Profil Kesehatan
. Yearl
y
R
eport. 201
1.
[2] Sis
w
ono.
K
e
matia
n
Ibu
Indon
esia
T
e
rtingg
i
di Asea
n
, http://
w
w
w
.
gizi.net/cgi-
bin/b
e
rita/full
ne
w
s
.cg
i
?
n
e
w
s
i
d
106
24
857
36,7
903
8.
[3]
Supri
y
a
n
ti R,
T
edd
y
S, Eko
M, Haris BW
.
Analytic
al H
i
er
archy Proc
ess
for Estimati
ng
Pregn
ancy
Diseas
es
. Proc
eed
ing
of Internatio
nal
Conf
e
r
ence o
n
Bi
omedic
a
l Sci
ence
and En
gi
neer
i
ng, Istanbu
l,
T
u
rky
,
Septem
ber. 201
4.
[4] May
o
C.
Hi
gh-risk pre
g
n
ancy: Know
w
hat to e
x
pect
, http://
www
.
m
ay
oclinic.org/healthy
-
livin
g/pre
g
n
anc
y-
w
e
ek-b
y-
w
e
e
k
/in-de
pth/hi
gh
-risk-
preg
na
nc
y/art 2
0
0
4
7
0
1
2? F
o
otprints
=
mine
,
accessed by
March 6
th
. 201
4.
[5] T
each
ICT
-
Com.
Advanta
ges
of
Expert S
ystem
, ht
tp://
w
w
w
.
t
each-
ict.com/as_a2
_
i
ct_ne
w
/
ocr/A2
_G063/3
3
4
_
a
p
p
licati
ons
_ict/e
xp
ert_s
y
stems/
mini
w
e
b/p
g8.ht
m,
accessed by
M
a
rch 6
th
. 2014.
[6]
Hasa
n MA, Khaja Md. Sher-E
-Alama
nd Ahs
an Ra
ja Ch
o
w
dhur
y. Hum
an
Diseas
e Di
agn
osis Usin
g
a
Fuzz
y
E
x
pert Sy
stem.
Jour
nal of
Computi
ng.
201
0; 2(6): 66-
70.
[7]
Prasadl BDCN, PESN Krishna P, Y Sagar. An A
pproac
h
to Dev
e
lop
Ex
pert S
y
stem
in Medical
Diag
nos
is Usi
ng Mach
ine L
earn
i
ng Al
gorit
hm (AST
HMA) and Pe
rform
ance Stud
y.
International
Journ
a
l of Soft Co
mp
uting (IJ
S
C).
2011; 2(1
)
: 26-33.
[8]
Mitra S. F
u
zz
y
MLP b
a
se
d E
x
pert S
y
stem f
o
r Medic
a
l
Di
ag
nosis.
F
u
zz
y
S
e
ts an
d Syste
m
.
19
94
; 65
:
285-
296.
[9]
Shah T
P
, Pooja JS. Connecti
onist E
x
pert S
ystem fo
r Medical Dia
gn
osis u
s
ing ANN
–
A case stud
y o
f
skin dis
eas
e Scabi
es.
Internatio
nal J
our
n
a
l of Adva
nce
d
Rese
arch i
n
Co
mp
uter
Scienc
e an
d
Softw
are Engin
eeri
ng.
20
13; 3
(
8): 227-2
30.
[10]
Schatz CV, F
abi
o KS.
In
te
l
l
i
g
en
t a
n
d
Expe
rt Syste
m
s
in
Me
di
ci
ne
– Are
v
i
e
w
. XVI
II Congres
o
Argenti
no d
e
Bioin
g
e
n
ierí
a SABI 2011 - VII J
o
rna
das
d
e
Ing
eni
ería Clínic
a
Mar del Pl
ata. 201
1.
[11]
Lantov
ics BL. Agent-Bas
ed
Medic
a
l Di
ag
n
o
sis S
y
stem.
Co
mp
uting an
d
Informatics.
200
8; 27: 59
3-
625.
[12]
Z
ahran
i NMA.
Breast Ca
ncer
Diag
nos
is a
nd
T
r
eatment of Proph
et
ic Med
i
ci
ne Us
ing E
x
pe
rt S
y
stem.
Journ
a
l of Infor
m
ati
on & Co
mmu
n
ic
ation T
e
c
hno
logy.
2
010;
4(2): 20-26.
[13] Hajji
AAA.
Ru
le-Bas
ed Exp
e
rt System fo
r Diag
nos
is
a
nd Sy
mpto
m
of Neur
olo
g
ica
l
Disor
de
r
s
Neur
olo
g
ist Ex
pert Syste
m
(
N
ES).
