The Application of Factor Analysis to Determine the Dominant

Causes of the Low Birth Weight Babies in

East Nusa Tenggara (ENT) Province

Astri Atti, Maria A. Kleden

and Maria Lobo

Department of Mathematics, Universitas Nusa Cendana, Jl Adisucipto, Kupang, Indonesia

Keywords: Factor Analysis, Low Birth Weight, Determinant Factors.

Abstract:

A research was carried out to determine dominant factors that affect the low birth weight babies in East

Nusa Tenggara (ENT) province Indonesia. The district Timor Tengah Selatan (TTS) was selected as the

target region for the research as this area has the highest number for the case. Samples were taken

purposively targeting the mothers whose babies have low birth weights. The data were collected using

questioners with 12 questions that have Likert Scale attributes ranged from 1 (strongly disagree) to 5

(strongly agree). Factor Analysis method was employed to examine the variables involved in the research.

The result shows that from the 12 variables that were grouped in 4 factors; internal diseases, environment,

birth defects and respiratory infection, the greatest threat to the low birth weight babies in ENT is internal

diseases with the percentage of variance is almost 40%.

1 INTRODUCTION

To develop a healthy society, there is a need to

integrate science and technology in order to promote

and improve health, to prevent diseases and to

recover and rehabilitate from illnesses. Social

economic problems are factors that are believed to

relate health issues in general and more particular to

the health of mothers and children. Therefore, health

is one of the indicators of the human welfare in any

country in the globe including Indonesia and East

Nusa Tenggara Province as one of its provinces.

ENT has a massive problems with regard to the

health of mothers and children and as it has

relatively high number of them died. The lack of

nutrition, health awareness, geographical obstacles

and lowincome families are believed to be the

factors that worsened the problems of the death of

mothers and children. One of the sub-districts that

has this big problem is the Timor Tengah Selatan

(TTS).

The death of babies is one of the indicators of

the health level of a nation. This is in agreement

with the one of the MDGs that aims at reducing the

number of babies died which was 23 in 1000 births

in 2015 in Indonesia).

There are two causes for the death of babies;

direct and indirect factors. Direct factors refer to

those brought by the birth babies such as low birth

weights and infections after births, whereas indirect

factors such as social and economic factors, health

services, condition of pregnant mothers and

environment. Deaths caused by maternal problems

are relatively high (Muchemi et al., 2015).

In Indonesia, the number of babies died during

and after delivery process, the first year and the first

five years is considered relatively high. Statistic

shows that 19 babies died in 1000 births during

neonatal period, whereas at the age of 2 to 11

months, there were an average of 15 deaths per

1.000 births and 10 deaths per 1.000 children at

the age group of 1–5 years (Yayasan Kesehatan

Perempuan (YKP) - MAMPU, 2017).

In ENT there was a fluctuate number of newborn

babies died during the period of 2011 – 2014. The

data shows that the number was 1210 in 2011, 1450

in 2012, 1286 in 2013 and 1.280 in 2014. These

numbers constitute about 12.8, 51.1, 13.6 and 14 in

1000 life births respectively. Compared to other sub-

districts in the ENT province, TTS has the highest

number of mortal babies which is 17 deaths in 1000

life births (Dinas Kesehatan Prov. NTT, 2017).

To overcome such problems, root causes should

be identified. Identification of root causes has

Atti, A., A. Kleden, M. and Lobo, M.

The Application of Factor Analysis to Determine the Dominant Causes of the Low Birth Weight Babies in East Nusa Tenggara (ENT) Province.

DOI: 10.5220/0010139600002775

In Proceedings of the 1st International MIPAnet Conference on Science and Mathematics (IMC-SciMath 2019), pages 227-231

ISBN: 978-989-758-556-2

227

definitely solved part of the problems. Having been

able to identify factors that cause the death of

babies, we will be able to control and improve such

factors which in turn reduce the number of death

babies.

The study was aiming at identifying all factors

that may cause the death of babies in ENT and then

finding the main factors that pose the greatest threat

to the problem using Factor Analysis Method.

