FACTORS ASSOCIATED WITH MOTHERS’ HEALTH SEEKING BEHAVIOUR ...

FACTORS ASSOCIATED WITH MOTHERS’ HEALTH
SEEKING BEHAVIOUR FOR TREATMENT OF
CHILDHOOD DIARRHOEAL DISEASE
IN INDONESIA

JONAS EDRIAN SUMAMPOUW

A THESIS SUBMITTED IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR

THE DEGREE OF MASTER OF ARTS
(POPULATION AND REPRODUCTIVE HEALTH RESEARCH)

FACULTY OF GRADUATE STUDIES
MAHIDOL UNIVERSITY
2000
ISBN 974-664-609-5

COPYRIGHT OF MAHIDOL UNIVERSITY

Thesis
Entitled

FACTORS ASSOCIATED WITH MOTHERS’ HEALTH SEEKING
BEHAVIOUR FOR TREATMENT OF CHILDHOOD
DIARRHOEAL DISEASE IN INDONESIA

Jonas Edrian Sumampouw
Candidate

Asst. Prof. Uraiwan Kanungsukkasem, Ph.D
Major Advisor

Lect. Alan Noel Gray, Ph.D
Co-Advisor

Prof. Liangchai Limlomwongse, Ph.D. Assoc. Prof. Chai Podhisita, Ph.D.
Dean Chairman
Faculty of Graduate Studies Master of Arts Programme in Population
and Reproductive Health Research
Institute for Population and Social Research

Thesis
Entitled

FACTORS ASSOCIATED WITH MOTHERS’ HEALTH SEEKING
BEHAVIOUR FOR TREATMENT OF CHILDHOOD
DIARRHOEAL DISEASE IN INDONESIA

was submitted to the Faculty of Graduate Studies, Mahidol University
for the Degree of Master of Arts

(Population and Reproductive Health research)
on

August 24, 2000

Jonas Edrian Sumampouw
Candidate

Asst. Prof. Uraiwan Kanungsukkasem, Ph.D
Chairman

Lect. Alan Noel Gray, Ph.D
Member

Ms. Supanee Jivasak-Apimas, Ph.D
Member

Prof. Liangchai Limlomwongse, Ph.D. Assoc. Prof. Bencha Yoddumnern-Attig, Ph.D.
Dean Director
Faculty of Graduate Studies Institute for Population and Social Research
Mahidol University Mahidol University

ACKNOWLEDGEMENTS

This is the LORD'S doing; it is marvellous in our eyes (Psalms 118:23-24).

First of all, I give thanks to God, He granted me the health, the ability, the zeal,
and the courage to successfully complete this 12 months Master of Arts Programme at the
Institute for Population and Social Research, Mahidol University, Thailand.

I would like to thank to Dr. Uraiwan Kanungsukkasem, my major advisor, and Dr.
Alan Noel Gray, my co-advisor, for their intensive guidance, being available and sparing
their valuable time to help me in all possible ways towards finishing this thesis. I am also
indebted to Dr. Supanee Jivasak-Apimas, my external reader, for her valuable comments
and suggestions at the completion of this thesis.

I wish to express my deepest gratitude to Dr. Bencha Yoddumnern-Attig, Director
of IPSR for providing me the opportunity to participate in this program and for her
continuous support during my study. I am highly indebted and warmly appreciate Dr.
Chai Podhisita, the Program Director, for his support and concern on my studies. My
earnest appreciation also goes to all lecturers of IPSR for their kindness and sharing of
their knowledge, and experiences during my study.

My sincere thanks to Khun Luxana Nil Ubol (M.A Program Coordinator), Khun
Padiwara Prasartku (Gring) and all the staff of IPSR for their warm co-operation and kind
assistance throughout my presence in Thailand. I would never forget all the support from
my classmates, especially my group discussions with Tita Lorna Perez, Cyprian Mpemba,
Glenrose Poyah and my English editor Bina Gubhaju.

I would like to express my gratitude to MEASURE Evaluation (USAID), for
providing me financial support to carry out my study at the IPSR.

I will not forget Dr. James Sinaya, and Dr. Mieke Andris, who recommended me to
come and study at this institute. To the Head of North Sulawesi Provincial Health Office
- Indonesia and all the staff there, thank you for all your support and prayers.

Last but not least, I owe to my beloved wife, Dr. Nora Louisa Sondakh, who not
only took care of our sons Ezra Louis Patrick and Jeremy Eliazer during my absence, but
also was a constant source of inspiration and loving support. I acknowledge your
sacrifice, encouragement and value most tremendously your prayerful support.

He hath made everything beautiful in His time (Ecclesiastes 3:11a).

Jonas Edrian Sumampouw

Fac. of Grad. Studies, Mahidol Univ. Thesis / iv

4238532 PRRH/M : MAJOR: POPULATION AND REPRODUCTIVE HEALTH
RESEARCH; M.A. (POPULATION AND REPRODUCTIVE
HEALTH)

KEY WORDS : CHILDHOOD DIARRHOEA / MOTHERS' CHOICE OF
CARE / INDONESIA

JONAS EDRIAN SUMAMPOUW: FACTORS ASSOCIATED WITH
MOTHERS HEALTH SEEKING BEHAVIOUR FOR TREATMENT OF CHILDHOOD
DIARRHOEAL DISEASE IN INDONESIA. THESIS ADVISORS: ASST. PROF.
URAIWAN KANUNGSUKKASEM, Ph.D., ALAN NOEL GRAY, Ph.D. 61 P. ISBN
974-664-609-5

Although many efforts have been made by the Indonesian Government to
control diarrhoeal disease, the morbidity of diarrhoea among children under five years
remains high. One of the reasons for such a high morbidity may be attributed to the
under-utilization of modern health facilities. This study has analyzed factors associated
with Indonesian mothers' health seeking behaviour for treatment of their children's
diarrhoea. Children data file from the 1997 Indonesian Demographic Health Survey
(IDHS). Only 1,603 children under five years, who had diarrhoea within 14 days prior to
the survey, were selected as the sample, and their mothers were given a questionnaire.

It was found that modern treatment was most frequently used by mothers
regarding their choice of care (54 per cent), followed by no treatment and self-treatment.
Traditional healers were least frequently reported as a choice of care (4.5 per cent).
Multinomial logistic regression was used to assess the net effect of several independent
variables on the dependent variable. The findings also revealed that of the eleven
independent variables, age of mothers, level of education, mothers' knowledge of ORT,
age of children, duration of diarrhoea, and symptoms and signs of diarrhoea were
statistically significant factors affecting mothers' choice of care. Other variables such as
mothers' place of residence, mothers' labour force status, mothers' religion, mothers'
accessibility to source of care and children's sex were not found to be significant.

The result of this study indicates that education plays a major role in mothers'
decision making to seek care for their children along with their knowledge of ORT. Thus,
it can be suggested that educating mothers is essential in order to increase their
knowledge, especially knowledge of how to deal with diarrhoeal disease as soon as
possible. To accelerate achieving the goals of the diarrhoeal disease control program in
Indonesia, one important element is to target younger mothers. Emphasis should be given
on educating them and providing them with knowledge, not only on how to use ORS or
ORT, but also on how to identify signs of dehydration, and to seek timely medical care.

