1 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) ISSN (Print) 2313-4410, ISSN (Online) 2313-4402 © Global Society of Scientific Research and Researchers http://asrjetsjournal.org/ Application of Binary Logistic Regression Model for Assessing the Caesarean Risk Factors in Bangladesh: A Case Study of Khulna and Gopalganj District Biplab Biswasa*, Md. Murad Hossainb, Mohammad Kamal Hossainc a,b,cDepartment of Statistics, Faculty of Science, Bangabandhu Sheikh Mujibur Rahman Science & Technology University, Gopalganj-8100, Bangladesh aEmail: bbiswas@isrt.ac.bd bEmail: sadid.shimul@yahoo.com cEmail: kamalbsmrstu@gmail.com Abstract The main focus of this study is to investigate the caesarean risk factors in a particular area of Bangladesh. The caesarean delivery rate is increasing day by day in most developing countries like Bangladesh and number of caesarean births has almost doubled in the last eight years in Bangladesh largely due to maternal, socio- economic and demographic factors. Instead of many disadvantages, caesarean deliveries are most common among women but it is not clinically justified. For improving the maternal health status, it is essential to determine the risk factors of caesarean delivery. For this study some hospitals have selected from Khulna and Gopalganj district. Our population is the total number of pregnant women admitted for delivery in the hospitals and 600 respondents were taken as sample. After collecting data, information were arranged in tables and analyzed. For the analysis, chi-square test and fisher’s exact test were performed to identify the significant association between delivery type (caesarean/non-caesarean) and maternal, socio-demographic and socio- economic factor’s respectively. A stepwise binary logistic regression analysis was carried out to identify the most impact factors on caesarean delivery. We found that 14 risk factors were statistically associated with delivery type out of 21 risk factors. From this study, it is clear to us that above influential factors may affects the mother’s health status in Bangladesh as well as Khulna and Gopalganj district. Keywords: Caesarean delivery; Fisher’s exact test; Chi-square test; Stepwise binary logistic regression. ------------------------------------------------------------------------ * Corresponding author. http://asrjetsjournal.org/ American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 2 1. Introduction Caesarean delivery or caesarean birth, also known as caesarean section (C-section) occurs through a scratch in the abdominal wall and uterus rather than through the vagina [1, 2]. In spite of many risks this surgical procedure is most common among women. There has been a gradual increase in cesarean births over the past 30 years in many developed and developing countries [1-2, 3, 4]. Over the last few years, caesarean delivery rates in South-Asian country increase annually for Bangladesh (1.19%), India (1.09%), Nepal (1.15%) and Pakistan (1.08%) [6, 9].The average CS rate in Asia is 15.9%. China, Malaysia (15.7%) [8], Hong Kong and Lebanon present the highest CS rates in Asia, with estimates of 40.5%, 27.4% and 23.3%, respectively. Nepal and Cambodia’s rates are lowest (1%), followed by Yemen (1.5%) [5, 6]. The caesarean delivery rate is increasing day by day in most developing countries like Bangladesh and number of caesarean births has almost doubled in the last eight years in Bangladesh largely due to maternal, socio-economic and demographic factors [1]. The figure rose to 690,000 or 23% of total deliveries last year from 355,000 or 9% of total deliveries in 2007, according to Bangladesh Demographic and Health Survey (BDHS) 2014 and this increase has not been clinically justified [3]. The World Health Organization recommends that a country should keep births through caesarean section below 15%, as it is a major operation that carries risks and takes a lot longer for a patient recover compared to a normal delivery [3, 6]. In Bangladesh, most of women during pregnancy are deprived from proper medical facility, balanced food and socio-economic facility as a whole [1]. Instead of some advantages of caesarean