2 RESEARCH METHOD

The research about the determinant factors of the

low birth weight babies (LBWB) was conducted

through several steps including observing causes

that cause the death of babies, analysing the factor

analysis model, interpreting the factor models and

finally forming the dead babies equation models.

This research was conducted in Timor Tengah

Selatan (TTS) District in the province of ENT. The

population of the research was all mothers who have

babies (below 5 years old) whereas the samples were

those mothers whose babies died. The samples were

taken using purposive sampling technique after

observation was undertaken to identify the number

of mothers with babies below 5 years. Research

instrument in the research was questioners with

ordinal Likert scale that consists of 12 questions

containing the attribute strongly disagree (1) to

strongly agree (5).

Factor Analysis method was employed to

analyse the data in this research. The collected data

was examined. When the data was incomplete then

the particular observation data was not included in

the modelling (Kline, 1994).

2.1 Basic Concept of the Factor

Analysis

Factor analysis is a multivariate statistical method

that used to analyse the relation between

independent variables to be then grouped in a lesser

number of variables (Johnson & Wichern, 2002).

These variables are not divided into dependent and

independent variables but the analysis is set to be

conducted in the manner of independence among

them.

The research was initiated by data exploration,

modelling factors equation, examining the factors

that cause the death of babies (initial factors) in TTS

and finding the new factors that contain a number of

initial factors that formed after the Factor analysis

and finally finding the factors that cause the death of

babies. The end stage of the research is the

interpretation of the model.

In short the research stages were including data

collecting and screening, data exploration or

description and data analysis.

2.2 Variables

There were 12 variables used in the research. They

were, invasive pneumococcal disease, diarrhea,

tetanus or postnatal infections., child abuse. birth

defects, measles, tuberculosis, meningitis, sudden

infant death syndrome (SIDS), severe respiratory

infection, malaria and low birth weight.

3 RESULTS AND DISCUSSION

3.1 General Description

3.1.1 Samples Characteristics

Table 1: Percentage of Sample characteristics on Sex

Basis.

Sex Percentage (%)

Females 53.38

Males 47.62

Total 100.00

Table 1 shows that the respondents were

dominated by females. This is most likely because of

the males normally work in their farm in the

morning when the interview was undertaken. It’s

also because the interviewer served by the same sex.

3.1.2 Characteristics of Respondents on Age

Parents of babies died were 19 - 37 years old.

Table 2: The Number of Respondents on Age Basis.

Age (y.o) Percentage (%)

19-25 52.38

26-32 33.33

33-39 14.29

Total 100.00

Table 2 shows that the highest percentage of the

parents from group of 19 – 25 y.o followed by 26-32

y.o and the lowest number was from the age group

33-39 y.o. This number confirms that the younger

the age of the parents the more likely to have their

babies die. Physical and emotional maturity is one of

IMC-SciMath 2019 - The International MIPAnet Conference on Science and Mathematics (IMC-SciMath)

228

the determinant factors to handle postnatal situation

(Jordan et al., 2018).

3.1.3 The Characteristic of Respondents on

Education Level Basis

Table 3: The Number of Respondents on education level

basis.

Respondents’ Status Percentage(%)

Not married 33.33

Married 66.67

Total 100.00

Table 3 shows that the biggest proportion of

parents whose babies died was from the primary

school group. The limitation in awareness as well as

knowledge on the health of expectant mothers,

babies’ health, nutrition and many other factors for

babies before and after births can be the biggest

concern for the death of the babies. Even though this

result may not be in line with the of the study

conducted by (Rosenthal, 2014) which concluded

that mothers’ education level did not affect child’s

rearing.

3.1.4 The Characteristics of Respondents on

Marital Status Basis

Table 4: The Number of Respondents based on Marital

Status.

Res

ondents’ Status Percenta

(

)

Not Marrie

33.33

Marrie

66.67

Total 100.00

Table 4 shows that 33,33% of the respondents

were not married. This implies that 33% of women

were single parents in handling the whole process

starting from pregnancy, give births and postnatal

process. Psychological pressure and stress can be the

factors of the death of the babies (Lamming, 2013).

