Modeling the Spread of Infectious Diseases after Flood
in the Rainy Season in South Bandung, Indonesia
Nuraeni, Shimaditya
Arai, Takeshi
Department of Industrial A dministration, Faculty of Science and Technology
Tokyo University of Science
2641 Y amazaki, Noda-shi, Chiba-ken 278-8510
Tel : 04-7124-1501
shimaditya@ sbm-itb.ac.id, tarai@ rs.noda.tus.ac.jp
ABSTRACT
Flood is the most frequent natural disaster in Indonesia. However mitigation program is still not effective,
especially the indirect impact related to human health. The limitation on clean water, sanitation facilities,
cold temperature and the quality of health surveillance might trigger the outbreak of infectious disease
and also ineffective the mitigation policy. Therefore in this paper we use System Dynamic that modeled
the two diseases that mostly emerge during rainy season, diarrhea and acute-respiratory-infections (ARI)
to give insight and feasible policy to mitigate and control the potential outbreak of infectious disease
Keywords : diarrhea, acute-respiratory-infection (ARI), system dynamics
INTRODUCTION Figure 1. Map of Indonesia
Indonesia, a country in Southeast Asia located / INDONESIA |
in equatorial, has 5,590 main river with 600 of | sancKoxo VIETNAM Bice be
them are potentially prone to floods. The PHNOM PENH HOCMMI ET enmas ft
average precipitation is 147 mm per month. Moe weed Cael ;
; maps ‘ Tota BANDAR SERI PALAU
With the significant change of rain pon
precipitation because of climate change, the SINGAPORE, Sea)
number and frequency of rivers that ost 0) 4
potentially flood increase. Even though the sie io ik aa H
rain consequently causes overflowing in the Sa oe :
— INDIAN Lombok Komodo EAST
surface water, human activities and changes of OCEAN &Rinca TIMOR
ne * Darwin
land uses often make the situation worse. é oe =
More than 58% of Indonesian people live in L°™===2°0 miss =
Java Island with 43% of them stay in West Java and make it the most populous province in
Indonesia. There is one main river in West Java Province, The Citarum River, flows through
several cities (one of them is Bandung city). It runs three dynamos for hydropower electricity in
three dams, provides water for industries and 21.3 million residents and irrigates 390,000 paddy
fields.
Figure 2. Citarum River Basin
Source : ICWMRIP
The Citarum River also is prone to flood. The flood history started in 1986 when there was land
conversion for urban development. At that time, the water inundated 7.450 ha area, sank almost
10 villages and 10 sub-district including Baleendah, Buahbatu, Dayeuhkolot, Majalaya,
Rancaekek, Banjaran, Pameungpeuk subdistrict. Risking 68,635 people and forced 38,672 of
them to evacuate. After the normalization process, in 1994 the inundation area was decrease to
5,100 ha, and continued to decrease until 3,200 ha in 2001. In 2003, after a long dry season, the
inundation area caused by Citarum River was only 820 ha. But, in 2005 the inundation area then
moved to Baleendah and Dayeuhkolot sub-district (Figure. 3).
At the beginning of 2010, the water from the Citarum River sank 4.474 houses in South Bandung
and forces 2,915 people from Baleendah and Dayeuhkolot sub-district to evacuate”. This worst
flood noted for the last two decades, also made the transportation to South Bandung and 17
textile industries temporarily paralyzed, risking 50,000 workers with potential loss of 20 billion
tupiahs, and sank 3,109 ha paddy fields with potential loss 28 billion rupiahs. This situation
continued almost three months.
Figure 3. Inundation Area of Citarum River
1 In 1978, 65% of Bandung area was covered by buildings and residences. In 1981 the covered area increased to 85.6% and
becoming 90% in 1986. The catchment area in north Bandung has been exploited for residences. So, less than 10% of open area
that can absorb the water is left when the rainy season comes.
