Determinats of commune health center (CHC) usage in long an province

INSTITUTE OF SOCIAL STUDIES

NATIONAL UNIVERSITY OF HCM
UNIVERSITY OF ECONOMICS

THE HAGUE
THE NETHERLANDS

VIETNAM

VIETNAM-THE NETHERLANDS
PROJECT FOR M.A ON DEVELOPMENT ECONOMICS

DETERMINANTS
OF COMMUNE HEALTH CENTER (CHC) USAGE
IN LONG AN PROVINCE

A thesis submitted in partial fulfillment of the requirements for the degree of
MASTER OF ARTS IN DEVELOPMENT ECONOMICS
BY
HUYNH DANG BICH VY

Academic Supervisor:
DR. NGUYEN VAN PHUC
BQ GIAO DVC VA 8AO -;-; ; .
TRVdNG E>H KINH TE TP.I·ic;~

THUVIEN

j;

. ( ~c

HO CHI MINH CITY, AUGUST 2007

}f-


ACKNOWLEDGEMENTS

Firstly, I would like to express my deep gratitude to all professors and
teaching staff in Vietnam-Netherlands programme on Development Economics
for their lectures, instructions and the best teaching conditions during my study
period from 2004 to present.
The author would like to give a special thanks to Dr. Nguyen Van Phuc, the
author's supervisor, for his scientific instructions and his valuable comments on
this study. In addition, the author owes special gratitude to Mr. Truong Dang Thuy
and Mr. Luong Vinh Quoc Duy for their enthusiastic help and criticism.
This research would have been completed difficultly were it not for the
kind and warm welcome of individuals from over 100 households in Can Duoc
and Can Giuoc district during the survey. Thanks are due to many friends that I
could not fully list here for their strong supports of the survey and their invaluable
encouragements and nice wishes.
And last but not least, all my love is devoted to my parents and brother who
always help and encourage me during my learning and doing this study.
Again, the author is really grateful to all people for their help. Any errors in this
research are my responsibility alone.


CERTIFICATION

I certify that the substance of this thesis has not already been submitted for
any degree and is not being current submitted for any other degree.

I certify that to the best of my knowledge any help received in preparing
this thesis, and all sources used, have been acknowledged in this thesis.

HUYNH DANG BICH VY


ABSTRACT

Many studies of the utilization of health care system have been done in
different countries over the world. In most developing countries like Vietnam,
people mainly live in rural area; hence rural health care sector plays an important
role. One of health care providers is commune health center that provides basic
health care. Therefore, this study's purpose is to investigate determinants of
commune health center usage; specifically it aims at examining the effects of
individual income.

Binary logit model was used to find out the answers for the questions of what
determinants of using health care services from CHC are and whether with higher
income people tend to use more CHC services or not. The author has applied the
method of multi-staged sampling to collect data in Can Duoc and Can Giuoc
districts, Long An province. The object of this research is individuals aged 15 and
older that are the adult population and have enough civil capacity to make their own
decisions.

The results show that income is a relatively important determinant of CHC
-----cnoice -in.-can Duoc and-Catlviuo-c-districts:-Furthermore;-other-factors-such-associo-demography, severity of illness and characteristics of the CHC provider are
significantly important. These results are useful to give several recommendations to

improve the quality of CHC services in order to satisfy particular income group.


TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION .................................................................................... I
1.1.PROBLEM STATEMENT ................................................................................................ 1
1.2 .RESEARCH OBJECTIVES ............................................................................................... 2
1.3 .RESEARCH QUESTIONS ................................................................................................ 2
1.4 .METHODOLOGY .......................................................................................................... 2
1.5. HYPOTHESIS .............................................................................................................. 2
1.6 .RESEARCH SCOPE ........................................................................................................ 3
1.7. THESIS STRUCTURE .................................................................................................... 3

CHAPTER 2: LITERATURE REVIEW ......................................................................... 5
2.1.DEFINITIONS ............................................................................................................... 5
2.2. THEORY OF CONSUMER BEHAVIOR .............................................................................. 6
2.3 .THEORIES OF FACTORS AFFECTING HEALTH CARE DEMAND .......................................... 7
2.4.MODEL ....................................................................................................................... 9
2.5.SUGGESTED RESEARCH VARIABLES ............................................................................ 12