3.2 Factor Analysis

Factor analysis is started with the determination of

the communal values that describe how the

proportion of variance of initial variables can be

explained or determined by the existing factors. The

communal values are displayed in Table 5.

Table 5: Communal values.

Variable Initial Extraction

Lungs Infec. 1.000 0.931

Diarhea 1.000 0.682

Tetanus 1.000 0.847

Child abuse 1.000 0.700

Birth defetcts 1.000 0.634

Measels 1.000

0.461

TB 1.000 0.882

Meningitis 1.000 0.882

SIDS 1.000 0.898

Resp. Infec. 1.000 0.827

Malaria 1.000 0.574

Low BW 1.000 0.648

From Table 5 it’s clearly seen that almost all the

variables have communal values extraction greater

than 0.5 except that of measles. This shows that the

relation between variables and factors is relatively

high which implies that the used variables have

obtained good factors. Communal values are the

sum of the square of each factor. For example; the

communal value for lungs infection (X

) = (0.958)

+ (-0.064)

+ (0.070)

+ (-0.070)

= 0.93166 or 0,931

which means that about 93.1% variance of X

can be

explained from the formed factors. Similar manner

applies to the other variables. From Table 5, we can

also observe that only variable Measels (X

) has the

communal value under 0.4 whereas lungs infection

has the highest communal value followed by the

variable SIDS.

Table 6: Matrix Component.

Variables

Component

1 2 3 4

Lungs infec. (X1) 0.958 -0.064 0.070 -0.070

Diarhea (X2) 0.441 -0.231 0.388 -0.533

Tetanus (X2) 0.907 -0.110 0.077 -0.078

Child abuse (X4) 0.314 -0.300 -0.689 0.190

Birth Defects (X5) 0.775 -0.067 -0.122 0.122

Measels (X6) 0.363 -0.530 0.220 0.013

TB (X7) 0.898 0.265 -0.011 -0.070

Meningitis (X8) 0.898 0.265 -0.011 -0.070

SIDS (X9) 0.182 0.859 -0.257 -0.249

Resp. Infec. (X10) 0.264 0.091 0.578 0.644

Malaria (X11) 0.532 -0.241 -0.332 0.351

LBW (X12) 0.100 0.732 0.133 0.291

The Application of Factor Analysis to Determine the Dominant Causes of the Low Birth Weight Babies in East Nusa Tenggara (ENT)

Province

229

Table 7: Factor extractions with Main component factors.

Comp

onent

Initial Ei

envalues Extraction Sums of S

uared Loadin

s Rotation Sums of S

uared Loadin

Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance

Cumulative

1 4.77 39.78 39.78 4.77 39.78 39.78 4.48 37.37 37.4

2 1.92 16.04 55.82 1.93 16.04 55.82 1.87 15.57 52.9

3 1.22 10.22 66.05 1.23 10.22 66.05 1.43 11.95 64.9

4 1.04 8.67 74.72 1.04 8.67 74.72 1.18 9.83 74.7

5 0.93 7.78 82.50

6 0.77 6.44 88.94

7 0.54 4.54 93.48

8 0.44 3.71 97.19

0.19 1.58 98.76

10 0.11 0.92 99.69

11 0.04 0.31 100.00

12 0.00 0.00 100.00

The analysis is then followed by the factorization

which is the most important step in the process. In

this research the method employed was Principal

Component Analysis. The criteria used to determine

the number of factors formed is the eigenvalue λ

≥

1. The eigenvalues of the factors are displayed in

Table 7.

It can be seen from Table 7, that the eigenvalue

for Factor 1 is 4.77 with variance of 39.78 % and

factor 2 is 16.04 with variance 1.92. Factor 12 has

0.0 eigenvalue with variance 0.000 %. Only 4

factors have eigenvalues greater than or equal to 1

(λ

≥ 1) with communal values 74.72 %. The scree

diagram for 12 eigenvalues is displayed in Figure 1.

Figure 1: Scree Plot for.12 eigenvalues.

Table 8: Loading factors after varimax rotation.

