Leveraging Machine Learning models to predict contraceptive nonuse among women of reproductive age in Rwanda: A machine learning approach

Leveraging Machine Learning models to predict contraceptive nonuse among women of reproductive age in Rwanda: A machine learning approach | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Leveraging Machine Learning models to predict contraceptive nonuse among women of reproductive age in Rwanda: A machine learning approach Norbert Nawe, Dieudonné N. Muhoza This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5300030/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background : Despite efforts and commitments put in place by the Rwandan Government and collaborating great health organizations, contraceptive prevalence rate (CPR) remains low in Rwanda (64% in 2019 for married women using any method and 58.% for only modern methods. CPR has however been increasing from 45% in 2010 and 48% in 2015. Consequently, unmet need for family planning dropped from 34% to 14% between 2010 and 2020. This study aims to leverage Machine Learning to predict contraceptive nonuse among women of reproductive age in Rwanda. Methods: A cross-sectional analysis of secondary data was conducted on 2020 Rwanda Demographic and Health Survey. We used six Machine Learning algorithms on the sample of 14,634 women of reproductive age, which were trained and evaluated using various metrics to know the best model. Moreover, multivariable binary logistic regression was used to determine key factors of contraceptive nonuse through Python software and to identify women at higher risk of not using contraceptives, providing valuable insights for targeted interventions and policy enhancements to improve access to reproductive health and family planning services for underserved populations in Rwanda. Results: Findings revealed that woman age, residence region, education, wealth status, marital status, urban-rural residence, total children ever born, working status, partner's occupation, and the desire for more children are key determinants of not using contraceptives. Younger women, particularly those aged 15-24, urban residents, wealthier women, and those desiring more children are at a higher risk of not using contraceptives. Furthermore, Support Vector Machine model performed better than other five classifiers in predicting nonuse of contraceptives status with an accuracy of 75%, giving a ROC-AUC score of 83%, and making it the best model to predict contraceptive behavior in Rwanda. Conclusion The results from this study suggest the use of Machine Learning to predict contraceptive outcomes accurately. Additionally, by tailoring focused interventions for identified women at higher risk of not using contraceptive could contribute to contraceptive use uptake in Rwanda. Contraceptives nonuse Women Rwanda Machine Learning Support Vector Machine Figures Figure 1 Figure 2 Introduction Contraception refers to the deliberate prevention of pregnancy through various methods, including devices, chemicals, sexual practices, drugs, or surgical procedures[ 1 ]. Common contraceptive methods include copper intrauterine devices (IUDs), vasectomy, injectable, combined oral contraceptives, progestin-only pills, and condoms, among others. In Rwanda, as in many developing countries, ensuring universal access to contraceptives and related services for women of childbearing age remains a significant challenge [ 2 ]. Despite efforts made by the government of Rwanda and various organizations to improve access to reproductive healthcare, disparities and reluctance to contraceptive use persist, affecting the health and well-being of women. Despite a continuing decrease of the proportion of women not using contraception, they still represent 42% among married women [ 3 ]. A study conducted Sebuhoro et.al.,(2016) [ 4 ], with aim to model the determinants of contraceptive choice in Rwanda using multinomial logistic regression on RDHS 2010 data came up with the following findings. Women aged 15–25, those with secondary or higher education, and those with more than three children were more likely to choose modern contraceptive methods (sterilization, barrier methods, and implants) or traditional methods over injectable. However, rural women, those involved in agricultural activities, and unemployed women were less likely to choose sterilization or implants over injectable. Among couples, both woman and husband approval played a critical role in the choice of modern methods over injectable. Another study, conducted in the United States, analyzed the characteristics of contraceptive nonusers among women aged 15–44 at risk for unintended pregnancies. Using data from the National Survey of Family Growth (2011–2017), [ 5 ] calculated unadjusted and adjusted prevalence ratios, considering p-values < 0.05 as statistically significant. The findings indicated that contraceptive nonusers were more likely be poor, adolescents, racial minorities, never married, non-English speakers, public insurance users or uninsured, non-U.S. natives, and those with zero or one child. Machine learning (ML), a subset of artificial intelligence, enables systems to learn from data and recognize patterns with minimal human intervention. In healthcare, ML algorithms can uncover novel patterns that might be difficult or impossible to detect manually, helping to develop predictive models [ 6 ]. According to the World Health Organization (WHO), the global contraceptive prevalence rate (CPR) among women of childbearing age was 64% in 2019. In 2022, the prevalence of any contraceptive method reached 65%, while modern contraceptive methods were used by 58.7% of women in unions globally. Furthermore, Sustainable Development Goal (SDG) subsection 3.7.1 indicates that from 2015 to 2022, approximately 77% of the demand for contraception was met by modern methods [ 7 ]. In Rwanda, UN reports that the contraceptive prevalence among sexually active women rose from 36.4% in 2008 to 64.1% in 2019. The report also highlights a decrease in unmet family planning needs from 34.4–13.6%, alongside a reduction in maternal mortality from 210 per 100,000 live births in 2015 to 203 per 100,000 live births in 2020 [ 8 ]. In Rwanda, the results of the analysis of the 5th General Population and Housing Census revealed that the population was 13,246,394 by August 2022, presenting an inter-censual annual growth rate of 2.3% from 2012 to 2022 with the fertility rate of 3.7 as of 2022 and a growing population density [ 8 , 9 ]. Particularly, low contraceptive prevalence and high fertility rate contributes to high population rate, which present a serious barrier to attain Rwanda’s sustainable development. Various studies have been conducted to predict contraceptive use among different age groups but most of these studies have used traditional methods while a few of them employed machine-learning algorithms to predict contraceptive use in different age groups [ 11 ]. The enhanced predictions and data analytics offered by ML techniques have become a very crucial tool and one of very interesting emerging technologies useful in different industries and over the whole test period from 2002 to 2016, ML machine models outperformed the traditional model [ 12 ]. Moreover, while traditional methods have struggled to capture these complexities, leveraging Machine Learning models offers the potential to address this gap by predicting contraceptive nonuse more accurately and they can help inform better-targeted interventions, improving contraceptive uptake and supporting the country’s broader development objectives by uncovering hidden patterns and insights from data. Failure to understand the determinants of contraceptive nonuse hinders efforts to increase contraceptive uptake in Rwanda, which is a severe concern for public health especially maternal and child health, directly undermining