CHAPTER 3: AN OVERVIEW OF HEALTH CARE PROVIDER IN VIETNAM ........ 19
3.1.BACKGROUND ON VIETNAMESE HEALTH CARE SYSTEM .............................................. 19

3.1.1.Achievements ....................................................................................................... 19
3.12. Shortcomings ....................................................................................................... 21
3.2.COMMUNE HEALTH CENTER ...................................................................................... 22

CHAPTER 4: RESEARCH METHODOLOGY, ESTIMATION AND RESULS •••••.••••• 27
4.l.DATA COLLECTION METHODS .................................................................................... 27

4.1.1.Sampling technique .............................................................................................. 27
4.1.2.Sample size ........................................................................................................... 28
4.2.DATA ....................................................................................................................... 29

4.2.1.The main contents of the questionnaire ................................................................ 29
4.2.2. Dependent variable ............................................................................................. 30
4.2.3. IndeJJendent variables .......... ._._._., .. ._~·
.._................................................. 30
4.2. 3.1. Socio-demographic characteristics ........ ~.-:_=30
4.2.3.2 Individual and household income ........................................................... 32
4.2.3.3 The characteristics of the perceived illness ............................................. 33
4.2.3.4. Attributes ofhealth care provider and the non-monetary cost ................ 35
4.3.STRENGTH AND WEAKNESS OF COLLECTED DATA ....................................................... 41
4.4.REGRESSION RESULTS ............................................................................................... 42

CHAPTER 5: CONCLUSIONS AND RECOMMENDATIONS ..................................... 50
5.l.CONCLUSIONS ........................................................................................................... 50
5.2.RECOMMENDATIONS ................................................................................................. 51

REFERENCES: ........................................................................................................... 52
APPENDICES: ............................................................................................................ 56
APPENDIX 1: ............................................................................................................. 56
APPENDIX 2: ............................................................................................................. 60
APPENDIX 3: ............................................................................................................. 64

-


LIST OF TABLES
TABLE 2.1: The expected signs ofdeterminants ofhealth care usage ................................ 17

TABLE 3.1: Infrastructure in three levels ofpublic health care system in 2006 in
Vietnam ............................................................................................................................... 23
TABLE 3.2: The proportion of out-patient health service contacts of each level of income
people at each type ofprovide ............................................................................................ 25
TABLE 4.0: Appropriate sample size ................................................................................ 29
TABLE 4.1: The number and the percentage ofobservations for each type ofproviders ..... 30
TABLE 4.2: Basic socio-demographic characteristics ofthe respondents .......................... 31
TABLE 4.3: Average income of respondent and average income per capita per month .... 32
TABLE 4.4: Mean individual income offive groups .......................................................... 33
TABLE 4.5 Choice ofprovider by income quintile ............................................................. 33
TABLE 4.6: Results ofthe survey CHC andnon-CHC respondents in terms oflDAY ........ 34
TABLE 4.7: Results of the survey CHC and non-CHC respondents in terms ofRDAY........ 35
TABLE 4.8: Availability ofdrugs in the CHC..................................................................... 36
TABLE 4.9: Total expenditure for consultation and drugs ................................................. 37
TABLE 4.10: The number of observations that people have used specific health care
provider ............................................................................................................................. 38
TABLE 4.11: Data definitions ........................................................................................... 39
TABLE 4.12: Logit regression results ............................................................................... 43
TABLE 4.13: The changes in the probability of using CHC provider
---- -ageCJrlvofductin;hs=.~,•

if having change in
........••-...,••,-.-.-.-.-.•••••••• 44--

TABLE 4.14: The changes in the probability ofCHC usage of male and female if age
increases by one year, given other factors ........................................................................ 45
TABLE 4.15: The changes in the probability ofCHC usage of male and female if the
ratio ofdistance increase by one unit, given other factors ................................................. .45
TABLE 4.16: The changes in the probability ofCHC usage ofmale and female if the user
fee ofCHC is cheaper, given other factors ......................................................................... 46