delivery, most of mothers and baby are sufferings from various health problems such as infection, hemorrhage, injury to organs, adhesions, reactions to medicine, maternal mortality rate, effects of future pregnancies, remove the uterus (hysterectomy) and premature birth, fetal injury, breathing problems respectively [1, 5]. So, increasing C-section rate is one of the main concern in most of the countries in the world and now, it is an identified as emerging “global epidemic” [10, 11]. The increase in caesarean deliveries has been attributed to multiple factors ranging from maternal, socio-demographic, socio-economic and institutional factors [3]. There are some common factors which are related to caesarean delivery such as maternal age at birth [12-14], order of birth [15], baby birth weight [16], socioeconomic status [17], high levels of maternal education [14, 18, 19], previous C-section [20-22], obstetric complications [23], high income level [14, 19, 24] etc. Age at marriage is also a significant cause of caesarean birth rates in the developing countries [25]. Many study were conducted about on this topic and many factors have been identified which are related to caesarean delivery but some factor were not included in the previous study. In the South-west region especially Khulna and Gopalganj district of Bangladesh, no early studies were carried out to examine the possible risk factors associated with the evaluated rates of caesarean birth. This study presents the most recent estimate of caesarean delivery in the south-west region of Bangladesh and identified some new factors which effect on the motivation of caesarean delivery by using data from sample survey. 2. Methodology 2.1. Conceptual Relationship We have to understand from the literature review, caesarean deliveries or caesarean birth rates (CBR) are influenced by various factors such as maternal, socio-demographic socio-economic and other relevant factors http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4454039/#ref12 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4454039/#ref14 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4454039/#ref15 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4454039/#ref16 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4454039/#ref14 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4454039/#ref14 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 3 [3]. Maternal factors are directly effect on caesarean delivery, socio-demographic and socio-economic factors are indirectly related to caesarean delivery but most of the factors are interrelated to each other. 2.2. Study Area and Study Population Both public and private hospitals have facility for caesarean and non-caesarean delivery in Bangladesh. For this reason study were conducted in the Khulna and Gopalganj district of Bangladesh for the public and private hospitals to represent the real scenario of caesarean section and to identify the risk factors which influence the caesarean delivery. Our study population is the pregnant women who were admitted for delivery in the selected public and private hospitals. 2.3. Data Collection In this study, we consider the proportion 𝑃𝑃 = 0.5 with an acceptable precision 𝐸𝐸 = 0.040, significance level of 𝛼𝛼 = 0.05 and 𝑍𝑍 = 1.96 . When the population size is unknown for Simple random sampling, the sample determinant formula is 𝑛𝑛 = 𝑃𝑃(1 − 𝑃𝑃) �𝑍𝑍 𝐸𝐸 � 2 [26]. The study sample contained of 600 women who had delivery either through normal or caesarean delivery from three private and three public hospitals. Questions were close- ended and questionnaire was filled up by direct interview methods. Most important criteria was that the pregnant women who were admitted to maternity ward of private or public hospitals for their deliveries. The study followed a cross-sectional design and participants were selected by simple random sampling during the period of February to April 2016. Since, 60% of total deliveries occurred during the period of February to April 2016. It can be ensured that data were sufficient to represent the population. All the Interviews were conducted in 1-3 days post deliveries. Among the 600 delivery cases, 435 were caesarean and remaining 165 were non-caesarean delivery. 