national economic development and ability to achieve sustainable growth. However, no previous studies have applied Machine Learning models to predict contraceptive nonuse and this highlights a significant gap in the current research. This project leverages six ML algorithms namely Logistic Regression (LoR), Decision Tree (DT), Naïve Bayes (NB), Random Forest(RF), Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) to develop a model that best predicts the likelihood of contraceptive nonuse among sexually active women in Rwanda. The model incorporates demographic, socioeconomic, and cultural factors. By analyzing available data on these factors, this research aims to identify women at higher risk of contraceptive nonuse, providing valuable insights for targeted interventions and policy enhancements to improve access to reproductive health and family planning services for underserved populations in Rwanda. Methods This study adopts a cross-sectional design to analyze secondary data from the Rwanda Demographic and Health Survey (RDHS) 2019/2020 dataset and develop a predictive model to identify key determinants of not using contraceptives among sexually active women in Rwanda. Data source and study population This study uses data from the 2019/2020 Rwanda Demographic and Health Survey (RDHS) and the target population comprises sexually active women aged 15 to 49 years. Determination of sample size The RDHS 2019/2020 employed a two-stage sampling design to ensure national-level estimates of key indicators. In the first stage, clusters made up of EAs were selected, with 500 sample points (clusters) chosen: 112 from urban areas and 388 from rural areas. In the second stage, households were systematically sampled from these clusters. 13,005 households were selected, with 12,951 of these occupied. Out of the occupied households, 12,949 participated in the interview, yielding a response rate of 100%.Among 14,675 women aged 15 to 49 years identified for individual interviews, 14,634 successfully completed the interviews, resulting in a response rate of 99.7% [13]. Consequently, the final sample size for this study is N = 14,634. Study variables Dependent variable The dependent variable in this study is contraceptive nonuse, a binary outcome variable used to predict nonuse of contraception among women aged 15–49 in Rwanda. The variable is classified as "Yes" (coded as 0) for women who did not use any contraceptive method at the time of the interview and "No" (coded as 1) for women who did use a method. Using the existing contraceptive use variable (V312) from the dataset, we derived a new variable termed contraceptive nonuse to serve as the primary outcome for the predictive modeling. Independent variables The independent variables of this study were adopted from various previous studies [2, 9-18]. These variables include current age of the respondent, marital status, wealth index, woman education level, partner’s education level, heard about Family planning on radio in last few months, heard about Family planning on TV in last few months and read about Family Planning in newspaper/magazine in last few months. Additionally, woman’s employment status, residence type, residence region, religion, woman’s ideal number of children, number of children ever born, partner’s occupation, visited health facility in last 12 months and at health facility told of Family Planning are used as independent variables. Data preparation and preprocessing Data preprocessing involves different data wrangling tasks such as data cleaning, data transformation in standard formats, data normalization, feature extraction and selection, among others. This phase significantly influences the generalization performance of supervised machine learning algorithms [24]. Data cleaning : Data cleaning is an important data analysis and machine learning preprocessing step, which aims to assure the quality of the dataset in relation to detecting errors and inconsistencies in the prepared data through corrections or rejection of the mistakes. Before data analysis, we navigated the dataset to explore missing data, duplicates and inconsistencies and we handled them by imputation with mean and with mode for numerical variables and categorical variables respectively. All these tasks aim at proving the data fit for analysis free of errors [25]. Data transformation : Data alteration is the procedure of altering fresh data into a logical format on which we can make useful conclusions. Through this process, we normalized the age variable alter data to fit within a suitable scale, we categorized and encoded categorical variables using map function. Feature engineering : Feature engineering is another critical task that we performed in data preprocessing stage by creating the new variable named “Contraceptive nonuse”, and some categories and mapping categorical variables to convert them into numerical format such as Residence type (‘Rural’=1, ‘Urban’=2), etc. to ensure ML algorithms can process them. Handling data imbalance : Naturally, real-world data often suffers from imbalance issue. Data imbalance is concerned with the situation where the number of occurrences of one class (Majority class) greatly exceeds the count of occurrences of the other class (Minority class). In this step, we identified by simple frequency counts and bar chats. The imbalance issue can result the minority class underrepresentation in the learning process of a ML model, which finally leads to favoring the majority class, exhibiting the skewed predictions and bias, which hence result the model poor generalizations [26]. To handle imbalance issue, we used Synthetic Minority Oversampling Technique (SMOTE), which generate synthetic samples for the minority class rather than making duplicates of existing samples. Methods of data analysis Multivariable binary logistic regression In this study, the multivariable binary logistic regression models the association between multiple predictors with a binary outcome, contraceptive nonuse for this case. It will predict the odds ratios of contraceptive nonuse based on independent variables. Odds ratios from the model describe the strength and direction of the association. Building Machine Learning predictive models This study used six most popular machine learning algorithms in predictive model building, which are LoR, DT, RF, SVM, Naïve Bayes and KNN [27]. After preprocessing the data, we split data into two sets: Training set (80%) and testing set (20%), then we built, trained and tested the models basing on the above-mentioned ML algorithms. To implement all these models and execute all related tasks, Python was a tool for writing necessary codes. It was chosen due to its comprehensive and extensive libraries such as Scikit-learn for machine learning tasks, Pandas for data manipulation and Matplotlib for data visualization [28]. Fig. 1 illustrates the main steps of the model development. Evaluation of the Model Model evaluation or test is the very crucial process through which a trained model is assessed to know how effective the model classifier is by measuring its prediction performance. We extracted from the confusion matrix various model evaluation metrics namely Precision, Accuracy, Sensitivity (Recall), Specificity, F1-score and Misclassification rate. We also employed Receiver Operating Characteristic-Area under the curve (ROC-AUC) for discriminative performance analysis and Cohen Kappa statistic to measure the degree of agreement for each ML algorithm. With these metrics, we easily identify which classifier outperformed on the others. Table 1 illustrate the confusion matrix. Table 1. Confusion matrix Predicted Class Positive Negative Actual class Positive True Positive (TP) False Negative (FN) Sensitivity= TP/(TP+FN) Negative False Positive (FP) True Negative (TN) Specificity=TN/(TN+FP) Precision=TP/(TP+FP) Negative