TABLE 4.17: The changes in the probability of CHC usage if changing severity of
illness, given other factors ................................................................................................... 47
TABLE 4.18: The changes in the probability ofCHC usage as having change in income,
given other factors .............................................................................................................. 48
TABLE 4.19: Descriptive statistics of variables ............................................................. ,.56
TABLE 4.20: Correlations among independent variables ...................................... ............ 58


LIST OF FIGURES
FIGURE 2.1: The behavioral model .................................................................................... 7
FIGURE 2.2: Choice ofa health care provider ................................................................... 8
FIGURE 3.1: Total expenditure on health (THE) as% ofGDP ......................................... 20
FIGURE 3.2: Total expenditure on health in million VND for 1996-2005 .......................... 20
FIGURE 3.3: Inequality in the use ofhealth services ......................................................... 22
FIGURE 3.4: Proportion ofout-patient health care service visits by type ofprovider ......... 24
FIGURE 3.5: Main reasons for choosing outpatient facility used ...................................... 25


ACRONYMS

CHC

Commune Health Center

GSO

General Statistics Office

LHS

Left Hand Side

MOH

Ministry of Health

MTEF

Medium-Term Expenditure Framework

RHS

Right Hand Side

SD

Standard Deviation

UNDP

United Nations Development Programme

UNICEF

United Nations Children's Fund

VDR

Vietnam Development Report

VLSS

Vietnam Living Standards Survey

VNHS

Vietnam National Health Survey

WB

World Bank

WHO

World Health Organization


CHAPTER I
INTRODUCTION

The introduction chapter consists of six parts. The problem statement is presented
in section 1.1. The following parts such as research objectives, research questions,
methodology, and hypothesis are in section 1.2, 1.3, 1.4, and 1.5, respectively.
Research scope as well as several reasons for choosing Can Duoc and Can Giuoc
districts as research places is revealed in section 1.6. The last section is thesis
structure.

1.1 Problem statement

Health sector is one of main concerns of many countries in the world as

well as in VietNam. During the period since Doi Moi (Renovation) to the present,
the government has invested resources and implemented many reforms in this
sector.
The rural heath care system is strengthened as the government's priority
because around 72.9% of 84,108 million Vietnamese people live in rural area
(UNDP, 2006). It is organized at four levels: national, province, district, and
commune (Tran Tuan, 2004). Together with private providers and self-medication,
almost all communes have a commune health centre (CHC). According to Nanak
Kakwani and Hyun H:-Son -(2U01i)~CH

seems to play an-impoftantrole_in__ -

providing basic health services to the poor in Viet Nam. In particular, CHC is
responsible for primary curative care service in a community, especially in rural
area. However, the role of the commune health center has declined. One of the
reasons of this is that a large part of primary curative care service is shared with
other providers. Another reason is that people tend to seek higher quality health
care services because of an improvement of living standard. The CHC system only
retains its dominant role in preventive care (Tran Tuan, 2004). This study
investigates determinants of CHC usage and finds out whether CHC is normal or
inferior goods, especially in two districts of Long An province, namely, Can Duoc
and Can Giuoc. This kind of information would help decision makers develop
1


appropriate plan for CHCs. Since, suppliers could provide appropriate health care
services and contribute to conduct an accessible and affordable health service to
consumers. Therefore, this research concentrates on investigating determinants of
commune health center usage in Can Duoc and Can Giuoc districts, Long An
provmce.

1.2 Research objectives

General objective is to examine the factors affecting commune health center
usage in Can Duoc and Can Giuoc districts, Long An province
Specific objectives are to measure the effect of individual income on CHC
usage and to propose appreciate plan for CHCs to satisfy the demand for health
care services of people living in the two research places.
1.3 Research questions

A general question is: what are determinants of using health care services
fromCHC?
And the other is: Are individuals with higher income more likely to use
health care service from CHC?
1.4 Methodology

Descriptive methods are used to review the background conditions of
Vietnamese health care system, especially health care information from survey in
Long An province. Moreover, an econometric model is formulated to examine the
relative effects of various determinants of CHC usage. In particular, the author
applied binary logit model to find out answers for the above research questions.
1.5 Hypothesis

The study hypothesis is that there is a negative relationship between
individual income and CHC usage in rural area. The higher income is, the higher
demand for health care facilities with high quality such as public or private
hospitals, private clinics is. As a result, demand for CHC usage will reduce.