Variable

Com

onent

1 2 3 4

Lungs infec.

(

)

0.939 -0.115 0.162 0.096

Diare

(

)

0.582 -0.363 -0.418 -0.192

Tetanus

(

)

0.890 -0.159 0.151 0.081

Child abuse

(

)

0.152 -0.092 0.783 -0.234

Birth Defects

(

)

0.694 -0.051 0.368 0.120

Measels

(

)

0.330 -0.582 0.056 0.101

TBC

(

)

0.898 0.225 0.131 0.092

Meningitis

(

)

0.898 0.225 0.131 0.092

SIDS

(

)

0.272 0.882 -0.098 -0.194

Resp. Infec.

(

X10

)

0.168 -0.070 -0.057 0.889

Malaria

(

X11

)

0.367 -0.141 0.634 0.132

LBW

(

X12

)

0.095 0.667 -0.099 0.430

Table 8 shows the leading factors after rotation.

The results show that Factor 1 consists of variables

, X

which is called as internal

IMC-SciMath 2019 - The International MIPAnet Conference on Science and Mathematics (IMC-SciMath)

230

diseases (poor areas). Factor 2 composed of

variables X

, X

dan X

named as birth defects.

Factor 3 is called an environment which consists of

dan X

. Factor 4 consists of variable X

and is

called reciptory infection factor.

The analysis then produced the following model

= 0.939X

+ 0.582X

+ 0.890X

+ 0.694X

0.898X

+ 0.898X

= -0.582X

+ 0.882X

+ 0.667X

= 0.783X

+ 0.634X

= 0.889X

The four models obtained can be used for further

analysis by computing the score of each factor on

the variables base.

4 CONCLUSIONS AND

RECOMMENDATIONS

4.1 Conclusion

Based on the multivariate analysis, it can be

concluded that there were 4 (four) factors that

caused the babies mortality. They were inner

diseases, birth defects, environment where the

babies born and breath infection. Among these 4

factors, inner disease is determined to be the

dominant factor with the variance percentage is

39.78%.

4.2 Recommendation

Things that can be recommended from this research

are:

1. Relevant government units should be more active

in assisting and giving helps to the mothers about

babies (children) rearing and the hygiene of the

surrounded environment.

2. Communities are encouraged to improve their

knowledge on expecting mothers and babies.

3. Government should encourage the society to

have house constructions that provide comfort to

babies and their mothers.

4. Nutrition intake and fresh air for babies are very

essential particularly for those who live in small

spaces with too many people in them.

ACKNOWLEDGEMENTS

We would like to thank the Universitas Nusa

Cendana for providing financial support for this

research.

REFERENCES

Dinas Kesehatan Prov. NTT. (2017). No Title.

http://dinkes.nttprov.go.id/

Johnson, R. A., & Wichern, D. W. (2002). Applied

Multivariate Statistical Analysis. Prentice Hall.

Jordan, R. G., Farley, C. L., & Grace, K. T. (2018).

Prenatal and Postnatal Care: A Woman-Centered

Approach. John Wiley & Sons.

Kline, P. (1994). An Easy Guide to Factor Analysis.

Psychology Press.

Lamming, G. E. (2013). Marshall’s Physiology of

Reproduction: Volume 3 Pregnancy and Lactation.

Springer Science & Business Media.

Muchemi, O. M., Echoka, E., & Makokha, A. (2015).

Factors Associated with Low Birth Weight among

Neonates Born at Olkalou District Hospital, Central

Region, Kenya. Pan African Medical Journal, 20(5).

https://doi.org/https://doi.org/10.11604/pamj.2015.20.

108.4831

Rosenthal, M. K. (2014). An Ecological Approach To the

Study of Child Care: Family Day Care in Israel.

Routledge.

Yayasan Kesehatan Perempuan (YKP) - MAMPU. (2017).

No Title. http://www.mampu.or.id/mitra-

kami/yayasan-kesehatan-perempuan/

The Application of Factor Analysis to Determine the Dominant Causes of the Low Birth Weight Babies in East Nusa Tenggara (ENT)

Province

231