predictive value= TN/(TN+FN) Accuracy=TP+TN/(TP+TN+FP+FN) Results Sociodemographic characteristics of women of reproductive age This study involves 14,634 women of reproductive age, of whom 9,043 (61.8%) do not use any form of contraception against 5,591 (38.2%) who are currently using some contraceptive methods. The descriptive results from Table 2 show that among young women aged 15-24 years, 3.9% use contraceptive methods against 96.1% who do not use any method. The women belonging to young adults’ category (25-34 years) cover 54% of nonusers to 46% of users. Women in rural area have the higher nonuse rate (66.3%) compared to use (33.7%) and for those who live in urban area, the rate of nonuse is higher (60.4%), compared to use (39.6%). The study further revealed that the proportion of nonusers is a bit higher among women with primary education (56.1%) than among those with no education (55.3%). However, women with secondary and higher education have a significantly higher proportion of contraceptive nonusers (73.8%). As far as wealth index is concerned, women living in rich families display a significantly higher percentage of (65.9%) of women not using contraception compared to those living in the middle wealth class (59 %) or in the poor category (58.4%). The corresponding percentages for contraceptive users were at 34.1%, 41% and at 41.6% for rich women, those in middle class and poor women respectively. For marital status, the rate of contraceptive nonuse among women who have never been in union is higher at 90.9% compared to married women (35.7%) and separated ones (72.4%). Concerning religion, women belonging to religions other than Christianity and Islam do not use contraception (62.0 %) more than Christians (61.8%) and Muslims (59.6%). Media exposure follows an interesting trend. The nonuse rate among women who had heard about FP on the radio the last 12 months was found to be lower (58.9 %) than women who did not (64.6 %). Moreover, contraceptive nonuse is lower (55.3%) among women who visited a health facility in the last 12 months than among those who did not (72%). Women who visited the health facility and were told of family planning show a proportion of nonuse rate of 44.7%, which is lower than among those who were not told of FP (65.8%). The Table 2 illustrates the full distribution of respondents according to contraceptive use status (Use/Nonuse) by independent variables. Table2. Distribution of respondents according to not using contraception and use by independent variable Predictors Categories Distribution (Frequency and %) Contraceptive nonuse (Yes, No) No (use) Yes (nonuse) Counts % counts % Wealth Index Poor 2310 41.61% 3241 58.39% Middle 1110 40.97% 1599 59.03% Rich 2171 34..06% 4203 65.94% Education level No education 604 44.67% 748 55.33% Primary 3733 43.92% 4767 56.08% Secondary + 1254 26.22% 3528 73.78% Partner’s education level No education 616 61.23% 390 38.77% Primary 4098 33.61% 8094 66.39% Secondary + 877 61.07% 559 38.93% Employment status Employed 4374 45.15% 5313 54.85% Unemployed 1217 24.60% 3730 75.40% Heard about FP on radio in last few months No 2633 35.43% 4798 64.57% Yes 2958 41.07% 4245 58.93% Heard about FP on TV in last few months No 5004 38.33% 8050 61.67% Yes 587 37.15% 993 62.85% Read about FP newspaper in last few months No 5202 38.89% 8173 61.11% Yes 389 30.90% 870 69.10% Visited_health_facility_in_last_12_months No 1597 28.02% 4102 71.98% Yes 3994 44.70% 4941 55.30% At health center, told of FP No 4043 34.17% 7790 65.83% Yes 1548 55.27% 1253 44.73% Partner’s occupation No profession 327 56.30% 254 43.7% Having profession 5264 37.50% 8789 62.50% Age Adolescent 130 3.93% 3178 96.07% Young adult 3018 45.96% 3548 54.04% Adults 2443 51.32% 2317 48.68% Total children ever born 0-2 2338 25.18% 6947 74.82% 3-5 2485 63.98% 1399 36.02% 6 + 768 52.42% 697 47.58% Marital status Never in union 549 9.06% 5511 90.94% Married 4688 64.31% 2602 35.69% Separated 354 27.57% 930 72.43% Type of residence Rural 1197 33.71% 2354 66.23% Urban 4394 39.65% 6689 60.35% Residence region Kigali 622 32.38% 1299 67.62% South 1356 38.94% 2126 61.06% West 1220 36.84% 2092 63.16% North 976 42.55% 1318 57.45% Est 1417 39.09% 2208 60.91% Religion Christian 5382 38.16% 8720 61.84% Muslim 116 40.42% 171 59.58% Others 93 37.96% 152 62.04% Ideal number of children 0-2 986 33.98% 1916 66.02% 3-5 4122 38.96% 6458 61.04% 6 + 483 41.93% 669 58.07% Identification of key determinants of contraceptive nonuse The table 3 reveals that marital status is the most significant determinant of contraceptive nonuse. The never in union women show the greatest likelihood of not using contraception. This category displays an odds ratio of 10.60 (CI: 9.3251-12.0513, p-value < 0.001) compared to married (ref. category). This means that women who have never been in union are more than ten times more likely to not use contraception than those in union. Separated women also have an increased risk of not using contraception (OR: 5.21, CI: 4.51 - 6.01, p < 0.001). Young women (15-24 years) taken as a reference category are at a relatively higher risk of contraceptive nonuse when compared to both adult women and aged ones. The odd ratios are 0.301 (95% CI: 0.26 - 0.35, p-value <0.01) for adult women, and 0.59 (95% CI: 0.49 - 0.70, p-value <0.01) for aged women. The OR less than 1 indicate a lower risk than the reference category. This would mean that adult and old women are less likely to not use contraceptives than young women. In addition, women who desire many children as six or above are more likely to experience contraceptive nonuse (OR: 2.10, CI: 1.77 -2.49 p-value < 0.001) than those who want fewer children (ref. category). Likely, women who want to have 3 to5 children also show more significant association with nonuse of contraception (OR: 1.39, CI of 1.24- 1.54, p-value < 0.001) than those wanting fewer, but less than those wanting more. This reflects how higher fertility preferences, determines the likelihood of not using contraception. The region of residence was also found to be one of the key predictors of contraceptive nonuse. Women residing in the West region are 1.32 times more likely not using contraception (95% CI, 1.17-1.48, p-value < 0.001) than those residing in the Eastern province (ref. category) and for Kigali, we have OR:1.33, CI: 1.13 - 1.55, p-value < 0.001). This suggests that regional factors significantly affect access to or attitudes toward contraception, hence leading to high rates of nonuse. Moreover, women in urban areas have an odd ratio of 1.25 (CI: 1.12 -1.40, p-value < 0.001). This means that they are about 25% more likely not using contraception compared to their rural counterparts. Somewhat surprisingly, rich women have an OR of 1.31 (CI: 1.16 -1.48, p-value < 0.001), when compared to middle class indicating that they are more likely to be nonusers. The study further found out that women with a primary education have an OR of 0.81 (CI: 0.71-0.92 p-value = 0.001), indicating lower odds of not using contraception compared to those with no education. Conversely, women with education higher than the primary level behavior as those without education. This result looks surprising since existing literature highlights high contraceptive nonuse among lees educated women. This is because most women with secondary education or higher are still students and so have not yet entered in childbearing, while women with no education are mostly among old women, married, sexually active, and therefore much more willing to control for their fertility. Lastly, women who desire more children (6 or more) are at increased risk of not using contraception. Religion also showed no significant impact, with Muslim women (OR = 0.77, p = 0.084) and women of other religions (OR = 1.09, p = 0.59) not differing significantly from Christians. In terms of wealth, the poorest women (OR = 1.06, p = 0.34) do