TABLE OF CONTENTS Page
iii
ACKNOWLEDGEMENTS iv
ABSTRACT
LIST OF TABLES viii
LIST OF FIGURE ix
CHAPTER 1 INTRODUCTION 1
1
1.1 The Problem's Background 2
1.2 Health Care System In Indonesia
1.3 Approaches Towards Diarrhoea Treatment in 3
5
Indonesia 6
1.4 Justification 7
1.5 Research Question
1.6 Research Objectives

CHAPTER 2 LITERATURE REVIEW 8
2.1 Essential Concept Concerning Diarrhoeal Disease 8
2.2 Approaches to Diarrhoea Treatment 10
2.3 Education, Knowledge and Health Seeking Behaviour
14
Toward Diarrhoeal Disease 17
2.4 Socioeconomic Status and Health Seeking Behaviour
2.5 Mothers' Perception of Diarrhoea Illness and Health 18

Care Seeking 20
2.6 Age of Mother and Health Seeking Behaviour to Treat 20

Childhood Diarrhoea
2.7 Residence and Health Seeking Behaviour

vi

Contents (Cont.) 21
22
2.8 Conceptual Framework
2.9 Research Hypothesis

CHAPTER 3 RESEARCH METHODOLOGY 24
3.1 Source of Data 24
3.2 Sampling Method 24
3.3 Sample Size 25
3.4 Research Instruments 26
3.5 Limitations of Study 26
3.6 Data Analysis 27
3.7 Operational Definitions 27

CHAPTER 4 FINDINGS 31
4.1 Background Characteristics 31
4.1.1 Mothers' Socio-demographic Characteristics 31
4.2 4.1.2 Knowledge of ORT, Accessibility to Sources of Care,
32
Symtoms and Signs, and Choice of Care 35
4.1.3 Children's Characteristics
Factors Associated With Mothers' Health Seeking Behaviour 36
For Treatment of Childhood Diarrhoea

CHAPTER 5 DISCUSSION, CONCLUSION AND RECOMMENDA-
TIONS
5.1 Discussion 44
5.2 Conclusions 44
5.3 Recommendations 44
48

vii

Contents (Cont.)

REFERENCES 50
APPENDICES 55

A. Individual Questionnaire in Indonesia Demographic and Health 55
Survey 1997 58
60
B. Specific Questionnaire Regarding Diarrhoea 61
C. Kroeger's Model
BIOGRAPHY

LIST OF TABLES

Table 1. Percentage distribution of socio-demographic background of 32

mothers

Table 2. Percentage distribution of knowledge of ORT, accessibility to 34
source of care, symptomps and signs, and choice of care

Table 3. Percentage distribution of children's characteristics 35

Table 4. Multinomial logit coefficients for the model of choice of care 37

Table 5. Predicted probability of choice of care 39

LIST OF FIGURE

Figure 1 Conceptual framework for studying factors associated with

mothers' health seeking behavior for treatment of childhood

diarrhoeal disease 23

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 1

CHAPTER 1
INTRODUCTION

1.1 The Problem's Background
Diarrhoeal disease is a leading cause of childhood morbidity and mortality. It is

also an important cause of malnutrition. Diarhhoea continues to be a public health
problem in many countries, particularly in developing countries. Despite the tremendous
advances in medicine and technology over the last decades, diarrhoeal disease is still the
main cause of deaths in children under the age of five. On average, children below three
years of age in developing countries experience three episodes of diarrhoea each year. In
1993, an estimated 3.2 million children below 5 years of age died from diarrhoea (WHO,
1995). The threat posed to children under the age of five years is also evident in the
number of episodes that most young children incur in a year. The available estimates
show that in many Third World countries, the annual diarrhoea incidence ranges from 2
to 12 episodes (Pitts et al., 1996 cited in Iyun, 1999). In Indonesia, children under five
years of age suffered diarrhoea 1.6 to 2.2 episodes each year (MOH, 1998). According to
1995 the Indonesia Household Health Survey, 29.3 per cent of deaths of children under
the age of five was caused by diarrhoea disease (MOH, 1998)

Jonas Sumampouw Introduction / 2

The main cause of death from acute diarrhoea is dehydration, which results from
the loss of fluid and electrolytes in diarrhoea stools. Maternal practices regarding health
care have been recognized as an important social and anthropological factor behind the
high mortality rates among children aged less than five years. Maternal literacy and
health education, socioeconomic status, culture, beliefs, practices, and access to health
care are among the factors contributing to the high mortality rates caused by diarrhoea
(Cuevas, et al. nd).

1.2 Health Care System in Indonesia
Indonesia's health care system consists of three major sectors: public, private and

non-governmental. The Ministry of Health is primarily responsible for designing a
national health development policy and strategy and developing plans of action for
achieving the health objectives. The efficiency and effectiveness of health efforts can be
achieved if the needs of manpower resources, funding and health facilities are fulfilled
(MOH, 1998). In Indonesia, to increase quality of health care and even distribution of
health care, efforts are emphasized at the primary health care. This has been done by
puskesmas or community health centers to give an integrated service. Puskesmas
provides the majority of health care services in Indonesia. These include health
promotion, prevention, rehabilitation, and curative therapies. The staff of each puskesmas
consists of one or two physicians and a team of eight to ten paramedics/nurses as medical
support staff. There are 7,243 puskesmas and 21,115 sub-health centres (pustu) in the
country (MOH, 1998). However, Indonesia consists of five main islands and more than

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 3

13,000 small islands, half of which are inhabited, with an area of 5,193,250 sq. km, the
distribution of people is very uneven. Many people who live in remote areas face
difficulties in accessing health centers or sub-health centers. Increasing people's
awareness of the importance of health will increase their demand to obtain better quality
health care from health facilities and providers such as doctors, nurses, midwives and
other health providers. A feature of the health programme is to recruit and place doctors
who have just finished university, and nurse/midwives, in the rural villages.

1.3 Approaches Towards Diarrhoea Treatment in Indonesia
In Indonesia, a diarrhoeal control program has been instituted by improving

environmental sanitation, improving treatment facilities in the hospitals, health centers,
sub-health centers and health posts (posyandu). The current general objective of the
diarrhoeal disease control program is to reduce the morbidity and mortality due to
diarrhoea. The specific objectives are to decrease the incidence from 280 to 220 per
thousand population; to reduce episodes of children's diarrhoea from 2.2 to 1 per year,
and to reduce the case fatality rate (CFR) during the outbreak to less than 1 per cent.
Providing training for doctors and nurses at regency level, establishing oral rehydration
centers in health centers and health posts, and introducing the use of ORT for treatment
of diarrhoea through mass media, especially television have also been established by the
program (CBS et al., 1998).