2.4. Data Description Dependent variable: Delivery type was the dependent variable and it was taken to be dichotomous in nature (coded by the values 1 if the respondents undergo caesarean deliveries and 0 non-caesarean deliveries). Independent variables: The maternal variables were baby birth weight (<2.5kg or more than equal 2.5 kg), previous C-section (Yes=1/No=0), death of previous children (Yes=1/No=0), previous delivery type (caesarean/normal), pregnancy induced swollen of leg (Yes=1/NO=0) (it is the condition of pregnant women that causes fluid retention and the inability of muscles to absorb liquid), pregnancy brought senseless (Yes/No), maternal high blood pressure (Yes/No), maternal asthma (Yes/No), malnutrition (Yes/No), lack of vitamin, calcium and iron (Yes/No) and breast feeding (Yes/No). The socio-demographic variables included maternal age at birth, age at marriage, occupation, education level, family planning (Yes/No), pregnant before age 18 (Yes=1/No=0), order of birth (one/two/three/four), husband age at birth, ultrasonography (Yes/No), maternal body mass index, balanced diet during pregnancy (Yes/No). The maternal age at birth was categorized into four groups (Years): <20, 20-24, 25-29 and more than 30. Age at American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 4 marriage was classified into three categories: less than 18 years, 18-22 years, 23 years and above. Most of the maternal were housewife and some were businessman or related to other occupations. Maternal body mass index which are related to maternal height and weight and it was classified into four categories: Under weights, Normal, Over weights and Obese according to less than 18.5, 18.5-25, 25-30, 31 and above. Education status is the highest level of schooling attained, measured as primary and below, secondary and higher. The socio- economic variables were husband occupation, family income and some other variables were included in the study. Husbands occupation were service holder, businessman, farmer and their income can be categorized as below 15000, 15000-25000, 25000-35000, 35000 and above. 2.5. Statistical Equipment After collecting data, information was organized in table and analyzed using some statistical tools. In the beginning some percentage distribution and graphical representation were presented to realize the scenario of caesarean delivery and its related factors. Bivariate analysis was implemented to detect the significant association between types of deliveries (caesarean/non-caesarean) and a series of independent variables. Categorical or nominal variables were analyzed by the Chi-square test or Fishers exact test according their suitability. Finally, binary logistic regression analysis was conducted to separate factors which are connected to caesarean delivery. Here, tolerable 5% level of significance was considered for each of the test. 2.5.1 Pearson’s Chi-square Test Chi-square test is a statistical test commonly used for testing independence and goodness of fit [27]. The independent test is applied when you have two categorical variables from a single population. Independence test is used to determine whether there is a significant association between the two variables and Goodness of fit test is used to determine whether observed frequency distribution matches the theoretical frequency distribution. In both the cases chi-square test statistic is, 𝜒𝜒2 = ∑ ∑ �𝑂𝑂𝑖𝑖,𝑗𝑗−𝐸𝐸𝑖𝑖,𝑗𝑗� 2 𝐸𝐸𝑖𝑖,𝑗𝑗 𝑐𝑐 𝑗𝑗=1 𝑟𝑟 𝑖𝑖=1 ~𝜒𝜒(𝑟𝑟−1)(𝑐𝑐−1),𝛼𝛼 2 (1) 𝜒𝜒2 =Pearson's cumulative test statistic, which asymptotically approaches a 𝜒𝜒2 distribution 𝑂𝑂𝑖𝑖,𝑗𝑗 = Observed frequency of type (𝑖𝑖, 𝑗𝑗) , 𝐸𝐸𝑖𝑖,𝑗𝑗 = Expected frequency of type (𝑖𝑖, 𝑗𝑗) , 𝑟𝑟 = Number of row in contingency table, 𝑐𝑐 = Number of column in contingency table, 𝛼𝛼 = Level of significance. To remove the overestimation of statistical significance Yate’s correction is needed. When at least one cell of a contingency table contain frequency less than 5 then Yate’s correction is appropriate to use. Yate’s corrected version of Pearson’s chi-square statistic: 