not differ from those in the middle wealth group, and exposure to family planning messages via radio (OR = 0.98, p = 0.63), TV (OR = 0.97, p = 0.72), or newspapers/magazines (OR = 1.03, p = 0.75) showed no significant impact on nonuse. Health facility visits (OR = 0.94, p = 0.19) and being informed about family planning at a health center (OR = 0.97, p = 0.59) were also not significant. Additionally, the education level of a partner, whether primary (OR = 1.004, p = 0.96) or secondary and higher (OR = 1.03, p = 0.79), showed no significant impact. Therefore, these factors were not key determinants of contraceptive nonuse, as their p-values indicate no statistically significant associations in this study. On the other hand, the key predictors of contraceptive nonuse are marital status, age, region, residence type, the number of children ever born, wealth status, employment status, partner's occupation, and the desired number of children. Table 3. Multivariable binary logistic regression analysis results of nonuse/use of contraception Variable ORs 95% CI P-value Age - Young adults (25-34) 0.301 0.257 - 0.352 <0.01 ** - Adults (35-49) 0.585 0.486 - 0.704 <0.01 ** * Adolescents (15-24) ( Ref ) 1 Region - Kigali 1.327 1.135 - 1.553 <0.01 ** - North 0.988 0.866 - 1.127 0.857 - South 1.107 0.984 - 1.246 0.092 - West 1.317 1.169 - 1.483 <0.01 ** * East (Ref) 1 Residence type - Urban 1.252 1.120 - 1.399 <0.01 ** * Rural (Ref) 1 Education level - Primary 0.81 0.713 - 0.922 <0.01 ** - Secondary and Higher 0.984 0.834 - 1.161 0.848 * No education (Ref) 1 Religion - Muslim 0.769 0.571 - 1.036 0.084 - Others 1.091 0.798 - 1.492 0.585 * Christian (Ref) 1 Wealth index - Poor 1.056 0.943 - 1.183 0.342 - Rich 1.31 1.157 - 1.483 <0.01 ** * Middle (Ref) 1 Total children ever born 3-5 0.445 0.396 - 0.501 <0.01 ** - 6 and more 0.611 0.518 - 0.720 <0.01 ** * 0-2 (Ref) 1 Heard about FP on radio last few months - Yes 0.979 0.897 - 1.068 0.629 * No (Ref) 1 Heard about FP on TV last few months - Yes 0.973 0.840 - 1.127 0.717 * No (Ref) 1 Heard about FP in Newspaper/Magazine last few months - Yes 1.027 0.870 - 1.212 0.754 * No (Ref) 1 Visited_health_facility_in_last_12_months - Yes 0.937 0.851 - 1.032 0.19 * No (Ref) 1 At health center, told of FP - Yes 0.972 0.874 - 1.081 0.598 * No (Ref) 1 Marital status - Never in union 10.601 9.325 - 12.05 <0.01 ** - Separated 5.205 4.511 - 6.006 <0.01 ** * Married (Ref) 1 Ideal number of children 3-5 1.387 1.248 - 1.541 <0.01 ** - 6 and more 2.103 1.774 - 2.494 <0.01 ** * 0-2 (Ref) 1 Partner’s education level - Primary 1.004 0.872 - 1.157 0.955 - Secondary and Higher 1.026 0.853 - 1.234 0.788 * No education (Ref) 1 Partner’s occupation - Not working 1.329 1.109 - 1.593 <0.01 ** * Having a profession (Ref) 1 Employment status - Unemployed 1.379 1.252 - 1.520 <0.01 ** Employed (Ref) 1 Models performance analysis and comparison From results in Table 4, Logistic Regression model constantly does poorly with 65.4% accuracy, with a misclassification rate standing high at 34.6%. The model's precision is 64.7%, recall is 65.4%, and it has an F1 of 64.9%. It also scores a Cohen's Kappa of 0.254, showing only fair agreement to the actual outcome. This model slightly underperforms in comparison to the others when evaluated for the job of classification, having very little efficiency. The performance of the logistic regression model reflects limitations. Its accuracy of 65.4% implies a struggle in the classification of the instances effectively, which in turn means that nearly one-third of the instances would be misclassified. This can be a problem in applications where accurate predictions are of utmost importance. Precision and recall of the logistic regression, both balancing at 65%, reflects that it has a moderate ability to identify positive cases, although leaving a high degree of probability of false positives and missing true positives. Gaussian Naive Bayes performs very well in the task of classification. It resulted in 71.2% accuracy, with a very small misclassification rate of 28.8% compared to DT and logistic regression. Its precision is 71.8%, with recall of 71.2% and F1 score of 71.4%. The value of 0.405 for Cohen's Kappa implies moderate agreement to actual results; this position is better than Logistic Regression but not the best. Random Forest improves accuracy even further to 73.3%, with a 26.7% misclassification rate, 73.4% precision, 73.3% recall, and an F1 score of 73.4%, along with an impressive Cohen's Kappa of 0.441. Decision Tree slightly did less on the metrics, with an accuracy of 70.1% and a misclassification rate of 29.9%. Naïve Bayes with precision is at 70.4%, recall at 70.1%, and an F1 is at 70.2%, with a Cohen's Kappa of 0.377 does better than Logistic Regression but does not beat Random Forest. The SVM model makes the best classifier, with an accuracy of 75.2%, yielding a misclassification as low as 24.8%. The precision is 78.3%, the recall is 75.2%, and the F1 score is 75.5%. At the top is a Cohen's Kappa of 0.509, indicating substantial agreement. The SVM model is strong in all measures. K-Nearest Neighbors (KNN) similarly performs well, having an accuracy of 72.9% and a misclassification level of only 27.1%. The classifier presents 73.4% precision, 72.9% recall, and F1 score 73.1%, with Cohen's Kappa 0.440. However, KNN performs like Random Forest but slightly lags in values of precision and recall. Therefore, considering these metrics, the best model was SVM because it outperformed the others in terms of all the measures evaluated. The table below highlights the model performance metrics. Table 4. Model evaluation metrics for six classifiers. Model Precision Recall F1_Score Misclassification ROC_AUC Accuracy Cohen’s Kappa Logistic Regression 0.65 0.65 0.65 0.35 0.74 0.65 0.25 Naïve Bayes 0.72 0.71 0.71 0.29 0.78 0.71 0.4 Random Forest 0.73 0.73 0.73 0.27 0.80 0.73 0.43 Decision Tree 0.71 0.71 0.71 0.29 0.70 0.71 0.39 SVM 0.78 0.75 0.76 0.25 0.82 0.75 0.51 KNN 0.73 0.73 0.73 0.27 0.78 0.73 0.44 Discriminative performance analysis: ROC curve-AUC Building upon our previous discussion of performance metrics derived from confusion matrices, the ROC-AUC analysis offers a different perspective on model performance, focusing on the trade-offs between sensitivity (true positive rate) and specificity (false positive rate). The ROC curve illustrates the model's ability to discriminate between positive and negative classes across various threshold settings, while the AUC quantifies this ability as a single value representing the model’s performance. We presented this on both the AUC results table and the ROC curves visualization. Table 4 shows various classification algorithms, ranked using their respective ROC AUC scores, showing a comparative assessment of models for predicting contraceptive nonuse among sexually active women in Rwanda. The Fig.2 visualizes the ROC curves and AUC scores for Logistic Regression, Random Forest, Support Vector Machine, Decision Tree, Naïve Bayes and K-Nearest Neighbor. The results on ROC-AUC scores presented in Table 5 show that among all models, Support Vector Machine proved very effective in the prediction of the target variable, returning a score of 0.83 on the ROC AUC. The Random Forest model followed very closely with a score of 0.80, which is very strong in terms of discriminatory power, while the K-Nearest Neighbor algorithm performed admirably well with a score of 0.79, showing its robustness for this classification task. On the other hand, the Naive Bayes model returned a respectable ROC AUC score of 0.78, slightly behind KNN but still effective in these circumstances of contraceptive nonuse prediction. Logistic Regression had some accuracy with a score of 0.74 but way off from the accuracies of SVM, RF, KNN and Naive Bayes. DT had the worst model performance, coming in at an ROC AUC score of 0.68, indicating it was very