? This number is taken from Indonesian Red-Cross (recorded 0n19/02/ 2010) because there are several versions on the amount of
houses and refugees. ANTARA (24/3/ 2010) reported there are 1500 houses sank and 4000 people have to evacuate. Surya
Online (25/3/2010) reported there are 7960 houses sank and 16037 people have to evacuate. But for the simulation, the amount of
affected population is taken from Department of Health in South Bandung regency.
{ Inundate area in 1986 LB tnundate area in 2003
HE Inundate area in 1994 LB Inundate area in 2005
Hh Inundate area in 2001
Source : Pikiran Rakyat Newspaper. Sunday, 8" May 2011
Although there was no victims, many workers were not able to go to work, the school-aged
children could not go to school, and limited access to clean both water and food as well as cold-
temperature might trigger the several potential infectious diseases as the surveillance on the
evacuation site reported there are many of the refugees get diarrhea, acute respiratory infection
(ARI) and experiencing skin itches.
According to Smith K et al (1998), there are direct and indirect losses caused by the flood, one of
the direct intangible aspect are ill-health and water-borne disease. At short term, the impact of
floods on the transmission of communicable diseases is limited, but an increased risk for water,
vector-bome diseases, and diseases associated with crowding definitely exists (WHO, 2006).
Diarrhea, hepatitis A and E, and leptospirosis are diseases related with contaminated water.
Measles and acute respiratory infection (ARI) are diseases associated with crowding. Even flood
water may wash away existing breeding site of mosquitoes, but the standing-water caused by it
can create new breeding sites. That situation may trigger the vector-bome diseases such as
malaria and dengue.
Based on the previous experiences, the primary problem related to the flood disaster and
probably many other natural disasters in Indonesia is lower performance of disaster management.
Emphasis on only the post-disaster treatments instead of the prevention or mitigation plan, and
unstandardize reports on the number of victims, refugees and their conditions might made the
recovery process slow.
As mentioned in Indonesia’s Mitigation Plan 2010-2014, related to the direct intangible impact
of natural disaster loss which is health aspect, there are 21 priority diseases which should be
mitigated because they potentially become epidemic. The top five diseases are: acute diarrhea,
malaria, dengue-fever suspect, pneumonia and bloody diarrhea.
The Government response related to the flood in South Bandung is try to normalize the river
again, hopefully the inundation area is decrease and the flood will not as frequent as it used to be.
However with the complicated procedures and bureaucracy, there is no assurance when the
project will be started. Even if the normalization starts in 2011, the process needs at least two
years of operation. So the houses in South Bandung in the next two year are going to be
inundated by flood.
Therefore this paper tries to propose a decision support system for decision makers (DMs)
related to the flood phenomena in South Bandung. System Dynamics (SD) as a tool is used to
model the dynamic of diarrhea and ARI in both areas in South Bandung. Then, some scenario is
developed to see the efficacy of the interventions.
TARGETED AREA - SOUTH BANDUNG
Bandung is the capital of West Java province that consist of four areas : North Bandung,
Bandung city, East Bandung (Sumedang) and Bandung Selatan (Figure 4). Since 2007, West
Bandung was established juridically and make South Bandung, now, consist of 31 sub-districts
(Figure 5) with 3,172, 860 populations.
Figure 4. Map of Bandung Figure 5. Map of South Bandung Regency
Cae
panongect:
simon uzun
enencin = 7) ages,
‘kADUEATEN anUT
Source : www.bandungkab.go.id
When the flood hit South Bandung in 2010, there were six sub-districts that inundated and
affected by it. The most houses inundated by flood are located in Baleendah and Dayeuhkolot.