𝜒𝜒𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌2 = ∑ ∑ �|𝑂𝑂𝑖𝑖,𝑗𝑗−𝐸𝐸𝑖𝑖,𝑗𝑗|−0.5� 2 𝐸𝐸𝑖𝑖,𝑗𝑗 𝑐𝑐 𝑗𝑗=1 𝑟𝑟 𝑖𝑖=1 ~𝜒𝜒(𝑟𝑟−1)(𝑐𝑐−1),𝛼𝛼 2 (2) 2.5.2 Fisher’s Exact Test http://stattrek.com/Help/Glossary.aspx?Target=Categorical%20variable American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 5 Fisher's exact test is a statistical significance test used in the analysis of contingency tables and it is employed when sample sizes are small, also valid for sample sizes. Fisher's exact test is a statistical test used to determine if there are nonrandom associations between two categorical variables [27]. Consider two variables 𝑋𝑋 and 𝑌𝑌 with 𝑚𝑚 and 𝑛𝑛 observed states, respectively. For 𝑚𝑚 × 𝑛𝑛 contingency table, the generalize form of hyper- geometric probability function is 𝑃𝑃𝑐𝑐𝑐𝑐𝑌𝑌𝑐𝑐𝑐𝑐𝑐𝑐 = (𝑅𝑅1!𝑅𝑅2!…𝑅𝑅𝑚𝑚!)(𝐶𝐶1!𝐶𝐶2!…𝐶𝐶𝑚𝑚!) 𝑁𝑁!∐ 𝑌𝑌𝑖𝑖𝑗𝑗!𝑖𝑖,𝑗𝑗 (3) Where, 𝑎𝑎𝑖𝑖𝑗𝑗 is the number of observations in the (𝑖𝑖, 𝑗𝑗)th cell, 𝑅𝑅𝑖𝑖 and 𝐶𝐶𝑗𝑗 are the row and column total respectively, N is the total number of observations. Now find all possible contingency table consistent with the row and column total. For each one, calculate the probability using the hyper-geometric probability function and each of them consider as a p-value. Then p-value of the test can be simply computed by the sum of all p-values which are ≤ 𝑃𝑃𝑐𝑐𝑐𝑐𝑌𝑌𝑐𝑐𝑐𝑐𝑐𝑐 .There would be a significant association between two variables if the p-value less than the level of significance. 2.5.3 Binary Logistic Regression Model Logistic regression model is useful to find out the best fitting, reasonable model to establish the relationship between dependent variable or response variable (dichotomous) and set of explanatory or independent variable (dummy or categorical) [28]. It is a multivariate technique for finding the probability that an event occur. For single variable the model can be defined as 𝑃𝑃𝑟𝑟𝑃𝑃𝑃𝑃(𝑒𝑒𝑒𝑒𝑒𝑒𝑛𝑛𝑒𝑒) = 1 1+𝑌𝑌−(𝛽𝛽0+𝛽𝛽1𝑥𝑥) (4) Where 𝛽𝛽0 and 𝛽𝛽1 are regression coefficients estimated from the data and 𝑥𝑥 is the independent variable. For more than one independent variable the model is 𝑃𝑃𝑟𝑟𝑃𝑃𝑃𝑃(𝑒𝑒𝑒𝑒𝑒𝑒𝑛𝑛𝑒𝑒) = 1 1+𝑌𝑌−�𝛽𝛽0+𝛽𝛽1𝑥𝑥1+𝛽𝛽2𝑥𝑥2+⋯……..+𝛽𝛽𝑝𝑝𝑥𝑥𝑝𝑝� (5) The log-odds and odds ratio can be defined 𝑙𝑙𝑛𝑛 � 𝑝𝑝𝑟𝑟𝑐𝑐𝑝𝑝(𝑌𝑌𝑒𝑒𝑌𝑌𝑒𝑒𝑌𝑌) 𝑃𝑃𝑟𝑟𝑐𝑐𝑝𝑝(𝑒𝑒𝑐𝑐𝑒𝑒𝑌𝑌𝑒𝑒𝑌𝑌𝑒𝑒𝑌𝑌) � = 𝛽𝛽0 + 𝛽𝛽1𝑥𝑥1 + 𝛽𝛽2𝑥𝑥2 + ⋯… … . . +𝛽𝛽𝑝𝑝𝑥𝑥𝑝𝑝 (6) 𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂 𝑅𝑅𝑎𝑎𝑒𝑒𝑖𝑖𝑃𝑃 = 𝑃𝑃𝑟𝑟𝑐𝑐𝑝𝑝(𝑌𝑌𝑒𝑒𝑌𝑌𝑒𝑒𝑌𝑌) 𝑃𝑃𝑟𝑟𝑐𝑐𝑝𝑝(𝑒𝑒𝑐𝑐𝑒𝑒𝑌𝑌𝑒𝑒𝑌𝑌𝑒𝑒𝑌𝑌) = 𝑒𝑒𝛽𝛽0𝑒𝑒𝛽𝛽1𝑥𝑥1 … … . . 𝑒𝑒𝛽𝛽𝑝𝑝𝑥𝑥𝑝𝑝 (7) Odds Ratio changes when the ith independent variable increases by one unit. If 𝛽𝛽𝑖𝑖 is positive then the odds Ratio are increased which indicates that the probability of occur an event is greater than the probability of occur non- event. If 𝛽𝛽𝑖𝑖 is negative then odds Ratio are decreased which specifies that the probability of occur an event is smaller than the probability of occur non-event. The Fadden and Nagelkerke test were also used to assess the https://en.wikipedia.org/wiki/Statistical_significance https://en.wikipedia.org/wiki/Contingency_table https://en.wikipedia.org/wiki/Sample_(statistics) http://mathworld.wolfram.com/StatisticalTest.html http://mathworld.wolfram.com/CategoricalVariable.html American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 6 overall goodness of fit of the logistic regression model. Finally, a best regression model was estimated by using Stepwise variable selection techniques. The data were analyzed using the Statistical R Packages. 3. Result and Discussion We conducted our study in the south-west region of Bangladesh and our sample data were contained of 600 mothers with the combined caesarean and non-caesarean rates among the contestants were 72.5% and 27.5% respectively. Simple frequency table and graphical representation are very widely used to identify the nature and various pattern of maternal, socio-demographic and other characteristic which is useful to clarify the relation with delivery type. From Table 1, we see that maximum of mother’s are lies between the age group 20-24 years and from Table 3, caesarean delivery rate is increasing before age 29, after that decrease it. From Figure1, it is found that 39% of mother education level is SSC and 24% are primarily educated