poor in class distinction. Given these findings, the SVM model indicated the best choice for this classification task, offering the best tradeoff between accuracy and discriminatory power, though Random Forest and KNN might be good alternatives in light of other practical considerations. Discussion The study was carried out with the aim to build a ML model that can predict contraceptive nonuse among women reproductive age in Rwanda and to identify the key predictors of the nonuse of contraception among the previously mentioned target group. It is in that regard that we trained and tested six ML classifiers, which are Logistic Regression, Naïve Bayes, Decision Tree, Random Forest, Support Vector Machine and K-Nearest Neighbor to compare their classification performance through different evaluation metrics. Using multivariable binary logistic regression, the study revealed the key predictors of the nonuse of contraceptives among women of reproductive age in Rwanda namely younger age, living in an urban area, having no education, higher status of wealth, never married or separated, higher desired number of children, not working status and having a not working partner. These results are consistent with prior research that shows regional disparities in access and utilization of the contraceptive method [ 29 ]. Similarly, the odds of nonuse were significantly higher among women living in urban environments, likely due to the very same pattern of urbanization, cultural factors, and the preference for certain methods of using contraceptives [ 30 ]. Education level also emerged as another significant determinant, as nonuse odds were lower among the women who had primary education compared to women who were not educated. This agrees with other studies where education has been shown to enhance a woman's ability to make active choices concerning family planning [ 31 ]. However, the lack of significance for women with secondary or higher education suggests that other factors may moderate the association between higher education and contraceptive use, such as cultural norms or spousal influence [ 32 ]. Regarding the model performance, Support Vector Machine outperforms the other models in metrics that we employed in evaluation with mainly the best accuracy of all the others at 75%, precision of 78%, Recall of 75%, F1 score of 76%, Cohen’s Kappa statistic of 0.51, the highest ROC AUC score (0.83) and the misclassification rate at 25%. Although SVM stands out the best performing, the two last metrics suggest that the model need further improvement and refinement to lessen errors, trying to achieve more agreement rate with the actual outcomes and to reach the least misclassification rate in classifying instances into positives and negatives more accurately. However, these results takes us to achieving the first objective of this research, which was to build a highly performing ML model that accurately predicts contraceptive nonuse in sexually active women in Rwanda. These results confirming outperformance of SVM were supported by the findings of the study that found out this model to be the most accurate in predicting the intention to use FP among reproductive –age woman in Ethiopia(with imbalanced data), while for the balanced dataset, they had found Random forest to be the best model [ 33 ]. Note that, on this turn, they had deployed five same as the current research’s models namely LoR, SVM, RF, KNN and NB with the other three models, which are Artificial Neural Network, XGBoost, and AdaBoost. In the current study, the use of ML models, particularly Support Vector Machines demonstrates the potential of these methods to enhance the predictive accuracy in health-related research especially contraceptive behavior. Limitations This study was limited to a cross-sectional analysis using six ML algorithms only and could neither provide insights into dynamic interplay of factors influencing contraceptive behavior nor understand causal relationships and identify the trends over time. Therefore, future researches could incorporate longitudinal data and qualitative insights to deepen the understanding of these determinants and further refine predictive models as well as using other algorithms than these six we used in this research. Conclusion In the current study, we identified the key predictors of contraceptive nonuse among sexually active women of reproductive age in Rwanda. These are woman age, residence region, education, wealth status, marital status, residence type, total children ever born, working status, partner's occupation, and the desire for more children. Younger women (particularly those aged 15–24), urban residents, wealthier women, and those desiring more children are at a higher risk of not using contraceptives. Furthermore, Support Vector Machine model performed better than other five classifiers in predicting nonuse of contraceptives with an accuracy of 75%, giving an ROC-AUC score of 82.96%, making it the best model to predict contraceptive behavior in Rwanda. This model would be deployed to predict contraceptive behavior with unseen data for future researches. The findings of this research will promote more focused and customized intervention strategies, particularly targeting younger women, women in urban areas, and women of high socioeconomic status, and this may ramp up contraceptive uptake to improve health and reproductive outcomes. Abbreviations SDGs: Sustainable development goals RDHS: Rwanda Demographic Health Survey NISR: National Institute of Statistics of Rwanda RPHC: Rwanda Population and Housing Census EAs: Enumeration Areas ML: Machine Learning FP: Family Planning KNN: K-Nearest Neighbor SVM: Support Vector Machine RF: Random Forest DT: Decision Tree NB: Naïve Bayes LoR: Logistic Regression ROC: Receiver Operating Characteristic AUC: Area Under Curve SMOTE: Synthetic Minority Oversampling Technique CI : Confidence Interval ORs: Odds Ratios Declarations Ethics approval and consent to participate Ethical considerations focused on using anonymized, publicly available data from the Rwanda Demographic and Health Survey (RDHS) 2019/2020, for registered users. The secondary data use did involve directly respondents. Consent for publication: Non applicable Data availability The data were accessed at: https://dhsprogram.com/data/dataset/Rwanda_Standard-DHS_2019.cfm?flag=1 after the official request and approval by DHS program. Funding This study did not obtain any funding. Conflict of interests The Authors declares no competing interest. Authors’ contribution Norbert Nawe designed the research, analyzed the results and drafted the manuscript. Dieudonné N. Muhoza provided guidance and participated to the design of the research, reviewed the analysis and the manuscript and approved the full paper. Acknowledgement We acknowledge DHS program for providing the approval with data access authorization, which made this study possible. Author details: Norbert Nawe is a graduate from African Centre of Excellence in data Science, College of Business and Economics, University of Rwanda, Kigali, Rwanda. Dieudonné N. Muhoza is an Associate Professor in the Department of Applied Statistics, School of Economics, College of Business and Economics, University of Rwanda, Kigali, Rwanda. References Jain R, Muralidhar S. Contraceptive Methods: Needs, Options and Utilization. J Obstet Gynaecol India. Dec. 2011;61(6):626–34. 10.1007/s13224-011-0107-7 . Schwandt H, et al. Family planning providers and contraceptive users in Rwanda employ strategies to prevent discontinuation. BMC Womens Health. Oct. 2021;21(1):361. 