With the average person per house in Baleendah and Dayeuhkolot’, then the amount of people
that affected and should to evacuate are 76,410 and 23,308. Based on the interview with
Regional Disaster Management Department (Badan Penanggulangan Bencana Daerah-BPBD),
the gap between the amount of people who evacuate and the one that should be evacuate because
3 Baleendah area is 4,155.54 ha with 192,480 populations. Total house is 17, 840 with average person per house is
10.8. Dayeuhkolot area is 1,102.91 ha with 121,224 populations. Total house is 36,384 with average person per
house is 3.3 (Health Profile of South Bandung Regency 2009, Department of Health of South Bandung Regency)
most of the people who decide not to evacuate has two-level house. So, they prefer to
temporarily do their activities in the second floor rather than move to the evacuation site.
Table 1. Number of Inundate House and People whose Evacuate
i ad Inundate Total Should
Subdistrict House House % Evacuate Evacuate %
Dayeuhkolot 7,063 36,384 | 19.4 23,308 8,849 | 37.97
Majalaya 2,492 15,872 | 15.7 25,418 3,233 | 12.72
Ciparay a 22,080 | 0.2 283 283 | 100
Banjaran 1,646 16,512 | 10.0 11,357 4,896 | 43.11
Baleendah 7,075 17.840 | 39.7 76,410 4,450 | 5.82
Rancaekek 1,779 27,530 6.5 10,852 10,852 | 100
In the evacuation area for both sub-district, Baleendah and Dayeuhkolot, Department of Health
of South Bandung Regency noted that there are several diseases suffered by the refugees. The
top three diseases are ARI (31%), dermatitis (21%) and rheumatisms (10%), as shown in Figure
6. However, the Department of Health (DOH) of South Bandung Regency only recorded and
focuses on two types of disease which are A RI and diarrhea. With different time for recorded the
number of refugees who get diarrhea and ARI, generally, the number of people who get both
diseases in Baleendah (as we later will be called Area-1 for the simulation) is higher than in
Dayeuhkolot (as we later will be called Area-2 for the simulation). The daily records of the
number of people who get diarrhea or ARI from both areas are shown in Figure 7. The
accumulative data of the total people who get diarrhea or ARI from both areas are shown in
Figure 8.
Figure 6. Disease Propotion at flooded-subdistricts
in South Bandung (J anuari-Februari 2010)
-data until 28th February 2010-
Conjunctivitis
Source : DOH South Bandung Regency, 2010
Figure 7. Number of People in Area-1 and Area-2who get Diarrhea and ARI
—+*—Diarrhea Area-l1 = —s— Diarrhea Area-2 = —®—ARIArea-l1 = —e®—ARI Area-2
lvaoluaa| art | 202 | 273 | ava | 205 | 216 | 2/7 | 278 | 209 larnofavaafari2|aralanalanis|2relzi7|2/18}a/19|2/20)2721|2/29]/23|224]2/25 2/26\2/27] 2/28]
—2— Diarshea Area-1 3 [1fio] 8 | 4 [a2]ia] 5 [i2]a7]22]a6]is [ia 15| 8 |e] ae] 19] 20] 73] 10/28] 26] 10] 6 | 9
—a— ART Areal 80 |159/ 169] 72 | 98 | 63 [317/184]112/102/340] 139] 280/203] 84 | 7a | 75 | 74
0
8 Diarrhea Area-2| 0
0
3
—2— ART Area-2
Figure 8. Cumulative Number of People in Area-1 and Area-2 who get Diarrhea and ARI
—+— Cum Diarvhea Area-1 —s— Cum Diarrhea Area-2 —a— Cum ARI Area-1 —e— Cum ARI Area-2
4000
3500 ee
3000
2500 ae Zail
2000 a
1500 =
1000 as
500 a
0 le ne tnt SS 98 8 etter aa
lurso|uaa| 21 | 272 | 273 | ara | 275 | 26 | 27 | 278] 209 rob arzlanaler alan sjarel217pr1sla/19|2/20)2r21 2/29)728)/24/2/25]26)/27)2/2
—2— Cum Diarrhea Area-T 0 | 3 [17 | 27{ 35 | a9] 51 | 65 | 70 | 82 | 99 [121]147|162]175|190]198] 237]288] 302] 322) 395 [405] 433|459| 469] 475 |484|401
a4 {21 [28] 32] 38] 41 | as | 57 | 64 | 72 | 81 [177|185]193/198]206|222|224]249|255| 261 |267 | 267 |267|267|267|267|267|
0
alé
—s—CumARTArea't | 0 | 0 | 58 [142/230] 331] 368] 489]585 /609|601 /776|856) 101 118) 125] 134|141]172|191|202|212|246 260 288 308) 317) 324] 332] 339]
—s—CumARIAreae2 [ala
—#— Cum Diarrhea Area-2
4 [12] 24] 35 | 46 | 58 | 61 | 77 | 98 [118]130]134/ 136|150| 162] 166|171|186] 195|228]275]285|285|285| 285|285| 285] 285
DIARRHEA and ACUTE RESPIRATORY INFECTIONS
According to WHO (2006), natural disaster are catastrophic events with atmospheric, geologic
and hydrologic origins. They include earthquake, volcanic eruptions, landslides, tsunamis, floods
and drought. From the past two decades, developing countries are disproportionately affected by
natural disaster because of their lack of resources, infrastructure and preparedness systems.