which indicate most of the mothers are less educated. Table 1: Percentage distribution of mother’s according to their age group. Age interval of mother’s (in year) Frequency % <20 150 25.0 20-24 220 36.7 25-29 155 25.8 30+ 75 12.5 Total 600 100 Figure 1: Percentage distribution of mother's educational status. Table 2: Percentage distribution of respondent husband occupation Husband Occupation Frequency % Service holder 234 39.0 Businessman 160 26.7 Farmer 90 15.0 Other 116 19.3 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 7 Figure 2: Percentage distribution of baby birth weight. From Table 2 and Table 3, it is evident that largest numbers of respondent husbands are those with service holder and 26.7% are businessman. So it may be said that caesarean delivery tendency more for higher class family. From Figure 2 and Figure 3, it is obvious that about 89.2% baby weight more than 2.5 kg and most of the family monthly income less than 15000. In recent years, recorded numbers of pregnant women go to clinic or hospital and facing with ultrasonography (Figure 4). From Table 3, maternal body mass index is one of the main factors for caesarean delivery and it also depend for previous delivery type (99% respondents in front of C-section whose were caesarean delivery in previous). Figure 3: Percentage distribution of mother’s family income. Figure 4: Percentage distribution of ultrasonography during pregnancy. 0 20 40 60 80 <15000 15000-25000 25000-35000 35000+ Monthly income patterns of mothers family YES NO 93.3% 6.7% Ultrasonography during pregnancy 0% 50% 100% >=2.5 Kg <2.5Kg 89% 11% Bar diagram of baby birth weight American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 8 Table 3 shows the percentage distribution of delivery type according to different variables with their categories and also assessing the association between delivery type and patients characteristics (maternal variables, socio- demographic variables and socio-economic variables). Here, we have considered the 𝐻𝐻0: There is no association between delivery type and selected variables VS 𝐻𝐻1: There is association between delivery type and selected variables. Table 3 also shows that out of 21 factors (variables) observed, 15 were statistically associated with the delivery type. Among all the maternal variables baby birth weight (P<0.001), previous delivery type (P<0.001), ultrasonography during pregnancy (P=0.002), Senseless During pregnancy (P=0.013), high blood pressure during pregnancy (P=0.001), Balanced Diet During Pregnancy (P<0.001), breast feeding (P<0.001) were highly significant and Medicine During Pregnancy (P=0.81), Pregnancy induced swollen of leg (P=0.78), Asthma during pregnancy (P=0.40) were not statistically associated with delivery type. Among the socio-demographic factors maternal age at birth (P=0.004), maternal body mass index (P=0.001), husband age at birth (P<0.001), pregnant before age 18 (P=0.002) were significant and some others factors were not associated. All the socio- economic variables were associated with delivery type. The odds ratios with 95% confidence intervals are obtained from the logistic regression model (for all variables) of caesarean delivery were obtained. From this analysis, we found that caesarean delivery rises significantly with having the maternal age at birth (most effective age group 25-29), baby birth weight, mother’s occupation (housewife), husbands age at birth, husband occupation, pregnant before age 18, death of previous children, Table 3: Percentage distribution of selected variables according to delivery type and test of association Selected variables Delivery Type p-value Caesarean Delivery Non-caesarean Delivery n % n % Maternal Age at Birth <20 20-24 25-29 30+ 95 165 125 50 63.3 75.0 80.7 66.7 55 55 30 25 36.7 25.0 19.3 33.3 0.004 Maternal Body Mass Index (BMI) Under Weight (<18.5) Normal Weight (18.5-24.9) Over Weight (25-29.9) Obese (30+) 55 240 125 14 61.1 70.6 80.7 93.3 35 100 30 1 38.9 29.4 19.3 6.7 0.001 Baby Birth Weight American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 9 < 2.5 kg 25 38.5 40 61.5 <0.001 Maternal Education Status Primary SSC Higher 150 190 95 