2

1.6 Scope of the research
Data for this research were collected from households in Can Duoc and Can
Giuoc districts, belonging to Long An province, in March 2007. There are 154
observations collected from over 100 households in the author's survey. The
author chose the two districts to be research place due to following reasons:
Firstly, Can Duoc is the author's home town that motivates me to choose it
in order to do this research. Furthermore, with the support of the author's friend
who lives in Can Giuoc district, to familiar with the geography, living conditions
and culture of local people made collection of data less difficult.
Secondly, Can Giuoc and Can Duoc districts have the same conditions such
as the number of communes that is 17 for each. In addition, the density in the two
districts is also similar, that is 781 and 787 people per square kilometer,
respectively.
Finally, they are rather near to Ho Chi Minh city that have many modem
health services. Respondents living in there can choose various health care
providers that may lead to reduce in CHC usage. Therefore, to examine
determinants ofCHC usage in Can Giuoc and Can Duoc districts is necessary.

1.7 Thesis Structure
The outline of this thesis is organized following:

Cliapter r:Tntroauction.-Tlle problem statement is presented-to---expresstrre---significant for investigating determinants of CHC usage and finds out the effect of
income on the CHC usage of individuals in Can Duoc and Can Giuoc districts,
Long An province.

Chapter 2: Literature Review. This chapter includes definitions of CHC,
theoretical background in terms of consumer behavior, theory and model of factors
affecting health care demand. In addition, results of empirical studies also are
presented in this chapter.

Chapter 3: Overview of Health Care System in Vietnam. This chapter
provides major information on rural health care system such as the gains as well as

3


shortcomings. In addition, infrastructure and reasons for choosing each level,
especially the CHC usages are also contained.
Chapter 4: Research Methodology, Estimation and Results. In this chapter,

there will be the sampling technique, sample size, data description as well as the
strengths and the weakness of data collected. Moreover, regression results and
interpretation of the study will be presented and analyzed.
Chapter 5: Conclusions and Recommendations. This chapter contains a

summary of the findings, and gives some recommendations.

4


CHAPTER2
LITERATURE REVIEW
This chapter includes five main parts. Section 2.1 is concepts that are
defined in order to avoid confusing the meanings of these concepts. The theory of
consumer behavior will be presented in section 2.2. The next part, section 2.3, is
theories of factors affecting health care demand. Section 2.4 deals with the model
of factors in determining household choice of health care provider. The final part,
section 2.5, is empirical studies that illustrate the results of various studies in order
to support the theoretical background.
2.1 Definitions

Commune health center (CHC)
Definition of CHC is based on the official explanation of the structure of
rural health care system in Vietnam by Hung, Anderson et al, in 2000 (Tran Tuan,
2004). CHC is the first level contact for health care in the public health system.
The CHC is responsible for delivery of primary health care in a community. It
includes preventive care, normal obstetrics, drugs, family planning and overall
health improvement in the community. The CHC is guided by the district health
center for technical matters and is supported by the Chairman of the Commune
People's Committee for health development activities in the commune.
---H-ealth-insurance-------------------- -------------------------- ----In Vietnam, according to the study of The Ministry of Health and General

Statistics Office

(201~),

health insurance is a form to carry out the social

welfare policy and considered as a mechanism to gain a just implementation in the
healthcare field through financially supporting people to access medical services,
to be free from any concerns of financial matter when falling ill, especially for low
and average income persons. It is indicated by the Regulations on health insurance
issued in attachment to the Decree No.

58/19N§~CP

of the Government on

health insurance and the Circular guiding health insurance development that there
are two health insurance forms: obligatory health insurance and voluntary health
insurance. Obligatory health insurance is the form in which the main contribution

5


2% is from the State Budget or the employers while the remaining 1% is from the
individuals or the Government in case of retirees, social policy persons. Voluntary
health insurance is the form applied for those people not belonging to the first one.
In addition, there are other forms such as Humanity Health Insurance bought by
international organizations for groups of people such as the Red Cross buys the
health insurance cards for particular sick people ...
In the context of this study, the term "health insurance" is defined to all
kinds of health insurance.
2.2 Theory of consumer behavior