The potential impact of communicable disease is often presumed to be very high in the chaos
that follows natural disaster, because the risk of the outbreak is associated with the size, health
status and living conditions of the population displaced by the natural disaster. Crowding,
inadequate water and sanitation, and poor access to health services, often characteristic of sudden
population displacement, increase the risk of communicable disease transmission.
Although the overall risk of communicable disease outbreak is lower than often perceived, but
the risk of transmission of certain endemic and epidemic-prone diseases can increase following
natural disasters.
Several diseases related to water-borne, vector-bome, diseases that associated with crowding and
diseases caused by corpse are summarize in Figure 9 below.
Figure 9. Communicable Diseases (WHO, 2006) Diarrhea
Diarrhea
Diarrhea is caused by
infectious organism, including
viruses, bacteria, protozoa, and
helminthes that are transmitted
from the stool of one
individual to the mouth of
another, termed _ fecal-oral
transmission. Some are well
known, others are recently
discovered or emerging new
agents, and presumably many
remain to be identified[4].
Water-bome Hepatitis A & E
Leptospires
Vector-bome
Communicable
5 Dengue
Disease gu
Associated with
Crowding
Acute Respiratory
Infections (ARI)
Tetanus
According to Caimcross
(2010) there are three factors that might reduce the risk of diarrhea. The risk reduction of
diarrhea associated with hand washing using soap is 42%, risk reduction associated with
improved water quality is 17% and 36% associated with sanitation/ excreta disposal.
Acute Respiratory Infections (A RI)
Acute respiratory infections (ARIs) are classified as upper respiratory tract infections (URIs) or
lower respiratory tract infections (LRIs). The upper respiratory tract consists of the airways from
the nostrils to the vocal cords in the larynx, including the paranasal sinuses and the middle ear.
The lower respiratory tract covers the continuation of the airways from the trachea and bronchi
to the bronchioles and the alveoli [5].
Pneumonia
Pneumonia can be categorized into lower respiratory tract infections. Both bacteria and viruses
can cause pneumonia. Bacterial pneumonia is often caused by Streptococcus pneumonia
(pneumococcus) or Haemophilus influenzae, mostly type b (Hib), and occasionally by
Staphylococcus aureus or other streptococci. In the developing country, pneumonia is the second
cause of death for children under 5 years old.
Beth On Ith Profil Table 2. Demographic of Area-1
aleendah (Area-1) Health Profile Male | Female | Total | %
* 2
Area 3 4,155.54 ha (16.82 km*) 14 7150 9438 | 165881 8.62
Population : 192,480 people 5-14 | 22308] 20592 | 42900 | 22.29
Household : 17,840 houses 3 15-44 | 45474] 48048] 93522 | 48.59
Density : 46.32 perha (11,443.52 per km*) 15-64 76588 | 15158) 31746 | 1649
Average person per household : 10.8 365 3147 | 4577 77241 4.01
From the all the house that monitored by DOH, Total 046671 978131 192480 100
45.81% (8484 household) show the hygiene
behavior such as hand wash with soap, 69.24% have clean water access, and 53.4% owned good
sanitation in their house.