68.2 80.9 65.5 70 45 50 31.8 19.1 34.5 0.001 Maternal Occupation Housewife Service Holder Businessman 425 10 1 74.6 40.0 20.0 145 15 4 25.4 60.0 80.0 <0.001 Husband Age at Birth <25 >=25 30 405 50.0 75.0 30 135 50.0 25.0 <0.001 Husband Occupation Service Holder Businessman Farmer Other 175 130 55 75 74.8 81.3 61.1 64.7 59 30 35 41 25.2 18.7 38.9 35.3 0.001 Husband Income <15000 15000-25000 25000-35000 35000+ 290 85 15 42 73.4 80.9 50.0 62.7 105 20 15 25 26.6 19.1 50.0 37.3 0.003 Early Age of Marrige <18 18+ 245 190 70.0 76.0 105 60 30.0 24.0 0.12 Order of Birth One 260 73.2 95 26.8 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 10 Two Three Four 130 30 15 74.3 60.0 75.0 45 20 5 25.7 40.0 25.0 0.23 Pregnant Before Age 18 No Yes 235 200 78.3 66.7 65 100 21.7 33.3 0.002 Ultrasonography During Pregnancy No Yes 20 415 50.0 74.1 20 145 50.0 25.9 0.002 Previous Delivery Type Caesar Normal 109 80 99.1 53.3 1 70 0.9 46.7 <0.001 Senseless During pregnancy No Yes 390 45 74.3 60.0 135 30 25.7 40.0 0.013 Pregnancy induced swollen of leg No Yes 230 205 73.1 71.9 85 80 26.9 28.1 0.784 Asthma During Pregnancy No Yes 355 80 71.7 76.2 140 25 28.3 23.8 0.400 High Blood Pressure No Yes 265 170 67.9 80.9 125 40 32.1 19.1 0.001 Death of Previous Children No Yes 375 60 70.3 92.3 160 5 29.7 7.7 0.001 Balanced Diet During Pregnancy No Yes 125 310 86.2 68.1 20 145 13.8 31.9 <0.001 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 11 Medicine During Pregnancy No Yes 360 75 71.4 72.7 30 135 28.6 27.3 0.81 Breast Feeding No Yes 195 240 92.8 61.5 15 150 7.2 38.5 <0.001 Family Planning No Yes 165 270 75.0 71.1 55 110 25.0 28.9 0.342 high blood pressure. The results are interpreted as maternal and his husband motivated to caesarean delivery when mother’s and father’s age become more than 25, for avoiding risk (if previous child were died). Early marriage and high blood pressure are most important cause of caesarean delivery. Balanced diet during pregnancy, family planning and breast feeding are effective factor of non-caesarean delivery. Table 4: Fitted logistic regression model by using stepwise variable selection techniques, odds ratio and their confidence interval. Selected Variables Odds Ratio[Exp(coefficients)] 95% Confidence Interval Maternal age at birth (<20) 20-24 25-29 30+ 1.000 12.39** 94.67** 26.71** 4.065-39.29 19.88-504.9 5.003-152.2 Baby Birth Weight (< 2.5 kg) >=2.5 kg 1.000 15.98** 7.113-38.15 Mothers Occupation (Service Holder) Housewife Businessman 1.000 6.632** 0.312* 1.841-25.35 0.012-3.924 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 12 Husband Age at Birth (<25) >=25 1.000 2.836** 1.033-8.101 Husband Occupation (Other) Farmer Businessman Service Holder 1.000 2.045** 6.446** 2.251** 0.903-4.776 2.813-15.47 1.080-4.725 Order of Birth (One) Two Three Four 1.000 0.176** 0.033** 0.066** 0.068-0.439 0.007-0.154 0.012-0.384 Pregnant Before Age 18 (No) Yes 1.000 2.879* 1.109-7.519 Senseless During pregnancy (No) Yes 1.000 0.387* 0.181-0.826 Pregnancy induced swollen of leg (No) Yes 1.000 0.583 0.328-1.029 High Blood Pressure (No) Yes 1.000 1.356* 0.327-1.163 Death of Previous Children (No) Yes 1.000 3.417* 1.232-11.34 Balanced Diet During Pregnancy American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 13 (No) Yes 1.000 0.177** 0.084-0.352 Breast Feeding (No) Yes 1.000 0.053** 0.022-0.119 Family Planning (No) Yes 1.000 0.332** 0.178-0.604 Intercept Fadden 𝑹𝑹𝟐𝟐 Nagelkerke 𝑹𝑹𝟐𝟐 0.167 0.381 0.523 *, ** indicate significant factors in the model, parenthesis indicates the reference category. To identify the most influential risks factors for caesarean delivery, we conducted a stepwise regression analysis on the variables and results are shown in the Table 4. By stepwise selection, the analysis discloses that fourteen remained significant independent variables to predict which patients were at high risk of caesarean delivery. These variables were maternal age at birth, baby birth weight, mother’s occupation, husband age at birth, husband occupation, order of birth, pregnant before age 18, senseless during pregnancy, pregnancy induced swollen of leg, high blood pressure, death of previous children, balanced died during pregnancy, breast feeding and family planning. In previous research, mother’s age at birth, husband occupation, Pregnancy induced swollen of leg, mother’s age, high blood pressure, order of birth, balanced diet during pregnancy were most influential factors of caesarean delivery. This evidence conform our study and also we have found some new most influential factor’s which were absent in the past study. 