According to Pindyck and Rubinfeld (2000, p.62), theory of consumer
behavior is the description of how consumers allocate incomes among different
goods and services. Given their preferences and limited incomes, the decision
maker always chooses the alternative for which the expected value of the utility is
maximum. In economics, utility is a measure of the relative happiness or
satisfaction gained. In this study, it is assumed that an individual faces budget
constraint and the choice between consuming two kinds of goods as health care
and other goods. The utility function on choosing a specific health care provider is

~-

considered. For maximizing utility, an individual's health care provider choice will
-~-

-~ -~

be CHC provider i if expected utility that person receives from CHC provider
---m--gller-thmrfrom-otherproviders:- -----

Un

=

max (Uu ) where j f. 1

Where i: typically individual i,
j: type ofhealth care providers.
j = 1 means individual choose treatment from CHC provider.
j

= 0 for other providers.

Estimate the probability of choosing CHC provider, we have:

Pr G= 1) = Pr (Uil > UiO)

6


2.3 Theories of factors affecting health care demand
Since the early 1960s, models of health care demand have evolved. Early
health care demand models were simple reduced-form equations that derived from
the assumption of utility maximization (Mwabu et al., 1993 ).
Andersens analyzed three groups of determinants of the choice of a health care
provider that were illustrated in his model of families' use of health services. Three
major components are, namely, predisposing, enabling, and need factors that

showed in figure 2.1 (Andersens, 1995).
Figure 2.1: The behavioral model
Predisposing
characteristics:
Demographic,

Enabling

_____. resource:
Personal/

Social Structure,

Family;

Health Beliefs

Community

_____.

Need:
Perceived

_____.

Use of health
services

Source: Andersens (1995)

In 1995, Andersen systematically reviewed the development of the

behavioral model of health service utilization. Predisposing characteristics
-------------m-cludtrrg-demo-graphic;-social-structure-and-health-beliefs-are-the-:first-gwup.-The
model has been criticized that it, using several variables represents for social
structure such as education, occupation, and ethnicity, did not pay enough attention
to social networks, social interactions and culture. However, Andersen (1995)
affirmed that the measurement of these concepts rightly fit into social structure
components. Besides, health beliefs, according to Andersen (1995), are ill people's
attitudes, values, and health or health-services-knowledge. He found that health
beliefs might be not important as they really be in predicting and understanding
use. The second group is enabling factors which include family and community
resources. He said that enabling resources had to be present for use. For example,
health personnel and facilities should be available at place where people live and

7


work. Moreover, he agreed that income, health insurance, and travel and waiting
times were some of the measures that could be important.
As time passes, many different determinants have been showed from other

researches on medical care demand patterns in developing countries. For example,
Decor and Ruttens (1998), Habtom and Ruys (2003) concentrated on three groups
of determinants of the choice of a health care provider that are showed in the
figure 2.2. The first group, household and individual characteristics, includes age,
sex, education, socio-economic status, income and occupation. The second group
refers to the characteristics of the perceived illness. The third group is the
characteristics ofthe health service: accessibility, quality of service, and user fees.

Figure 2.2: Choice of a health care provider
Population at risk

Variables

affecting

Possible choices
(Dependent variables)

choice
(Independent variables)

Those

All

members

~

becoming characteristics (age, sex,

of

sick

households

injured

in

the

Household & individual r+ Government health care

or education,

facilities

mcome,

occupation,

socto-econ

status)

community ------------ -cnaracteristics ---uf___tlre- - p.-Mtssiun-- health-----careperceived illness (Mild,
Moderate

or

facilities

High

severity of illness)
Characteristics

health

of the
service

(accessibility:

distance,

travel

transport

time,

cost; quality and cost of
care)
Source: Habtom G. and Ruys P. (2003)

8

--. Private

health

care

facilities
--. Self-treatment
(self-care + traditional

healer)