Despite the measures instituted by the program, health services have been under-
utilized in Indonesia. Grace (1998), discussing the result of her study in Lombok, West

Jonas Sumampouw Introduction / 4

Nusa Tenggara province, mentioned that the level of utilization of modern curative health
services for treatment of childhood diarrhoea in Indonesia is questionable. The majority
of villagers still seek treatment from traditional healers rather than utilize modern
curative treatment. The Indonesian government health officials attribute the under-
utilization of health services to the level of education and modernity of the people.
Health officials believe that villagers will not abandon traditional healers for doctors,
midwives and nurses until they are educated and modernized. Although western
allopathic medicine is increasingly becoming widely accepted in Indonesia and used
more frequently than traditional medicines, traditional medicines are taken as a remedy
against diarrhoea (Supardi et al., 1997).

A study conducted by Utomo and Iskandar (1986) has shown that only 6.6 per cent
of rural people in Indonesia visit a 'western' medical doctor each year. Probably the
current situation might be different from that result since many efforts have been made
since that time in order to achieve better health status. Although the percentage may have
increased from 1986, treatment at a modern health care facility is still low in Indonesia,
particularly in rural areas. A study of deaths among children under five years of age,
conducted in rural Indonesia (Sutrisna et al., 1993), found that only 36 per cent of all
children who died had sought care from a doctor, nurse or midwife in the private or
public sector. Even though the majority of households had close access to a government
health post, this option was not chosen. This may be due to the understaffed facilities,
overworked, and under-motivated personnel that may result in long waiting periods.

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 5

Also, the inaccessibility of the facilities to the villagers, who are ill but not able to travel
long distances in comfort, may be the cause of non-utilization of modern health facilities.

Seeking treatment from local practitioners saves a woman time that might be
spent travelling to a far-away health care facility. It also helps maintain her
confidentiality by eliminating the need to explain a long absence from home or routine
duties. However, this approach can be harmful for children who are suffering from
diarrhoea complications. Many of the village-based midwives are untrained in
recognizing symptoms of critical conditions. Thus, they often cannot offer the children
the treatment they need or refer them to an appropriate treatment facility, such as a
puskesmas. This delay in receiving treatment can be dangerous and even fatal, since early
recognition and treatment of symptoms can prevent many of the more serious
complications.

1.4 Justification
The Indonesian government has made many efforts to decrease the child mortality

caused by diarrhoeal disease. Improving treatment from health facilities and health
providers are among the measures used. However, under-utilization of health services
particularly in rural areas still exists. Of course there are numerous factors that can affect
mothers' decision making about when and where to seek help for their children's
diarrhoea, which is known as health seeking behaviour. Probably the availability of
services, acceptability of the services, level of confidence of mothers in a health

Jonas Sumampouw Introduction / 6

providers' diagnosis or the affordability of consultation or medicine are among the
decisive factors.

Generally, the implicit assumption, that increased utilization of modern curative
services and the abandonment of traditional care practices will reduce child mortality, is
accepted. However, in some parts of Indonesia, harmful practices of traditional healers
are still in practice. Mothers often turn first to them for advice when a child has diarrhoea
because of various reasons. Practices of mothers to buy medicines from pharmacies or
shops to administer self-treatment for their children's diarrhoea, which is harmful, are
also still found.

Research is needed to identify what factors affect Indonesian mothers' health
seeking behaviour in controlling their children's diarrhoea. The results of the current
study will help decision-makers in Indonesia to take a holistic approach in dealing with
problems such as improving the planning and implementing of more effective
management of diarrhoeal problems to prevent diarrhoea diseases in the future. Finally,
the existing high mortality rate due to diarrhoea can be reduced to a lower level.

1.5 Research Question
What factors affect Indonesian mother’s health seeking behaviour for treatment of

their children's diarrhoeal diseases?

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 7

1.6 Research Objectives
1. Immediate objective.
To assess factors affecting Indonesian mothers' health seeking behaviour for

treatment of childhood diarrhoeal disease.

2. Ultimate objective.
To provide information for policy makers in order to improve the performance of
the diarrhoea control program in Indonesia

Jonas Sumampouw Literature Review / 8

CHAPTER II

LITERATURE REVIEW

2. 1 Essential Concepts Concerning Diarrhoea
Diarrhoea is a complex of symptoms and signs, usually defined as an increased

number of stools passed. According to World Health Organization (WHO, 1995),
Diarrhoea is the passage of loose or watery stools, usually at least three
times in a 24-hour period. However, it is the consistency of the stools
rather than the number that is most important. Frequent passing of formed
stools is not diarrhoea. Babies fed only breastmilk often pass loose,
"pasty" stools; this also is not diarrhoea. Mothers usually know when their
children have diarrhoea and may provide useful working definitions in
local situations.

The word 'diarrhoea' is also used in programmatic and public health context,
although not typically in clinical context, to embrace dysentry. Dysentry is usually
characterized by the presence of blood in the stools, with or without excessive looseness
of frequency (Martines, 1993). Most diarrhoeal episodes occur during the first two years
of life and the incidence is highest in the age group 6 to 11 months, when weaning often
occurs (WHO, 1992). Although diarrhoeal episodes tend to be self-limiting and generally
lasting for less than seven days, however, a small delay in initiating appropriate treatment
can result in the illness becoming so severe that infants cannot be saved. Studies in
several developing countries have shown that three to twenty per cent of acute diarrhoeal

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 9

episodes in children under five may become persistent, lasting for at least fourteen days

(WHO, 1992).

Practically, diarrhoea treatment is based on the clinical type of illness without

need of laboratory studies. World Health Organization (WHO, 1995) recognised

diarrhoea as four types, which are:

1) acute watery diarrhoea which lasts for several hours or days, dehydration is
the main danger and weight loss that occurs if feeding is not continued

2) acute bloody diarrhoea which is also called dysentery; intestinal damage,
sepsis and malnutrition are the main dangers

3) persistent diarrhoea which lasts 14 days or longer with malnutrition and
serious non-intestinal infection dehydration are the main dangers

4) diarrhoea with severe malnutrition (marasmus and kwashiorkor). Conditions
of diarrhoeal disease may be present with atypical or masked signs so that
assessment and aggressive treatment should be taken carefully

Diarrhoea disease is of course very serious among newborns and young infants

who cannot tolerate even small upsets in fluid and electrolyte balance (Ebrahim, 1993).

Diarrhoea disease is rare among infants who are entirely breast fed, but very common

among those who are not breastfed or are artificially fed, especially if the standards of

hygiene are poor. According to World Health Organization (WHO, 1992), specific

behaviours such as failing to breast-feed exclusively for the first four to six months of

life, using infant feeding bottles, storing cooked food at room temperature, using

drinking-water contaminated with faecal bacteria, failing to wash hands after defecation,

after disposing of faeces or before handling food, have been recognized as practices that

promote the transmission of enteric pathogens and thus increase the risk of diarrhoea.

Jonas Sumampouw Literature Review / 10

Unfortunately, the management of illness in young infants continues to pose
major clinical challenges. However, WHO and UNICEF have developed a guideline to
assist first level facility health worker to assess, classify and treat infants diarrhoea
(WHO, 1995). That guideline described procedures for dealing with common conditions,
and recommended urgent referral to the hospital if the infants presents ‘danger sign’ such
as dehydration or severe persistent diarrhoea.