Proce
e
d
i
ng
of Inter
nati
ona
l C
onfer
en
ce o
n
C
o
mmu
nicati
ons
an
d
Information T
e
chnology
(ICCI
T
)
, Hamma
met,
T
unisian, Jun
e
26-2
8
. 201
2.
[14]
F
euerh
a
ch
RD
, T
e
resa LP.
Buil
din
g
an
E
x
pert S
y
stem
A
S
y
stematic A
ppro
a
ch
to D
e
velo
pin
g
an
Instrument for Data Ex
trac
tion From the Literature.
Inter
natio
nal
Jour
n
a
l of N
u
rsin
g
Care Qu
ality.
200
3; 18(2): 12
9-13
8.
[15]
Supri
y
a
n
ti R, Hitoshi
H,
M
a
satsugu
K, Sa
toru N
agata.
A Si
mpl
e
and
Rob
u
st Meth
od to
Scree
n
Cataract us
in
g
Spec
ular
Refl
ection A
p
p
ear
a
n
ce.
Proc
ee
di
ng of SPIE M
e
dical
Imagi
ng
Confer
ence,
San Di
ego, C
a
l
i
forni
a
. F
ebruar
y. 20
08.
[16]
Supri
y
a
n
ti R, Hitoshi
H, Mas
a
tsugu K, Sato
ru N.
Extracting Appe
aranc
e Information i
n
s
i
de the Pu
pil
for Cataract Screen
ing.
Proc
eed
ing
of IAPR Confer
enc
e
on Ma
chi
ne
Visio
n
Appl
ica
t
ion, T
o
kyo,
Japa
n. Ma
y
.
2
009.
[17] Supri
y
a
n
ti
R,
Hitoshi
H, Mas
a
tsugu K, S
a
to
ru N.
Co
mpact
Cataract Scre
eni
ng Syste
m
:
Desig
n
a
n
d
Practical D
a
ta Acquis
i
tion.
Proceed
ing
of Internationa
l C
onfer
e
n
ce on Inst
rumentati
o
n
,
Commun
i
cati
o
n
, Informatio
n
T
e
chnolo
g
y
and B
i
om
edic
a
l En
gi
neer
in
g (ICICI-BME), Band
ung,
Indon
esi
a
. Nov
e
mber. 20
09.
[18]
Supri
y
a
n
ti R, Dhe
a
AP, Eko M, Haris BW
.
Compari
ng E
dge D
e
tectio
n Methods to L
o
c
ali
z
e
Ut
eru
s
Area on Ultras
oun
d
I
m
a
ge.
In
ternatio
nal
C
o
n
f
erence
o
n
Inst
rumentati
on, C
o
mmunic
a
tio
n
,
Informatio
n
T
e
chnolog
y
an
d Biome
d
ica
l
Engi
neer
in
g,
Bandu
ng, Indo
ne
sia. Novem
ber.
2013.
[19]
Supri
y
a
n
ti R, Elvin P, Yogi R, T
u
tik
IR.
Separabil
i
t
y
F
ilter for
Local
izin
g Ab
normal P
upi
l: Identific
atio
n
of Input Image.
T
e
lkomnik
a
Internati
ona
l Jour
nal.
20
13; 11(
4
)
: 783-79
0.
[20] Saat
y
RW
.
Val
i
dati
ng the
Ana
l
ytic Hier
a
rchy/
N
etw
o
rk Proce
ss.
Proceed
ing
s
of ISAHP, Vina d
e
l Mar
,
Cile. Au
gust. 2007.
[21]
Qi-chua
n T
,
Bin C,
Da-sh
en W
,
W
en-g
uan
g
F
A
N. Vi
rtual Sl
ice E
x
t
r
action B
a
se
d
on
Hermi
t
e
Interpol
atio
n.
T
e
lko
m
nika Ind
o
nesi
an Jour
nal
of Electrical E
ngi
neer
in
g.
20
12; 10(6): 1
430
–14
38.
[22]
Rad AE, S
hafr
y
M, R
a
h
i
m M
,
Norouz
i A. D
i
gital
De
ntal
X-Ra
y Imag
e S
egme
n
tatio
n
a
nd F
e
ature
Ex
traction.
T
e
l
k
omnik
a
Indo
n
e
sia
n
Journ
a
l o
f
Electrical Eng
i
ne
erin
g
. 201
3; 11(6): 310
9–
3
114.
Evaluation Warning : The document was created with Spire.PDF for Python.