10.1186/s12905-021-01503-1 . NISR. The 2019-20 Rwanda Demographic and Health Survey (RDHS) | National Institute of Statistics Rwanda. Accessed: Oct. 20, 2024. [Online]. Available: https://statistics.gov.rw/publication/1724 Sebuhoro D, Ndengo M, Mung’atu J, (PDF) MODELLING DETERMINANTS OF CHOICE OF CONTRACEPTIVE METHODS IN RWANDA.. Accessed: Sep. 15, 2024. [Online]. Available: https://www.researchgate.net/publication/310422798_MODELLING_DETERMINANTS_OF_CHOICE_OF_CONTRACEPTIVE_METHODS_IN_RWANDA Frederiksen BN, Ahrens K. Understanding the extent of contraceptive non-use among women at risk of unintended pregnancy, National Survey of Family Growth 2011–2017. Contracept X. Jan. 2020;2:100033. 10.1016/j.conx.2020.100033 . Flam S. Machine Learning in Healthcare: Guide to Applications & Benefits, ForeSee Medical. Accessed: Sep. 02, 2024. [Online]. Available: https://www.foreseemed.com/blog/machine-learning-in-healthcare WHO. Family planning/contraception methods. Accessed: Sep. 02, 2024. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/family-planning-contraception UN. World Contraceptive Use | Population Division. Accessed: Sep. 02, 2024. [Online]. Available: https://www.un.org/development/desa/pd/data/world-contraceptive-use NISR. Rwanda’s population reaches 13,2 million in 2022 | National Institute of Statistics Rwanda. Accessed: Sep. 02, 2024. [Online]. Available: https://statistics.gov.rw/publication/Rwanda_population_2022 WB. Rwanda - Fertility Rate, Total (births Per Woman) – 2024 Data 2025 Forecast 1960–2022 Historical. Accessed: Sep. 02, 2024. [Online]. Available: https://tradingeconomics.com/rwanda/fertility-rate-total-births-per-woman-wb-data.html Pal S. A Comparative Analysis of Machine Learning Algorithms for Predictive Analytics in Healthcare, vol. 72, pp. 10–25, Mar. 2024. Martin L-C. Machine Learning vs Traditional Forecasting Methods: An Application to South African GDP, Work. Pap. , Art. no. 12/2019, 2019, Accessed: Oct. 11, 2024. [Online]. Available: https://ideas.repec.org//p/sza/wpaper/wpapers326.html NISR. The 2019-20 Rwanda Demographic and Health Survey (RDHS) | National Institute of Statistics Rwanda. Accessed: Sep. 02, 2024. [Online]. Available: https://www.statistics.gov.rw/publication/1724 Dambo N, Jeremiah I, Wallymahmed A. Determinants of contraceptive use by women in the central senatorial zone of Bayelsa State, Nigeria: A cross-sectional survey. Niger Med J. Jan. 2017;58:26. 10.4103/0300-1652.218409 . Tesema ZT, Tesema GA, Boke MM, Akalu TY. Determinants of modern contraceptive utilization among married women in sub-Saharan Africa: multilevel analysis using recent demographic and health survey. BMC Womens Health. May 2022;22(1):181. 10.1186/s12905-022-01769-z . Jean de Dieu H, Lambert M. Predictors of contraceptive discontinuation in Rwanda: evidence from demographic and health survey 2019–2020. Contracept Reprod Med. Apr. 2024;9(1):19. 10.1186/s40834-024-00282-y . Otim J. Contraceptive nonuse among women in Uganda: a comparative assessment of predictors across regions. BMC Womens Health. Dec. 2020;20(1):275. 10.1186/s12905-020-01148-6 . Kebede SD, Sebastian Y, Yeneneh A, Chanie AF, Melaku MS, Walle AD. Prediction of contraceptive discontinuation among reproductive-age women in Ethiopia using Ethiopian Demographic and Health Survey 2016 Dataset: A Machine Learning Approach. BMC Med Inf Decis Mak. Jan. 2023;23(1):9. 10.1186/s12911-023-02102-w . Haq I et al. Machine Learning Algorithm-Based Contraceptive Practice among Ever-Married Women in Bangladesh: A Hierarchical Machine Learning Classification Approach, in Artificial Intelligence Annual Volume 2022 , IntechOpen, 2022. 10.5772/intechopen.103187 Mulugeta SS, Muluneh MW, Belay AT, Yalew MM, Agegn SB. Reason and Associated Factors for Nonuse of Contraceptives Among Ethiopian Rural Married Women: A Multilevel Mixed Effect Analysis. SAGE Open Nurs. Jan. 2023;9:23779608221150599. 10.1177/23779608221150599 . Habyarimana F, Ramroop S. The Analysis of Socio-Economic and Demographic Factors Associated with Contraceptive Use Among Married Women of Reproductive Age in Rwanda. Accessed: Sep. 02, 2024. [Online]. Available: https://www.openpublichealthjournal.com/VOLUME/11/PAGE/348/FULLTEXT/ Zeleke GT, Zemedu TG. Modern contraception utilization and associated factors among all women aged 15–49 in Ethiopia: evidence from the 2019 Ethiopian Mini Demographic and Health Survey. BMC Womens Health. Feb. 2023;23(1):51. 10.1186/s12905-023-02203-8 . Demeke H, Legese N, Nigussie S. Modern contraceptive utilization and its associated factors in East Africa: Findings from multi-country demographic and health surveys. PLoS ONE. 2024;19(1):e0297018. 10.1371/journal.pone.0297018 . Kotsiantis SB, Kanellopoulos D, Pintelas PE. Data Preprocessing for Supervised Leaning, 2007, Accessed: Sep. 02, 2024. [Online]. Available: https://publications.waset.org/14136/data-preprocessing-for-supervised-leaning Wes M. Python for Data Analysis, 2nd Edition[Book]. Accessed: Sep. 02, 2024. [Online]. Available: https://www.oreilly.com/library/view/python-for-data/9781491957653/ AWE OO. IASC - Computational Strategies for Tackling Imbalanced Data in Machine Learning | ISI. Accessed: Sep. 01, 2024. [Online]. Available: https://www.isi-web.org/webinar/iasc-computational-strategies-tackling-imbalanced-data-machine-learning Tajdini F, Kheiri M-J. (PDF) Recent advancement in Disease Diagnostic using machine learning: Systematic survey of decades, comparisons, and challenges. Accessed: Sep. 02, 2024. [Online]. Available: https://www.researchgate.net/publication/366086539_Recent_advancement_in_Disease_Diagnostic_using_machine_learning_Systematic_survey_of_decades_comparisons_and_challenges Pedregosa F et al. Jan., Scikit-learn: Machine Learning in Python, J. Mach. Learn. Res. , vol. 12, 2012. Westoff CF. Unmet need for modern contraceptive methods., ICF Macro , Oct. 2012, [Online]. Available: http://www.measuredhs.com/publications/publication-AS28-Analytical-Studies.cfm Cleland J, Harbison S, Shah IH. Unmet need for contraception: issues and challenges. Stud Fam Plann. Jun. 2014;45(2):105–22. 10.1111/j.1728-4465.2014.00380.x . Bongaarts J. The impact of family planning programs on unmet need and demand for contraception. Stud Fam Plann. Jun. 2014;45(2):247–62. 10.1111/j.1728-4465.2014.00387.x . Stephenson R, Baschieri A, Clements S, Hennink M, Madise N. Contextual Influences on Modern Contraceptive Use in Sub-Saharan Africa. Am J Public Health. 2007;97(7):1233–40. 10.2105/AJPH.2005.071522 . Adem JB, et al. Explainable machine learning algorithm to identify predictors of intention to use family planning among reproductive-age women in Ethiopia: Evidence from the performance monitoring and accountability (PMA) survey 2021 dataset. Jan. 2024;25. 10.21203/rs.3.rs-3848375/v1 . Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5300030","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":370971972,"identity":"9a4b6743-bc88-4d3c-80c0-f506cec85161","order_by":0,"name":"Norbert Nawe","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYJACZjAp/7DxAZDi4SNWiwQDQ3KzAUgLGwla0tskQDyCWsylm499Lqi4V8fPcLCt8muOnQwbA/PDRzfwaLGccyx59owzxRKSjY1tt2W3JQMdxmZsnINHi8GNHGNm3rYECYPDjG23JbcxA7XwsEkT1vIPqOUYY1ux5LZ6YrU0ALWcYWxj/LjtMDFa0pKZZxxLkJw5g7FZmnHbcR42ZoJ+ST7MXFCTwM8vwf7w489t1fb87M0PH+PTggKYecAkscpBgPEHKapHwSgYBaNgxAAAxiRBoBr7TS0AAAAASUVORK5CYII=","orcid":"","institution":"University of Rwanda","correspondingAuthor":true,"prefix":"","firstName":"Norbert","middleName":"","lastName":"Nawe","suffix":""},{"id":370971973,"identity":"538af721-e663-43ab-b21e-21f04c701ff5","order_by":1,"name":"Dieudonné N. Muhoza","email":"","orcid":"","institution":"University of Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Dieudonné","middleName":"N.","lastName":"Muhoza","suffix":""}],"badges":[],"createdAt":"2024-10-20 21:53:04","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5300030/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5300030/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":68695575,"identity":"ac4ee909-b07f-47e1-97d5-0e76d158e906","added_by":"auto","created_at":"2024-11-11 06:39:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":67647,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart of the development of six ML models\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-5300030/v1/9bd3e13885da75cd8a7f5aa0.png"},{"id":68695574,"identity":"51aaa447-e5d6-4434-b6a4-8e0d9925cf3b","added_by":"auto","created_at":"2024-11-11 06:39:03","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":68224,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVisualization of ROC-AUC Curve\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-5300030/v1/0d50df0be736f35cbe6967bf.png"},{"id":93156643,"identity":"371bac45-8b82-48ce-babe-92429aab2d0c","added_by":"auto","created_at":"2025-10-09 15:38:47","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1876571,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5300030/v1/932f2c7f-f8bd-4073-a0f6-50c41a21a715.