Table 3. Demographic of Area-2
Dayeuhkolot (Area-2) Male | Female | Total %
Area : 1,102.91 ha (4.46 km’) 1-4 2736 | 3010| 5746 | 4.74
Population : 121,224 people 5-14 11766) 11219] 22985] 18.96
Household : 36,384 houses 15-44 33110 | 32291] 65401 | 53.95
Density: 109.91 per ha (27,180.27 per km”) 45-64 12313 | 12315 | 24628 | 20.32
Average person per household : 3.3 >65 1918 546 2464 | 2.03
In Dayeuhkolot area, only 11,361 houses are Total 61843 | 59381) 121224 100
monitored by DOH. Assuming this number represent the population, then it can be concluded
that 57.21% household in Dayeuhkolot show the hygiene behavior. The number of house that
have clean access is 83.82% and 86.89% owned good sanitation in their house.
MODEL DESCRIPTION AND SIMULATION
The basic structure of both diarrhea and ARI model are adapted from Kermack-McKendrick
with some modification. The population is represented as stock and will be represent the health
state but susceptible The initial populations modeled in the simulation are 4450 and 8849 for
area-1 and area-2, respectively.
In diarrhea model, the probability of susceptible population being infected by diarrhea microbe
depends on the incidence rate. The rate can be decrease through three factors such as the access
to safe drinking water, hygiene behavior rate and basic sanitation rate. For the population that in
the infected state can be recover immediately if they understand about the characteristic of the
disease so necessary and immediate treatment can be done. Also depend on the time of
infectivity which is in the period of 1-2 days diarrhea will reduce. For the people who don’t have
the knowledge about disease characteristic and how to overcome the primary symptom will get
sid; which is represented as conveyor stock. From that stock the number of people who recover
will be depend on the recovery rate, which affected by the number of clean water and
rehydration availability in the evacuation site. The dynamic model of diarrhea can be
represented in Figure 10.
Figure 10. Dynamics Model of Diarrhea
ORS & Zine
Efficacy!
ma @)
Sick!
—— vA
coos \ pe
ORS & Zine
StoRD1 Rate!
Rate!
Get Sickt
Clean Water
Rate?
Incidence
Rate!
fi, ascinicctviny /*etPont acta
en Clean Water
: Citceer
HealthyPopt
Different characteristic with diarrhea, the probability of the population get infected by ARI
microbe is depends on the incidence rate and contact rate. Since the disease of ARI is
transmitted through droplet or air-transmission (Figure 11).
Figure 11. Dynamics Model of ARI
HealthyPops
@ cums V
Get Iffected3 Get Sick8_—
. al
cumsick? Recovered
we)
(sess
Good
ee
Se
Rated
_ 2.
Coverages
Vaccination
Efficacy
OF ised a athyPopd
Increasing
Hazard3
Prevalence — T5495
Probinfected Get Worse3
}
Transmission
Probabilty3
ContactRate3
The recovery rate in ARI’s model depends on the level of how good is the nutrition of the
population and vaccination coverage. From both areas, area-1 is tends to have worse condition
than area-2 in the level of risk-reduction for diarrhea. Even the vaccination coverage in area-1 is
higher than area-2, but the number of people who get pneumonia in area-1 is higher than area-2.
The simulation was run on two conditions. First condition is status quo, which is represents the
previous situation. The second condition of simulation is run after interventions are introduced in
the model.