2.4 Model

According to Linderlow (2004 ), recent economic analysis of health care
choices have been derived from the human capital and household production
literature. For example, Grossman, in his studies which were done in 1972,
presented a model in which health entered the utility function. Healthcare, time
and other inputs, and factors in a health production function and the demand for
medical care represented the rational response to a health shock. The health stock
makes people reduce consumption of other goods for medical care in order to
improve their health. Grossman said that an increase in income could lead to an
increase in demand for medical care. There were many empirical studies in
developed countries that have been motivated by Grossman's model. However,
according to Linderlow, in developing countries, data limitations have not
permitted examining based on Grossman's framework.
Therefore, in this research, economic model is based on a simpler static
framework (e.g. Bedi et al., 2003). Bedi et al. (2003), using a standard framework
from Gertler, Locay and Sanderson (1987), Mwabu, Ainsworth and Nyamete
(1993), have examined some factors in determining household choice of health
care provider.
Bedi et al. (2003) have based on different assumptions. The first is that
--------typically individual-onlyconsum:es-two-kinds-of-goods-as-health-care-that-can-be
defined by the expect health status and other goods within income constraint. The
second assumption is that consumers attempt to maximize their utility by choosing
a product with attributes, in this case, choosing among alternative health care
providers that will provide them with the highest amount of utility. The third is the
expect health status of individual which depends on the features of person along
with the quality of treatment received by individual.

Focusing on the individuals with illness, they have used the utility function.
It includes some elements like the personal health status represented by Hij and the
consumption of all other goods denoted by Cij along with Tij as the non-monetary
cost. The utility function on choosing a specific health care provider is given by:

9

-------


Uu = u (Hu, Cu, Tu)

(2.1)

Where i: typically individual i,
j: type ofhealth care providers.
Due to the third assumption, Bedi et al. (2003) have given the form of the
function of health production:

Hij = H (Xb Zj)

(2.2)

Where: Xi and Zj are the attributes of individual

1

and provider j,

respectively.

Moreover, the first assumption leads to the individual or household income
denoted by Yi includes spending on user fees in terms of a visit to health care
provider Pij and consumption of other goods Cij. Consequently,

Cij

= Yi -

Pij

(2.3)

Substituting (2.3) and (2.2) into (2.1) results in the form of utility function
be:

Uij = U (Xb Zj, Yi - Pu, Tij)

(2.4)

Separating the utility function into two parts as systematic part denoted by
vij

that depends on components in equation (2.4) and idiosyncratic part

represented by Eij will results to another form of utility function that is defined as:

Uij- Vu =t-Eij --------{2--:SJ________________________________ _
Following the early literature on health care demand in developing
countries such as Akin et al.(1984), Mwabu (1986), the empirical specification is
based on a linear utility and health production function (Linderlow, 2004) as we

have:

(2.6)
For maximizing utility, an individual's health care provider choice will be

CHC provider if expected utility that person receives from CHC provider highest
among other providers:

UiO = U (Xb Zo, Yi - Pio, T ij) max (Uij ) where j f:. o

10

(2.7)


Estimating the probability of choosing particular health care provider, such
as private government, mission and self-care, Bedi et al. (2003) have used the
multinomial logit model. Although, individuals in rural area, especially in
communes, can choose from among the different options in terms of CHC, private
provider and self-treatment and traditional medical practitioners, the context of the
current study is only about CHC and non-CHC. Non-CHC includes selfmedication, and professional care other than CHC as private treatment. Therefore,
the author used the binary logit model
The problem of choice of health care providers will be solved by using the
above form of utility function for individuals.

(2.6)
The equation (2.6) could be estimated by regressing the following function:

uij =a+ [31Xi + [32 yi + [33Zij + [34Pij + [3s Tij + Eij

(2.8)

The utility function describes the relationship between utility and its
determinants. The former becomes a dependent variable, while the latter become
independent variables for the model.

uij = a + [31Xi + [32 yi + [33Zij + [34Pij + [3s Tij + Eij
The left hand side

(!-HS)_~

dependent variables which are__ dummY

variables. If respondent chooses CHC provider, dependent variable will be
assigned a score of one; and the score of zero represents non-CHC treatment
including private facility, self-treatment and others.
The right hand side (RHS) is independent variables as follows:
Xi: the attributes of individual i
Yi : the individual or household income
Zj : the attributes of provider j
Pij : user fees in terms of a visit to health care provider
Tij: the non-monetary cost
Sij:

idiosyncratic part

11


In summary, choosing a specific health care provider of individuals is

defined by different factors such as the characteristics of individuals,
characteristics of providers, individual income, and user fee in terms of a visit to
health care provider, and non-monetary cost.
2.5 Suggested research variables