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eLeveraging Machine Learning models to predict contraceptive nonuse among women of reproductive age in Rwanda: A machine learning approach\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eContraception refers to the deliberate prevention of pregnancy through various methods, including devices, chemicals, sexual practices, drugs, or surgical procedures[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Common contraceptive methods include copper intrauterine devices (IUDs), vasectomy, injectable, combined oral contraceptives, progestin-only pills, and condoms, among others. In Rwanda, as in many developing countries, ensuring universal access to contraceptives and related services for women of childbearing age remains a significant challenge [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Despite efforts made by the government of Rwanda and various organizations to improve access to reproductive healthcare, disparities and reluctance to contraceptive use persist, affecting the health and well-being of women. Despite a continuing decrease of the proportion of women not using contraception, they still represent 42% among married women [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eA study conducted Sebuhoro et.al.,(2016) [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], with aim to model the determinants of contraceptive choice in Rwanda using multinomial logistic regression on RDHS 2010 data came up with the following findings. Women aged 15\u0026ndash;25, those with secondary or higher education, and those with more than three children were more likely to choose modern contraceptive methods (sterilization, barrier methods, and implants) or traditional methods over injectable. However, rural women, those involved in agricultural activities, and unemployed women were less likely to choose sterilization or implants over injectable. Among couples, both woman and husband approval played a critical role in the choice of modern methods over injectable.\u003c/p\u003e \u003cp\u003eAnother study, conducted in the United States, analyzed the characteristics of contraceptive nonusers among women aged 15\u0026ndash;44 at risk for unintended pregnancies. Using data from the National Survey of Family Growth (2011\u0026ndash;2017), [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e] calculated unadjusted and adjusted prevalence ratios, considering p-values\u0026thinsp;\u0026lt;\u0026thinsp;0.05 as statistically significant. The findings indicated that contraceptive nonusers were more likely be poor, adolescents, racial minorities, never married, non-English speakers, public insurance users or uninsured, non-U.S. natives, and those with zero or one child.\u003c/p\u003e \u003cp\u003eMachine learning (ML), a subset of artificial intelligence, enables systems to learn from data and recognize patterns with minimal human intervention. In healthcare, ML algorithms can uncover novel patterns that might be difficult or impossible to detect manually, helping to develop predictive models [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. According to the World Health Organization (WHO), the global contraceptive prevalence rate (CPR) among women of childbearing age was 64% in 2019. In 2022, the prevalence of any contraceptive method reached 65%, while modern contraceptive methods were used by 58.7% of women in unions globally. Furthermore, Sustainable Development Goal (SDG) subsection 3.7.1 indicates that from 2015 to 2022, approximately 77% of the demand for contraception was met by modern methods [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. In Rwanda, UN reports that the contraceptive prevalence among sexually active women rose from 36.4% in 2008 to 64.1% in 2019. The report also highlights a decrease in unmet family planning needs from 34.4\u0026ndash;13.6%, alongside a reduction in maternal mortality from 210 per 100,000 live births in 2015 to 203 per 100,000 live births in 2020 [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. In Rwanda, the results of the analysis of the 5th General Population and Housing Census revealed that the population was 13,246,394 by August 2022, presenting an inter-censual annual growth rate of 2.3% from 2012 to 2022 with the fertility rate of 3.7 as of 2022 and a growing population density [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Particularly, low contraceptive prevalence and high fertility rate contributes to high population rate, which present a serious barrier to attain Rwanda\u0026rsquo;s sustainable development.\u003c/p\u003e \u003cp\u003eVarious studies have been conducted to predict contraceptive use among different age groups but most of these studies have used traditional methods while a few of them employed machine-learning algorithms to predict contraceptive use in different age groups [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. The enhanced predictions and data analytics offered by ML techniques have become a very crucial tool and one of very interesting emerging technologies useful in different industries and over the whole test period from 2002 to 2016, ML machine models outperformed the traditional model [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Moreover, while traditional methods have struggled to capture these complexities, leveraging Machine Learning models offers the potential to address this gap by predicting contraceptive nonuse more accurately and they can help inform better-targeted interventions, improving contraceptive uptake and supporting the country\u0026rsquo;s broader development objectives by uncovering hidden patterns and insights from data. Failure to understand the determinants of contraceptive nonuse hinders efforts to increase contraceptive uptake in Rwanda, which is a severe concern for public health especially maternal and child health, directly undermining national economic development and ability to achieve sustainable growth. However, no previous studies have applied Machine Learning models to predict contraceptive nonuse and this highlights a significant gap in the current research.\u003c/p\u003e \u003cp\u003eThis project leverages six ML algorithms namely Logistic Regression (LoR), Decision Tree (DT), Na\u0026iuml;ve Bayes (NB), Random Forest(RF), Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) to develop a model that best predicts the likelihood of contraceptive nonuse among sexually active women in Rwanda. The model incorporates demographic, socioeconomic, and cultural factors. By analyzing available data on these factors, this research aims to identify women at higher risk of contraceptive nonuse, providing valuable insights for targeted interventions and policy enhancements to improve access to reproductive health and family planning services for underserved populations in Rwanda.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eThis study adopts a cross-sectional design to analyze secondary data from the Rwanda Demographic and Health Survey (RDHS) 2019/2020 dataset and develop a predictive model to identify key determinants of not using contraceptives among sexually active women in Rwanda.