The intervention of both prevention and mitigation are applied on the model by increased the rate
that related to prevention factors and mitigation factors. Since the aim of this paper is to propose
decision support system, then it is necessary to have currency based. To calculate the currency
on certain intervention, the reduction between targeted rate and basic rate (status quo) multiplied
by acost based on DALY. After that combining the total budget needed from the targeted rate
with the budget allocation policy will give the efficacy of the interventions.
Table 4. Intervevntion for Diarrhea
Hygiene Behavior | Safe Drinking Water Basic Sanitation
Area
Current | Target Current Target Current | Target
Area-1 45.81% 90% 69.24% 90% 53.4% 75%
Area-2 57.21% 90% 83.82 % 90% 86.9% 75%
SIMULATION RESULT
With assumption of 2% budget allocation for both prevention and mitigation intervention, the
efficacy of prevention program is 15% while the mitigation 2%. That efficacy shows only for
diarrhea patients which reduce 25% from the basic situation. While for ARI, there is no different
between the basic situation and after intervention. The not-significant result for ARI disease
might be caused by the high-cost for intervention program, which mention before that the
mitigation program for ARI is by introduce vaccination where most of the pneumococcal
vaccination is not well known and expensive in Indonesia.
The result of simulation can be shown in Figure 12 below.
Figure 12. Simulation Results
BData OSimulation # Intervention
4500
4000
3500
3000
2500
2000
1500
1000
500 i = —
2 Diarrh Diarrh
iarrhea iarrhea
‘Area-1 ‘Area-2 ARI Area-1 | ARI Area-2
BData 491 267 3396 285
OSimulation 516 313 4088 323
Intervention 387 238 4088 323
10
DECISION SUPPORT SYSTEM
The aim of this paper is to propose a decision support system (DSS) for decision makers (DMs)
on creating policy regarding the health aspect that affected by flood in South Bandung.
Based on model construction there are three aspects that DMs can considered in creating the
policy :
1. The budget allocation.
2. The prevention intervention
3. The mitigation intervention
Those three aspects are interconnected. The budget allocation will affect on the efficacy of both
prevention program and mitigation program. The target area or rate on both prevention and
mitigation programs correlated with the total cost needed, which also related to the capability of
current budget availability. The most effective cost and also intervention will depend on this
three aspects. How much budget can be allocated, to which program, and for how many targeted
area or rate.
LIMITATION AND FUTURE RESEARCH
The limitation of this research is that the populations that being modeled are assumed to be
homogenous, no age cohort which might have different probability on being infected (i.e
children under 5 years old is more vulnerable to get infected by diarrhea and pneumonia).
The intervention programs are input simultaneously. It means that the trade-off between each
intervention programs are not calculated to see which program has the most cost-effective effect
to prevent or mitigate both diseases.
Therefore, it is a hug opportunity for future research on developing the simulation model to be
more sophisticated and complex.
11
References
[1] Brailsford, S.C., Berchi, R., Angelis, V.De., and Mecoli, M. (2007). “System Dynamics
Models to Assess the Risk of Mosquito-borne Diseases and to Evaluate Control Policies”.
International Conference of the System Dynamics Society, 2007.
[2] Ritchie-Dunham, James L., and Mendez Galvan, Jorge F. (1999).“Evaluating epidemic
intervention policies with systems thinking : A case study of dengue fever in Mexico”.
System Dynamic Review Vol. 15 No.2
[3] Pryut, Erik. “Cholera in Zimbabwe’/ International Conference of the System Dyanamics
Society, 2009.
[4] World Bank. (2006). Disease Control Priorities in Developing Countries, ond Edition, Chapter
19 Diarrheal Diseases. Oxford University Press.
[5] World Bank (2006). Disease Control Priorities in Developing Countries 2"¢ Edition, Chapter
25A cute Respiratory Infection in Children. Oxford University Press
[6] USAID (2007). Assessment for the Introduction of Zinc in Improved Management of
Diarrhea in Indonesia.
[7] Department of Disaster and Mitigation Plan Republic of Indonesia. National Mitigation Plan
2010-2014.
12