2. 2 Approaches to Treatment
According to Paredes (1992), families knowledge of how to manage diarrhoea at

home with fluids and food and recognizing of when children need treatment will
determine the success of community based program of diarrhoeal disease control.
Families often try different approaches, both modern and traditional to treat childhood
diarrhoea. Furthermore, Paredes (1992), listed the choices that mother sought to treat
their children with diarrhoea, as follow;

(1) home care without drugs,
(2) home care after asking the advice of relatives or neighbours,
(3) home medication with purchased drugs or home treatment using traditional

remedies,
(4) visiting a traditional healer,
(5) seeking advice and/or prescribed treatment from a pharmacist
(6) seeking advice and/or prescribed treatment from a health worker in

government or private practitioner.

To improve health, programmes often try to increase accessibility of the modern
medical care such as hospitals or clinics. But the fact that hospitals or clinics are there

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 11

does not necessarily mean that people will use it or accept it. Many will continue to seek
traditional healers and give the home remedies for their children.

According to Freund (1992), in many cultures, particularly in the developing
countries and for a variety of reasons, families often turn first to traditional healers for
advice when their children have diarrhoea. It would seem sensible to try to ensure that
traditional healers give the right advice about fluids and feeding, so that the harmful
practices can be discouraged.

Despite progress made to treat diarrhoeal disease, this common childhood illness is
still taking a heavy toll. Diarrhoeal disease is a leading cause of childhood morbidity and
mortality and an important cause of malnutrition in developing countries. Tandon et al
(1984, cited in Edmunson, 1992) reported that diarrhoea contributed to the deaths of 43
per cent of 4,300 Indian children with clinical protein-calorie malnutrition. WHO (
1995) reported that 8 out of 10 childhood deaths caused by diarrhoea occured in the first
two years of life.

The provision of health care was primarily considered the domain of health
providers in the past, however, the current trend has recognized the importance of
mothers and the family in identifying, caring for and preventing children's illness
(Cuevas, et al nd). It is commonly known that maternal practices regarding health care
have been recognized as important social and anthropological factors, explaining high

Jonas Sumampouw Literature Review / 12

mortality rates due to diarrhoea disease among children aged less than five years.
Maternal literacy and health education (Mull, 1988), parental age, family's
socioeconomic status (Scrimshaw, 1988,) and access to health care (Weiss, 1988) are
among the factors mentioned. Furthermore, Cuevas et al (nd) reported that lack of
information to identify complications, such as dehydration, limited use of oral
rehydration solutions, inadequate maternal health seeking behaviour and dietary
modifications, are the factors that have prompted decision-makers to involve the family
particularly mothers to reduce the burden of diarrhoea disease.

Feyisetan (1997) noted that in Yorubaland-Nigeria, besides inadequate availability
of health care services in many areas, especially the less developed countries, certain
diseases-specific cultural beliefs related to diarrhoea disease might influence people's
health seeking behaviour. Health services may be underutilized and several health and
child care instructions may be ineffective or ignored in traditional and transitional
societies where people's ideas and behavioural pattern conflict with the knowledge being
passed to them (ibid). The understanding or knowledge that diarrhoeal disease is caused
by supernatural power may lead them to seek treatment from traditional healers. This
situation still occurs in many parts of Indonesia, although western allopathic medicine is
increasingly becoming more widely accepted. The people still believe in supernaturalism
so they view western medicine as a system that merely provides symptomatic or
supportive therapy. Therefore, improving health facilities, training health providers and
other efforts to control diarrhoea such as changing people's behaviour, without

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 13

considering and understanding the culture of the population probably will not be
sufficient.

It is obvious that in many countries, as well as in Indonesia, traditional healers are
often the first providers of health care for many families in both rural and urban areas
(Arustiono, 1999). The exact number of traditional healers in Indonesia is not clearly
known, but it is well known that in some places, traditional healers are present in far
greater numbers than modern medical personnel. The advantages of traditional healer are
they are already treating many cases of diarrhoea and traditional healers work in the
community and are familiar with the way people think about illness. They are also
familiar with the attitudes of mothers. However, according to Hayfron, 1992, there are
also many disadvantages of traditional healers, which caused difficulties to collaborate
with them. If traditional healers do not treat diarrhoea correctly, children may develop
more serious dehydration and be taken to the clinics too late. The potential for
collaboration between traditional healer and modern medical systems is being explored in
some countries. According to Freund (1992), in Zambia, the Ministry of Health created a
Department of Traditional Medicine, and supported the setting up of traditional
practitioners' association. Ultimately, they discovered that most healers could recognize
the signs of dehydration and said that they would be willing to advise giving oral
rehydration therapy.

Jonas Sumampouw Literature Review / 14

2. 3 Education, Knowledge and Health Seeking Behaviour Toward Diarrhoeal
Diseases
The importance of mothers' education regarding infant and child mortality has

been well established and widely accepted. It is generally agreed that maternal education
acts as an independent determinant of infant and child mortality. Maternal education
influences child survival through many pathways, such as enhancing socio-economic
status, greater health choice for children, including interaction with medical personnel
(Rajna, 1998). The education advantage is through greater access to child health services
and environmental factors. Nevertheless, many studies have found that education is not
associated with the mother’s decision making regarding treatment (Granich, et al., 1999,
Cuevas, nd, Langsten, 1995, Emch, 1999, Bhandary, 1995). The gaps between mothers'
education and their decision making to control diarrhoeal disease in these studies were
probably influenced by the cultural belief context of the population under study. It has
been noted that health services may be underutilized and several health and child care
instructions may be ineffective or ignored in traditional and transitional societies where
people's ideas and behavioural patterns conflict with the knowledge being passed to them
(Feyisetan, 1997). However, compared to educated mothers, less educated mothers are
more likely to seek care, particularly from a government clinic facility, pharmacy, or
other informal sources as Langsten (1995) has argued. Langsten found that mothers'
education makes little difference to whether or not care is sought from a private
physician.

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 15

It is obvious that education increases mother’s knowledge about diarrhoea and
increases mother’s perception to recognize the disease. Feyisetan (1997) indicates that
mothers' education and the likelihood that she would know the correct causes of the
diseases are positively correlated. Knowing the causes, signs, and symptoms of diarrhoea
will help mothers to take care their children.

A kind of misperception of the causes of diarrhoea, identifying it as a general
teething problem, still occurs among mothers, as Iyun and Okes’s (1999) research had
found in Ondo State, Nigeria. Pitts et al, (1996) reported that education affects mother’s
knowledge about diarrhoea, causes and prevention. Pitts's et al. conclusion from an
ethnographic study in Shona people, Zimbabwe, mentioned that less educated mothers
believed that diarrhoea is caused by superstitious forces, hence they tended to seek
traditional healers in order to control the diseases, or utilized inappropriate treatments.

Although mothers’ knowledge may influence their practice to manage their
children’s diarrhoea, there is evidence of how mothers' knowledge about diarrhoea was
not appropriate with their practice to manage their children’s diarrhoea. Research which
was conducted by Iyun and Oke (1999) found that a few women in salt water swamps
area (which forms part of the oil-rich coastal region of Nigeria), who claimed to have
knowledge about salt-sugar solution (SSS), could not provide correct description of how
to prepare SSS and usage of SSS is still nil. These findings highlight the importance of

Jonas Sumampouw Literature Review / 16

educating mothers about home treatment of diarrhoea and when to seek medical care for
their children although the area is economically viable for multinational oil companies.