4. Conclusion and Recommendations From the above discussion, we have to conclude that caesarean delivery is one of the most serious health issues. To reduce the caesarean delivery rate in developing countries as well as south-west region of Bangladesh especially in Khulna and Gopalganj district will require widespread tactics to introduce maternal variables, proper care practices and government policies to control the health facility of public and private hospitals. In our study, fourteen significant factors can be considered to predictors of caesarean delivery which were notified in the above discussion. Finally, from the statistical point of view, this study suggests that our finding factor may influence the health of mother. The following suggestions may be recommended in view of our observation: • In this study, the rate of caesarean delivery is higher for the maternal whose age at birth less than 18 years and more than 25 years. This indication may suggest that government should take necessary steps American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 14 to prevent early marriage. • Balanced and nutritional diet may be beneficial for pregnant women. For this purpose Govt. and non Govt. organization should conduct healthcare programs. • Increasing awareness of women on suitable delivery type may influence the motive of caesarean delivery. • Provide complete and reliable information to the mothers so that they do not opt for caesarean section in a state of panic or ignorance. • Moreover, public awareness of harmful and long term effects of caesarean deliveries are needed to conform a stable mother’s health. 5. Limitations of the Study The study has been conducted in a small region of Bangladesh especially in Khulna and Gopalganj district for some public and private hospitals and the sample size is relatively small. It may be possible to get better and more reliable result to increase the sample size. We have tried to recognize a good questionnaire but there may be missing some variables or factors which are associated with delivery type. There may be some ethical questions in our questionnaire. In this case, the respondents were not spontaneously answers these questions. Acknowledgement This study is conducted by the pregnant women who were admitted in hospitals for delivery in the south-west region of Bangladesh especially in Khulna and Gopalganj district. So thanks the respondents and authority of hospitals to given with their personal information.` References [1] Van Ham MA, Van Dongen PW, Mulder J. “Maternal consequences of caesarean section. A retrospective study of intraoperative postoperative maternal complications of caesarean during a 10- year period.” European Journal of Obstetrics & Gynecology and Reproductive Biology, vol. 74 (1), pp. 1-6, Jul. 1997. [2] Declerq ER, Sakala C, Corry MP. “Listening to Mothers: Report of the first national survey U.S. survey of women childbearing experience.” New York: Maternity Centre Association, Oct 2002. [3] Gomes UA, Silva AA, Bettiol H, Barbieri MA. “Risk factors for the increasing caesarean section rate in Southeast Brazil: a comparison of two birth cohorts, 1978–1979 and 1994.” International Journal of Epidemiology, vol. 28 (4), pp. 687-694, 1999. [4] Leung GM, Lam TH, Thach TQ, Wan S, Ho LM. “Rates of caesarean birth in Hong Kong: 1987- 1999.” Birth Issues In Perinatal Care, Vol. 28, pp. 166-172, Sept. 2001. [5] Caesarean births in Bangladesh up mostly on greed, The China Post 6th July, 2015. [6] Betran AP, Merialdi M, Lauer JA, Shun WB, Thamas J, Look PV, Wagner M. “Rates of caesarean section: analysis of global, regional and national estimates.” Paediatric and Perinatal Epidemiology, vol. 21 (2), pp. 98-113, Mar. 