The first group of attributes of individuals - focus on the sociodemographic factors of individuals that comprise gender, age, marital status, and
education.
Trivedi (2003) based on the health components of the 1998 Vietnam Living
Standards Survey (VLSS) in order to emphasize on econometric modeling of
demand for different types of health care in Vietnam, especially, for CHC
provider. The negative relationship between CHC usage and age as well as
education is founded. CHCs are typically not used by better educated people.
Moreover, Vietnamese older individuals seem to avoid using CHCs. Similar result
was also indicated by Damen in Ethiopia in 2003 (Habtom, Ruys, 2003). He
concluded that when individuals became older, they were more likely to visit
distant and higher quality of health care facilities than local health stations.
Effects of socio-demographic factors on choices of a health care provider
were founded in many studies. However, not all results are the same. Whether the
- ---------effecrofgenaer, being a m-ale, -onusigh~altcrepvd

-

depends on a particular provider the researcher considers. For example, in rural
Cote d'Ivoire, male effect is positive and significant in the case of doctor visits
while being negative and insignificant in the nurse choice (Dor and Van der Gaag,
1987). It shows that males are more likely to obtain higher quality heath care.
While in Kenya, women were more likely to be treated by all types of health care
provider than self-medication (Mwabu et al., 1993). Therefore, there does not
appear to be any strong link between gender and the type of health care that people

choose.
Besides, other personal factors like religion and ethnicity may effect on
individuals' decisions among treatments.

12


The second group - Attributes of health care provider (Zj) includes
various variables.
Drug availability

There were many indicators of service quality, including the availability of
drugs, the qualification of health care staff, and the availability of equipment
(Mwabu, 1993 ). Because of data limitation of service quality, many researches
used data of the availability of drugs in order to examine the health care demand
model. Betdi et el. (2003) who used information on about 10,000 households and
over 50,000 individuals from almost all districts in Kenya in 1994 showed clearly
the effect of drug unavailability in government facilities on use of health facilities.
They concluded that an increase in the unavailability of drugs sharply reduced the
use of public facilities, leading to an increase in self-treatment.
According to Okello et al. (1998), the hospitals are chosen by people living
both urban and rural area in Uganda rather than the nearest health care because of
unavailability of drugs in the local health center. Another reason for this is to look
for higher level of health quality.
In this research, the author will analyze the effect of availability of drugs on
health care usage and expect that it is positive effect
User fee

--------user fee-is- an ifupoftanCfactor of
moaelf-thcr~;Sv

studies have analyzed the impact of increases in prices on the use of health care.
This negative relationship was proved in Eritrea by Habtom, Ruys (2003), and in
Kenya by Bedi et al. (2003). In Zaire, for instance, a rapid increase in user fees for
health care resulted in a sharp fall in the demand for curative services. A similar
drop in usage was observed in Mozambique (Dercon, Ruttens, 1998). Moreover,
Bedi et al. (2003) also showed that the increase in user fees in public health
facilities might be matched by an increase in the use of private and mission
facilities. As a result, an expected effect of user fees is negative.

13


Distance
The other characteristic of health services is distance, that is, the length
from the person's houses to the particular health care facility. In rural areas,
especially, distance and poor quality of services at commune health center have
been cited as possible reasons for continued growth of using self-treatment. It is
said that distance played an important role in choosing health care providers
(Acton, 1975). This variable is expected to affect negatively on CHC usage.
The non-monetary cost - waiting time
It is well known that even in absence of user fees, access to the health

services is not equal due to non-monetary factors such as travel time (Acton,
1975). Therefore, non-monetary factors are very important for health care usage
model. Economic theory indicates that the more time is spent on a health unit, the
lower the probability of patients choosing that provider. Hobtom and Ruys (2003)
considered total time spent on a health unit made of waiting time, treatment time,
laboratory examination time, and time spent at the pharmacy as non-monetary
factors. In Eritea, travel time has a negative effect on the choice of a health care

service provider for most rural residents although it has a positive sign for public
health facilities, especially for hospital.
In this research, the waiting time variable is a proxy for the non-monetary
cosr.-CHC-isalsoa-type of-h--e-alth-care-provider;-if-waiting-time-for-using-€HG increases, it is expected that there could be negative effects on usage of CHC.
The third group includes characteristics of the perceived illness,
insurance and income variable.
Characteristics of the perceived illness is an important factor because it is
described as perceived level of sickness which is expressed as need factor by
Anderson (Habtom and Ruys, 2003). Several studies of Deolalikar and Berhman
(1989), Gertler and van der Gaag (1990) concluded that ill people, especially the
poor, did not go to the facility until when their illnesses were already very serious
(Dercon S. and Ruttens C., 1998).