Another research done by Cuevas et al., (nd) in Tlaxcala, Mexico, found that the
reason for not looking for medical care was related to mother’s perception that the illness
was mild or self-limited. Nevertheless, Granich et al. (1999) found from their study in
Chiapas, Mexico, that parents education was not significantly associated with difference
in home remedies utilized, delay in seeking out-of-home care, number of out-of-home
options utilized or utilization of traditional versus allopathic medical options. However,
inappropriate treatment of children’s diarrhoea given by mothers is not merely caused by
lack of mothers’ knowledge towards diarrhoea, but also from health providers such as
physicians. Some physicians in New South Wales, Australia (three to six per cent)
prescribed drugs like anti diarrhoeals and anti emetics (Porteus et al., 1997), which is not
recommended by WHO. WHO (1994) stated, "Anti motility drugs like opiates, codeine,
loperamide are NOT indicated in the treatment of diarrhoea. They greatly slow intestinal
peristalsis, delay the elimination of causative organism and can be very dangerous (even
fatal) in infants". In many countries retail pharmacies or drugs stores are visited
frequently by people with diarrhoeal problem but without a prescription from a doctor,
particularly in rural areas with fewer doctors (WHO, 1994). This behaviour could have
happened because retail pharmacies or drugs stores are widely distributed geographically.
Lack of knowledge is one of the main reasons why people, either the drug sellers or
customer, act as they do.

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 17

2. 4 Socioeconomic Status and Health Seeking Behaviour
While socioeconomic factors explain the under-utilization of government health

services in rural Indonesia, they do not explain the patterns of utilization or the reasoning
which underlie them (Grace, 1998). To study the level of utilization of various health care
options to manage diarrhoeal disease, some researchers (Bojalil et al., 1994, Hudelson,
1993) have raised the question as to why parents seek treatment for their infants and
young children at local healers, health services, or other health care options.

According to Gochman (1988), attempts to relate socioeconomic status and health
behaviours have employed essential measures that reflect individual income or
occupational level or both. Social status is an individual's position in any system of social
ranking. People can be stratified according to such dimensions as class, ethnicity, age and
gender (Freund, 1995). Social class influences how much power individuals have to
manage their bodies and their external environment (ibid), and also influence their health
seeking behaviour for their children’s illness. Granich et al. (1999) and Develay (1996)
found that socioeconomic status (SES) influenced mother’s health seeking behaviour.
Lower SES households tend to rely heavily on local options (local clinics, traditional
healers, and grocery stores) while higher SES households, especially those with vehicles,
utilize more distant options, especially private physicians. By contrast, Costello (1996)
found in Philippines, that lower SES couples accept oral rehydration therapy (ORT) and
community health stations more than wealthier and better-educated couples.

Jonas Sumampouw Literature Review / 18

A study by Kaye et al (1994) in two slum communities in Jakarta found that there
was a close relationship between family income and incidence of diarrhoea. There was
also the possibility of poor mothers who have children with diarrhoea being more likely
to bring their children to the posyandu (community health integrated service post).
Access to health services and ability to pay for the visit or prescribed drugs probably are
explanations for the difference as to why people with higher SES status seek private
physicians or clinics and people with lower SES tend to use government health facilities.
However, that difference also depends on the severity of the disease. If the child was
severely dehydrated by diarrhoea, preference for care from a private physician was
particularly strong. If the child was viewed as very sick, mothers were more likely to take
their children to a private doctor or to a government health facility. The underlying logic
of the decision parents make about when and where to seek treatment for their children
must be sought not only through socio-economic factors, but also in local perception of
practitioners and the appropriateness of the treatment they give for particular illnesses.

2. 5 Mothers' perception of illness and care seeking.
It is generaly agreed that a symptom is finally acknowledged by mothers as illness

and medical care sought when mothers can no longer tolerate or accommodate it. Prior
studies had been conducted to obtain explanation about mothers' perception of illness and
their decision making to treat their children with diarrhoea. According to de Zoysa, 1998,
mothers in urban slum India recognized the diarrhoea based on their identification of a

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 19

pattern of signs, such as a change in the frequency, consistency, colour or smell of stools.
Mothers in that area were not concerned if the young infant passed one or two watery
stools a day, as the stools of young infants who are exclusively breastfed are often
frequent and runny. But they worried, if the stools were more frequent, or if there were
other signs such as vomiting or fever, or if the episode did not resolve after two or three
days. However, according to Granich et al. (1999), perceived diarrhoeal severity did
influence out-of-home health seeking behaviour particularly for the most severe type of
diarrhoea.

Using traditional healers was the initial option sought by most people. This is
consistent with other studies, showing that aetiology and severity influence treatment
selection, especially the use of traditional healers (Mull and Mull, 1988). Levine (1992),
found that in Chile symptomswhich caused the family to seek outside help, from a
modern practitioner or a traditional healer, include vomiting, listlessness or loss of
appetite. But the duration of diarrhoea was also one of important factors affecting
mothers' choice of care. The longer the diarrhoea continued, the more likely it was that
the family took the child to a doctor. In Indonesia, a morbidity and diarrhoea care survey
in 1993, which was conducted by Ministry of Health, showed that the signs of children's
diarrhoea that prompted mothers to take them to a health provider is watery stools,
followed by not getting well, marked thirst and blood in stools (MOH, 1996).

Jonas Sumampouw Literature Review / 20

2. 6 Age of Mothers And Health Seeking Behaviour to Treat Childhood Diarrhoea
Younger mothers and those with no education are less likely to give increased

fluids to their children during a diarrhoeal episode (Al-Mazrou, 1995), so that may
increase the mortality of child caused by diarrhoea. Langsten (1995) also found that
younger mother in Tlaxacala, Mexico, somewhat less likely to use ORS than older
mothers. Neverthelss, Langsten found that majority mothers who were older, have no
education and from poorer family sought care from the modern treatment.

2. 7 Residence and Health Seeking Behaviour
Increased distance between residents and health care providers is commonly

It is obvious that in .(2000 ,Nemet) thought to decrease the utilization of health care
health facilities and health providers as well as pharmacies are ,many countries
This situation explained why mothers particularly who lived .concentrated in urban areas
tional healer than other health carein the remote areas tend to bring the children to a tradi
a study which was conducted by Develay ,Nevertheless .options in case of emergency
mentioned that no significant association was found between the choice of health ,(1996)
.s expectations'o the Develaywhich was contrary t ,care and residential zone

found that there was a general pattern of (1995 ,Langsten) A study in rural Egypt
Mothers’ health seeking behaviour .care and treatment in Upper Egypt and Lower Egypt
t was found from a study thatA similar resul .was improved by increasing the knowledge
sample 747 per cent of 92 They mentioned that .(nd) et al ,was conducted by Cuevas

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 21

there is no ,Thus .mothers in one state of Mexico used liquids to prevent dehydration
mothers in rural and urban difference in pattern of health seeking behaviour between
These studies have also shown that pattern of health seeking behaviour was .areas
mothers in the urban areas are more ,However .influenced by increased the knowledge
.(1997 ,Feyisetan ) slikely to have adequate knowledge of the causes of the disease
.Having the knowledge of causes of the disease will make them seek appropriate options

Supardi (1997), found that in Indonesia, particularly in the two villages at South
Sumatera province, self-medication might potentially be an efficient self-help means for
the community before getting the opportunity to reach the primary health center
(puskesmas). Drugs were used more frequently than the traditional medicines.
Nevertheless, the study result above is contradicted with a study that was conducted in
two villages in a coastal area of Java (Satoto, 1988). In that area, mothers prefer to use
traditional herbs to treat their children's diarrhoea, administered either by rubbing onto
the skin or orally and they will seek modern medical treatment only if the traditional
herbs do not work after one or two days.