2007. http://www.sciencedirect.com/science/journal/03012115 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 15 [7] Leone T, Padmadas SS, Matthews Z. “Community factors affecting rising caesarean section rates in developing countries: an analysis of six countries.” Social Science & Medicine, Vol. 67 (8), 1236- 1246, Oct. 2008. [8] Ravindran J. “Rising caesarean section rates in public hospitals in Malaysia 2006.” Med J Malaysia, vol. 63 (5), pp. 434–435, Dec. 2008. [9] BMMS. Bangladesh maternal mortality and health care survey 2010: summary of key findings and implications, Dhaka, Bangladesh 1-12, 2010. [10] Savage W. The caesarean section epidemic. Journal of Obstetrics & Gynecology, vol. 20, pp. 223-225, 2000. [11] World Health Organization. “Appropriate technology for birth.” Lancet, vol. 2, pp. 436-437, 1985. [12] Parrish KM, Holt VL, Easterling TR, Connell FA, LoGerfo JP (1994). Effect of changes in maternal age, parity, and birth weight distribution on primary caesarean delivery rates. JAMA, vol. 271 (6), 443- 447, Feb. 1994. [13] Ecker JL, Chen KT, Cohen AP, Riley LE, Lieberman ES. “Increased risk of caesarean delivery with advancing maternal age: indications and associated factors in nulliparous women.” Am J Obstet Gynaecol, vol. 185, pp. 883-887, 2001. [14] Kun H, Fangbiao T, Brian F, Joanna R, Rachel T, Shenglan T et al. “A mixed-method study of factors associated with differences in caesarean section rates at community level: the case of rural China.” Midwifery vol. 29(8), pp. 911-920, Aug. 2013. [15] Mossialos E, Allin S, Karras K, Davaki K. “An investigation of caesarean section in three greek hospitals: the impact of financial incentives and convenience.” Eur J Public Health, vol. 15, pp. 288- 295, Jun. 2005. [16] Onwude JL, Rao S, Selo-Ojeme DO. “Large babies and unplanned caesarean delivery.” European Journal of Obstetrics & Gynecology and Reproductive Biology, vol.118 (1), pp. 36-39, 2005. [17] Skalkidis Y, Petridou E, Papathoma E, Revinthi K, Tong D, Trichopoulos D. “Are operative delivery procedures in Greece socially conditioned?” Int J Qual Health Care, vol. 8, pp. 159-165, 1996. [18] Taffel SM. “Caesarean delivery in the United States, 1990.” Vital Health Statistics”, vol. 51, pp. 1-24, May 1994. [19] Tatar M, Gunalp S, Somunoglu S, Demirol A. “Women's perceptions of caesarean section: reflections from a Turkish teaching hospital.” Social Science & Medicine, vol. 50, pp. 1227-1233, 2000. [20] Signorelli C, Cattaruzza MS, Osborn JF. “Risk factors for caesarean section in Italy: results of a multicentre study.” Public Health, vol.109, pp. 191–199, 1995. [21] Spaans WA, Sluijs MB, Van Roosmalen J, Bleker O. “Risk factors at caesarean section and failure of subsequent trial of labour.” Journal of Obstetrics & Gynecology and Reproductive Biology, vol. 100, pp. 163–166, 2002. [22] Lynch CM, Kearney R, Turner MJ. “Maternal morbidity after elective repeat caesarean section after two or more previous procedures.” Journal of Obstetrics & Gynecology and Reproductive Biology, vol. 4320, pp.1-4, 2002. [23] Joshua LW, Fergal DM, Danielle E, Robert HB, David AN, Christine HC et al. “Obesity, obstetric complications and caesarean delivery rate–a population-based screening study.” Am J Obstet http://www.sciencedirect.com/science/journal/03012115 http://www.sciencedirect.com/science/journal/03012115 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2016) Volume 26, No 4, pp 1-16 16 Gynaecol, vol. 190(4), pp. 1091-1097, 2004. [24] Shearer EL. “Caesarean section: medical benefits and costs.” Social Science & Medicine, vol. 37, pp. 1223–1231, 1993. [25] Rahman M, Shariff AA, Saaid R, Shafie A. “Age at marriage, maternal age and caesarean delivery of first birth in the northern region of Bangladesh: a study on curve estimation.” Man in India vol. 92(1), pp. 93-113, 2012. [26] Cochran WG. Sampling techniques. 3rd ed. New York: John Wiley & Sons; 1977, pp. 72-76. [27] Gibbons J D, Chakraborti S. Nonparametric Statistical Inference, New York: Chapman and Hall/CRC; 2010, ISBN-10: 1420077619. [28] Hosmer DW, Lemeshow S. Applied Logistic Regression, New York: John Wiley & Sons; 2005, Print ISBN: 9780471356325.