14


Two variables as the number of days of illness and number of days of
limited activity were used to measure perceived severity of illness (Trivedi, 2003).
The results of Trivedi's research are that the number of days of illness has negative
effect while the number of days of limited activity has positive effect on choosing
CHC usage in Vietnam. On the contrary, the study of Albar et al. (1996) found that
in rural areas of Niger, most serious injury cases, either life threatening or longterm disability creating, were treated by traditional medical practitioners (Habtom
and Ruys, 2003)
In conclusion, whether severity of illness has negative or positive
coefficient on CHC usage must be considered.
Health insurance variable
Similar to other determinants, health insurance also affects on health care
choices. Chang and Trivedi (2003) with the study about self-medication in
Vietnam gave strong suggestion that health insurance diverted demand away from
self-medication and toward higher quality healthcare, for example, government

hospitals. In terms of using CHC provider in Vietnam, according to Trivedi (2003)
the relationship between health insurance status and this facility is positive but
with a low level of precision. Trivedi (2003) explained that CHCs is considered as
drugs providers of insured people.
-------- -------lnilividmiis-and-buusehold-income----------- It could be uncertain that whether the effect of the respondent and

household income on the choice of health care treatment is positive or negative.
The sign of income coefficients may be positive if the particular health care
provider is considered by respondents as normal goods. Conversely, the income
coefficient would be negative if people perceived it to be inferior goods.
Trivedi (2003) concluded that Vietnamese CHC is treated by users as an
inferior good. The marginal impact of rising household income on both the
probability and level of usage is negative and significant. However, Chang and
Trivedi (2003) used data from World Bank's Living Standards measurement
survey of Vietnam, 1997-1998 in order to investigate the role of income on the use

15


of health care facilities. It is concluded that at all income levels, pharmacy is the
most frequently contacted care provider for the sick.
Change in income affects considerably on usage of non-government health
facilities in Kenya. In particularly, an increase in household income would lead to
a rise in health care demand in mission and private clinics (Mwabu et al., 1993).
Income also has a considerable impact on the choice of health care service
provider in Eritrea. Habtom and Ruys (2003) pointed that a one-unit change in
income had increased significantly the probability of selection modern health care
service providers and decreased significantly the probability of self-treatment.
These researchers divided household income into five groups. The results of their
study show that the probability of selection of private health care provider increase

tremendously as the household income increases by one-unit from middle to the
fourth and from the fourth to the highest group.
However, there are other conclusions that income was a relatively
unimportant determinant of health care choices, particularly in Mozambique
(Lindelow, 2005). Research done by Akin et el. (1986) in the Philippines also
showed that, contrary to prediction of the theory of consumer behavior, economic
variable like household income had little or no impact on health care use.
Therefore, effect of income on health care choices, especially on CHC
usage is amoiguous.
--- --- -- --- -------The expected signs of variables are presented in table 2.1.

16


Table 2.1: The expected signs of determinants of health care usage
EXPECTED

DETERMINANT

SIGN

Gender

Ambiguous

Socio-demographic

Marital status

Ambiguous

factors

Age

-

Education

-

Insurance

+

The number of days of illness

-

insurance

Health
status

Characteristics of the

The number of days of limited

perceived illness

activity

Individual

and

Ambiguous

household income

Attributes of health
care provider

The
cost

non-monetary

-

Drug availability

+

Distance

-

User fee

-

Waiting time

-

In summary, this chapter presents literature review of health care

determinants. The utility function in theory of consumer behavior is considered to
be the main measurement of consumer's satisfaction. Function form of
17

B¢ GIAO D!)C VA 8.~0
T~.O
,
TRUClNG f:H KINH TE TP.HCM

r;(A~cY

THUVI:¢N