2. 8 Conceptual Framework

Many models had been developed by researchers to explain health behaviour. This
conceptual framework of this study is based on Kroeger’s (1983) model, cited in
Develay, et al.(1996), which is usually employed for developing countries to explain the
process by which a sick person will select health care among the choices available. The

Jonas Sumampouw Literature Review / 22

model has shown that factors, which might influence health-seeking behaviour, are socio-
demographic characteristics, characteristics of the perceived illness and health services
characteristics (appendix 3). Kroeger's model is utilized for this study since it is quite
simple to implement and almost all of the variables needed are available from this data
set, except the health services characteristics, which is one of the limitations of this study.

This study expects that mothers' choice of care to control their children's diarrhoea such
as modern treatment, traditional healer, self-treatment and no advice/treatment is
influenced by mothers' socio-demographic characteristic such as age, residence,
education, religion, occupation, children characteristics such as sex, age, duration or days
of children's illness, and diarrhoea related factors such as knowledge about oral
rehydration therapy (ORT), accessibility of source of care and symptoms signs related to
diarrhoea that mother can recognize, as shown in figure 1.

2. 9 Hypothesis
The study hypothesis is:
The mothers' choice of care to treat children's diarrhoeal disease is affected by

mothers' characteristics such as age, education, occupation, residence, religion; children’s
characteristics such as sex, age, days of illness; and the mothers’ diarrhoea related factors
such as knowledge of oral rehydration therapy (ORT), accessibility of sources of care,
and symptoms and signs.

Fac. of Grad. Studies, Mahidol Univ. M.A. (Pop. & Repro.H. Res.) / 23

INDEPENDENT VARIABLE DEPENDENT VARIABLE

Mothers' socio demographic CHOICE OF CARE
characteristics - Modern treatment
- Self Treatment
- Age - Traditional Healer
- Education - No Advice / Treatment
- Residence
- Religion sought
- Labour force status

Children's characteristics

- Sex
- Age
- Days of illness

Diarrhoea related factors

- Knowledge About ORT
- Accessibility of source of

care
- Symptoms and signs

Figure 1: A conceptual framework for studying factors associated with mothers' health
seeking behaviour for treatment of childhood diarrhoeal disease.

Jonas Sumampouw Research Methodology / 24

CHAPTER III
RESEARCH METHODOLOGY

In this chapter the source of data, sampling method, sample size, research
instruments, limitations of study, data analysis and operational definitions of the variables
are discussed.

3. 1 Source of Data
This study utilized data from The Indonesian Demographic and Health Survey

(IDHS). The IDHS was conducted in September to December 1997. The 1997 IDHS was
designed as a collaborative effort of four institutions: the Central Bureau of Statistics
(CBS), the State Ministry of Population/National Family Planning Coordinating Board
(NFPCB), the Ministry of Health, and Macro International Inc. Maryland USA. The
sample was designed to produce estimates at the national, urban-rural, regional and
provincial level.

3. 2 Sampling Method.
The sample is stratified by province and by urban and rural domain within each

province. The sample was selected in three stages. In the first stage, census enumeration
areas (EAs) were selected systematically with probability proportional to population size.
The second stage, in each EA, segments of approximately 70 contagious households with

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 25

clear boundaries were formed, and only one segment was selected with a probability
proportional to size. In the third stage, 25 households were selected from each segment
using a systematic sampling. A complete listing of all households in the selected
segments was carried out prior to the selection of households. Since the sample was
designed to produce estimates at the provincial level, the households selected at the
provincial level did not constitute a proportional representation at national level. The
results presented in this study are based on data that were weighted to take account of
differential sampling probabilities and non response at both the household and individual
levels. The weights are used to produce estimates that are representative at the
appropriate level of aggregation ( for example, provincial, regional, and national).

3. 3 Sample Size.
The 1997 Indonesian Demographic and health Survey (IDHS) covered 27

provinces and a total of 35,362 households were selected for the survey, of which 34,656
were found. Of the encountered households, 34,255 were successfully interviewed. In
these households, 29,317 eligible women were identified, and complete interviews were
obtained from 28,810 women. In all, 15,433 children under the age of five were covered
during the survey.

For this thesis, the children data file from DHS was utilized. However, only a
sample of 1,603 children under the age of five who had diarrhoea two weeks preceding
the survey were selected for this study.

Jonas Sumampouw Research Methodology / 26

3. 4 Research Instruments
The IDHS 1997 used three questionnaires: the household questionnaire, the

questionnaire on family welfare, and the individual questionnaire for ever-married
women 15-49 years old. Only the women questionnaires were used for this study
(Appendix A). The details of the specific question used can be seen in appendix B. Data
were collected by teams of interviewers, who had been trained for 16 days during June to
August 1997. Data collection took place from September to December 1997.

3. 5 Limitations of Study
Mothers' health seeking behaviour to control their children's diarrhoea is not an

individual phenomenon. Other factors such as socio-cultural beliefs, as well as the
influence of parents and peers that affect mother's decision making about choice of care.
The unavailability of information regarding these factors is one limitation of this study.
When drawing a general conclusion about patterns of mothers' health seeking behaviour
towards diarrhoea, these limitations must be taken into consideration.

Another limitation of this study has to do with the sequence of actions that mothers
chose in order to treat their children's diarrhoea. The mother's second option of
treatment, after the first option was not improving the child's condition, was not included
in the data set. Additionally, the 1997 IDHS no longer investigated the availability of
family planning and health services as in the 1994 survey. Since information about the

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 27

sequence of treatment and health services characteristics are not available, therefore it is
difficult to fully explain all factors that may affect mothers' choice of care in general.

3.6 Data Analysis
The Statistical Package for Social Science (SPSS) and STATA have been used for

the purpose of data analysis. SPSS has been used to describe the background
characteristics of the sample population. In addition, STATA was utilized to perform
multinomial logistic regression analysis to assess the effect of the mothers' and children's
characteristics on mothers' health seeking behaviour for treatment of childhood diarrhoea,
which is a multicategory dependent variable. The usual presentation of the multinomial
logit result includes a table of coefficients, with each choice being compared with a
single reference category. It is more desirable to present the implications of the
multinomial logit coefficients in terms of expected probabilities. The table of
probabilities gives a more intuitive meaning by presenting the change in probability of
each category of the dependent variable.

3.7 Operational Definitions
Choice of care : . Choice of care is mothers’ utilization of health care options to

manage their children’s diarrhoea. It is categorized into four groups, such as modern
treatment (includes health facilities namely, hospital, health center (Puskesmas), health
post (Posyandu), private clinic, village delivery post; and health providers namely,
doctor, nurse, midwife, and health cadre); traditional healers (spiritual healers, religious

Jonas Sumampouw Research Methodology / 28

healers, or herbalist healers); self-treatment (buy the medicines from pharmacy or shop),
and no advice/treatment sought at all.

Age: Age refers to the current age of mothers who had children with diarrhoea
within two weeks prior the survey. The first category is age 15 to 19 years, the second
category is 20 to 24 years, the third category is 25 to 29 years, the fourth category is 30 to
34 years, and the last category is 35 to 49 years.

Labour force status: Mothers' status of labour force has two categories, not
working and working (includes professionals, technicians, managers and administrators,
clerical, sales, services, agricultural worker, and industrial worker).

Education: Mothers' highest level of education is classified into three categories:
no education, some primary education, and completed primary education or higher.

Residence: It is categorized as rural and urban.

Religion: Religion is classified as Moslem and non-Moslem. Other religions such
as Protestant/Christian, Catholics, Hindus and Buddhists are categorized as Non-Moslem
since there were too few cases to be categorized separately.

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 29

Age of child: Age of child with diarrhoea has five categories, children less than
12 months, children 12 to 23 months, children 24 to 35 months, children 36 to 47 months,
children 48 to 59 months.

Days of illness: Days of illness refers to the duration of illness that children had
two weeks prior the survey. It is classified as within 24 hours (one day), two to six days
and seven days or more.

Knowledge about ORT: Knowledge about ORT refers to mothers’ knowledge
about oral rehydration therapy for diarrhoea treatment. It was classified as yes, for
mothers who have heard of or seen ORS packets or ever prepared ORT, and no,
otherwise.

Accessibility to source of care: Mothers' difficulty of accessing a source of care
to obtain ORALIT (packets of oral rehydration salts commonly used to treat diarrhoea in
Indonesia) for diarrhoea was used as proxy indicator to measure mothers’ accessibility of
health care options to manage their children’s diarrhoea. It is categorized as easy and
difficult.

Symptoms and signs: Symptoms and signs refers to mothers’ recognition of the
symptoms and signs of their children’s diarrhoea that they should be given treatment.
This variable has seven categories, namely repeated watery stools with/without blood,

Jonas Sumampouw Research Methodology / 30

repeated vomiting, fever, not eating/drinking/marked thirst, getting sicker, not getting
better, and no response.

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 31

CHAPTER IV
FINDINGS

This chapter presents the findings of the study. The first section presents the
background of the sample children and their mother. Description of diarrhoea related
factors was also included in this section. The second section presents the results of the
multinomial logistic regression model, which is used to assess the effect of the
independent variables on mothers' health seeking behaviour for treatment of childhood
diarrhoea, which is a polytomous dependent variable. The selected independent variables
include mothers' characteristics, such as the socio-demographic background, accessibility
of source of treatment, mothers' knowledge of oral rehydration therapy (ORT), and
children's characteristics such as children's sex, children's age, days of illness and their
symptoms and signs of diarrhoea episode.

4. 1 Background characteristics
4. 1.1 Mothers' socio-demographic characteristics.

The majority of mothers lived in rural areas (76 per cent). Most of them were
Moslems (88 per cent) and their average age was 28.4 years. Their education was still
quite low. Only 28 per cent completed primary education or higher and about two fifth
were working.

Jonas Sumampouw Findings / 32

Table 1. Percentage distribution of socio-demographic background of mothers

Socio-demographic number per cent
characteristics

Age 15-19 121 7.5
20-24 398 24.8
25-29 407 25.4
30-34 364 22.7
35-49 313 19.5
x = 28.39 S.D.= 6.74
1,603 100.0
Total

Residence Urban 384 23.9
Rural 1219 76.1
Total 1,603 100.0

Education No education 122 7.6
Some primary educa- 1,029 64.2
tion
Completed primary 452 28.2
education or higher
Total 1,603 100.0

Religion Moslem 1.410 88.0
Non Moslem 193 12.0
Total 100.0
1,603

Labour Not working 981 61.2
force Working 621 38.8
status
Total 1,603 100.0

4. 1.2 Knowledge of ORT, Accessibility of Sources of Care, Symptoms and Signs,
and Choice of Care
The recommended treatment for diarrhoea is oral rehydration therapy (ORT),

including a solution prepared from ORS packets (prepackaged oral rehydration salts).

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 33

The majority of mothers knew about ORT. More than three fourths of mothers reported
that they had ever heard or seen or prepared ORALIT.

The ease or difficulty of getting to a place which provides ORALIT, the brand of
oral rehydration solution most commonly used in Indonesia, was used as a proxy
indicator to measure mothers' accessibility of health options to manage their children's
diarrhoea. The majority of mothers (71 per cent) expressed that it was easy to go to a
place that provided ORALIT.

A question was asked to find out where mothers sought advice or treatment for
their children's diarrhoea. The responses were classified into four categories, namely
modern treatment, which included health facilities and health providers; self-treatment
that included buying medicines from pharmacies or shops; traditional healer, and no
advice/treatment sought.

About one third of mothers perceived that watery stools were the symptom which
indicated that children should be given treatment, followed by not getting better and
getting sicker.

Table 2. shows the options that mothers sought to treat their children with
diarrhoea. Over half of mothers (54 per cent) used modern treatment for their children's
diarrhoea whereas 19 per cent bought medicine from pharmacies or shops to treat their

Jonas Sumampouw Findings / 34

children's diarrhoea by themselves. Only few mothers (4.5 per cent) went to traditional or
other healers, and about one fourth gave no treatment or sought no advice.

Table 2. Percentage distribution of knowledge of ORT, accessibility to source of

care, symptoms and signs, and choice of care.

number per cent

Knowledge of ORT Yes 1,257 78.4
No 346 21.6
Total 100.0
1,603
Accessibility of Easy 70.6
source of care Difficult 1132 29.4
Total 471 100.0
1,603
Symptoms and watery stools with 29.84
signs /without blood 478
vomiting 6.00
fever 96 7.87
not eating / thirst 126 4.59
getting sicker 74 21.09
not getting better 338 26.05
orther 418 4.55
response/don't
know 73 100.00
Total
1,603 54.2
Choice of care Modern treatment 18.7
Self treatment 870 4.5
Traditional or other 300 22.5
No advice/treatment 100.0
Total 72
361
1,603

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 35

4. 1.3 Children's characteristics.
Table 3 shows children's characteristics related to diarrhoeal disease. There were

more male than female children (56 per cent : 44 per cent) who had episodes of
diarrhoea. Over three fourths of them were under 3 years old.

The proportion of getting diarrhoea for older children of age 36 to 47 months and
48 to 59 months decreased respectively. This result can be interpreted that children under
three years of age have higher risks of getting diarrhoea disease compared to older
children.

In the majority of cases, the duration of children's diarrhoea was between two to
six days with an average of 3.44 days.

Table 3. Percentage distribution of children's characteristics

Children characteristics number per cent
903 56.3
Sex Male 670 43.7
Female 100.0
Total 1,603
24.8
Age ≤ 11 months 397 28.2
12-23 months 452 23.5
24-35 months 377 12.7
36-47 months 204 10.8
48-59 months 173 100.0
Total 1,603
13.35
Days of 0-1 day 214 72.34
illness 2-6 days 1160 14.31
7 days or more
x = 3.44 S.D =2.60 229 100.0
Total
1,603

Jonas Sumampouw Findings / 36

4.2 Factors associated with mothers' health seeking behaviour for treatment of
childhood diarrhoea.

Choice of care is the dependent variable in this study. It consists of four categories,
namely, no treatment, self-treatment, traditional healer and modern treatment. In this
study respondents who used more than one form of treatment are not identified. Since the
dependent variable is a categorical variable with four unordered mutually exclusive
categories, the multinomial logit technique is appropriate (Agresti 1990).

In the multinomial logit model, to estimate the effects of independent variable on
different categories of the dependent variable, one category is used as the reference
baseline. The choice of the reference category is irrelevant for estimation and should be
determined on substantive grounds (Liao, 1994). In this study, modern treatment is used
as the reference category. This category has the highest frequency among all the options.
It is also more reasonable to explain the option that mothers made according to modern
treatment, which is an ideal option. The likelihood of choice of no treatment over modern
treatment, the likelihood of choice of self-treatment over modern treatment, the
likelihood of choice of traditional healers over modern treatment will be compared.

Table 4 presents the full set of multinomial logit coefficients estimates respectively
for the analysis of choice of care.

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 37

Table 4. Multinomial logit coefficients for the model of choice of care

Independent variables No treatment Self treatment Traditional healer
Vs Modern Vs Modern Vs Modern
Age of mother treatment treatment treatment
15 - 19 years
20 - 24 years 0.874 -0.435 -1.604
25 - 29 years 0.064 -0.309 -1.008 *
30 - 34 years 0.298 -0.604 * -0.960
35 - 49 years @ 0.306 -0.234 -0.558

Education 0.750 0.070 2.104 ***
0.415 0.049 0.009
No education

some primary education

completed primary education
or higher @

Residence 0.020 -0.412 -0.011

Urban
Rural @

Labour force status -0.183 -0.355 -0.089

Not working
working @

Religion -0.157 -0.433 -0.507

Moslem
Non Moslem @

Sex of children 0.130 0.213 0.295

Male
Female @

Children's age 1.317 ** -1.035 * 0.989
less than 11 months 0.224 -0.465 0.892
12 - 23 months 0.425 0.379 0.802
24 - 35 months 0.617 0.379 0.079
36 - 47 months
48 - 59 months @

Jonas Sumampouw Findings / 38

Table 4. (Continued) No treatment Self treatment Traditional healer
Vs Modern Vs Modern Vs Modern
Independent variables treatment treatment treatment

Duration of illness -1.791 *** -0.818 * -1.658 **
0 - 1 day -0.828 * -0.436 -0.293
2 - 6 days
7 days or more @ -0.538 -0.772 -2.170 **

Knowledge of ORT 0.356 0.067 -1.163
No knowledge
Have knowledge @ -0.337 0.382 -0.492

Accessibility of sources of care -0.029 -0.208 -0.048
Easy -0.411 -0.399 -2.442 *
Difficult @ -0.260 -0.863 0.064

Symptoms and signs 0.148 -0.189 0.288
Watery stools with/without 0.288 0.135 -1.697 *
blood
Repeated vomiting
Fever
Not eating/drinking/marked
thirst
Getting sicker
Not getting better
No responses @

Notes: *, ** and *** indicate statistical significance at the 0.05, 0.01 and 0.001 respectively
@ is reference category.

Table 4 shows the multinomial logit coefficients of the likelihood of choosing a
source of care. Of the eleven independent variables, mothers' age, highest level of
education, knowledge of ORT, children' age, duration of illness and symptoms and signs
were statistically significant affecting mothers' choice of care. Mothers aged 20 to 24
years were less likely than mothers aged 35 to 49 years to seek traditional healer than

Fac. of Grad. Studies, Mahidol Univ. M.A.(Pop. & Repro.H. Res.) / 39

modern treatment, while mothers aged 25 to 29 years were less likely to administer self
treatment for their children's diarrhoea than modern treatment.

Table 5 shows that for mothers aged 15 to 34 years the probabilities of mothers
buying medicine from pharmacies and shops to administer self-treatment for their
children's diarrhoea decreases as their ages increase. In addition, the older the mothers
are, the higher probabilities of seeking no treatment

Table 5. Predicted probabilities of choice of care

Independent variables No treatment Self Traditional Modern
treatment healer treatment

Age of mother 0.399 0.114 0.009 0.478
15 - 19 years 0.419 0.093 0.004 0.483
20 - 24 years 0.482 0.056 0.002 0.462
25 - 29 years 0.538 0.042 0.001 0.419
30 - 34 years 0.170 0.233 0.078 0.518
35 - 49 years

Education 0.297 0.144 0.168 0.392
No education 0.323 0.140 0.161 0.376
some primary education
completed primary education or 0.181 0.195 0.035 0.587
higher

Residence 0.307 0.178 0.156 0.359
Urban 0.328 0.128 0.163 0.381
Rural

Labour force status 0.335 0.141 0.161 0.363
Not working 0.316 0.112 0.166 0.407
Working

Jonas Sumampouw Findings / 40

Table 5. (Continued) No treatment Self Traditional Modern
treatment healer treatment
Independent variables
0.315 0.116 0.171 0.398
Religion 0.319 0.090 0.126 0.465
Moslem
Non Moslem 0.221 0.177 0.042 0.560
0.231 0.200 0.050 0.520
Sex of children
Male 0.420 0.066 0.073 0.442
Female 0.440 0.041 0.113 0.405
0.477 0.044 0.152 0.327
Children's age 0.589 0.043 0.124 0.245
less than 11 months 0.137 0.239 0.027 0.595
12 - 23 months
24 - 35 months 0.456 0.100 0.003 0.441
36 - 47 months 0.500 0.102 0.003 0.396
48 - 59 months 0.130 0.156 0.037 0.677

Duration of illness 0.284 0.094 0.354 0.268
0 - 1 day 0.314 0.085 0.098 0.504
2 - 6 days
7 days or more 0.314 0.085 0.098 0.504
0.256 0.054 0.017 0.674
Knowledge of ORT
No knowledge 0.183 0.236 0.033 0.548
Have knowledge
0.187 0.207 0.033 0.574
Accessibility of sources of care 0.155 0.172 0.004 0.669
Easy
Difficult 0.141 0.089 0.005 0.764

Symptoms and signs 0.156 0.077 0.007 0.761
Watery stools with/without 0.190 0.083 0.001 0.726
blood 0.237 0.177 0.071 0.532
Repeated vomiting
Fever
Not eating/drinking/marked
thirst
Getting sicker
Not getting better
No responses