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eData source and study population\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study uses data from the 2019/2020 Rwanda Demographic and Health Survey (RDHS) and the target population comprises sexually active women aged 15 to 49 years.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDetermination of sample size\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe RDHS 2019/2020 employed a two-stage sampling design to ensure national-level estimates of key indicators. In the first stage, clusters made up of EAs were selected, with 500 sample points (clusters) chosen: 112 from urban areas and 388 from rural areas. In the second stage, households were systematically sampled from these clusters. 13,005 households were selected, with 12,951 of these occupied. Out of the occupied households, 12,949 participated in the interview, yielding a response rate of 100%.Among 14,675 women aged 15 to 49 years identified for individual interviews, 14,634 successfully completed the interviews, resulting in a response rate of 99.7% [13]. Consequently, the final sample size for this study is N = 14,634.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStudy variables\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eDependent variable\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe dependent variable in this study is contraceptive nonuse, a binary outcome variable used to predict nonuse of contraception among women aged 15\u0026ndash;49 in Rwanda. The variable is classified as \u0026quot;Yes\u0026quot; (coded as 0) for women who did not use any contraceptive method at the time of the interview and \u0026quot;No\u0026quot; (coded as 1) for women who did use a method. Using the existing contraceptive use variable (V312) from the dataset, we derived a new variable termed contraceptive nonuse to serve as the primary outcome for the predictive modeling.\u003c/p\u003e\n\u003ch3\u003e\u003cem\u003eIndependent variables\u003c/em\u003e\u003c/h3\u003e\n\u003cp\u003eThe independent variables of this study were adopted from various previous studies [2, 9-18]. These variables include current age of the respondent, marital status, wealth index, woman education level, partner\u0026rsquo;s education level, heard about Family planning on radio in last few months, heard about Family planning on TV in last few months and read about Family Planning in newspaper/magazine in last few months. Additionally, woman\u0026rsquo;s employment status, residence type, residence region, religion, woman\u0026rsquo;s ideal number of children, number of children ever born, partner\u0026rsquo;s occupation, visited health facility in last 12 months and at health facility told of Family Planning are used as independent variables.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData preparation and preprocessing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData preprocessing involves different data wrangling tasks such as data cleaning, data transformation in standard formats, data normalization, feature extraction and selection, among others. This phase significantly influences the generalization performance of supervised machine learning algorithms\u0026nbsp;[24].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eData cleaning\u003c/em\u003e\u003c/strong\u003e: Data cleaning is an important data analysis and machine learning preprocessing step, which aims to assure the quality of the dataset in relation to detecting errors and inconsistencies in the prepared data through corrections or rejection of the mistakes. Before data analysis, we navigated the dataset to explore missing data, duplicates and inconsistencies and we handled them by imputation with mean and with mode for numerical variables and categorical variables respectively. All these tasks aim at proving the data fit for analysis free of errors\u0026nbsp;[25].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eData transformation\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e:\u003c/em\u003e\u003c/strong\u003e Data alteration is the procedure of altering fresh data into a logical format on which we can make useful conclusions. Through this process, we normalized the age variable alter data to fit within a suitable scale, we categorized and encoded categorical variables using map function.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eFeature engineering\u003c/em\u003e\u003c/strong\u003e: Feature engineering is another critical task that we performed in data preprocessing stage by creating the new variable named \u0026ldquo;Contraceptive nonuse\u0026rdquo;, and some categories and mapping categorical variables to convert them into numerical format\u0026nbsp;such as Residence type (\u0026lsquo;Rural\u0026rsquo;=1, \u0026lsquo;Urban\u0026rsquo;=2), etc. to ensure ML algorithms can process them.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eHandling data imbalance\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003eNaturally, real-world data often suffers from imbalance issue. Data imbalance is concerned with the situation where the number of occurrences of one class (Majority class) greatly exceeds the count of occurrences of the other class (Minority class). In this step, we identified by simple frequency counts and bar chats. The imbalance issue can result the minority class underrepresentation in the learning process of a ML model, which finally leads to favoring the majority class, exhibiting the skewed predictions and bias, which hence result the model poor generalizations\u0026nbsp;[26]. To handle imbalance issue, we used Synthetic Minority Oversampling Technique (SMOTE), which generate synthetic samples for the minority class rather than making duplicates of existing samples.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods of data analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eMultivariable binary logistic regression\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, the multivariable binary logistic regression models the association between multiple predictors with a binary outcome, contraceptive nonuse for this case. It will predict the odds ratios of contraceptive nonuse based on independent variables. Odds ratios from the model describe the strength and direction of the association.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eBuilding Machine Learning predictive models\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study used six most popular machine learning algorithms in predictive model building, which are LoR, DT, RF, SVM, Na\u0026iuml;ve Bayes and KNN\u0026nbsp;[27]. After preprocessing the data, we split data into two sets: Training set (80%) and testing set (20%), then we built, trained and tested the models basing on the above-mentioned ML algorithms. To implement all these models and execute all related tasks, Python was a tool for writing necessary codes. It was chosen due to its comprehensive and extensive libraries such as Scikit-learn for machine learning tasks, Pandas for data manipulation and Matplotlib for data visualization\u0026nbsp;[28]. Fig. 1 illustrates the main steps of the model development.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eEvaluation of the Model\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eModel evaluation or test is the very crucial process through which a trained model is assessed to know how effective the model classifier is by measuring its prediction performance. We extracted from the confusion matrix various model evaluation metrics namely Precision, Accuracy, Sensitivity (Recall), Specificity, F1-score and Misclassification rate. We also employed Receiver Operating Characteristic-Area under the curve (ROC-AUC) for discriminative performance analysis and Cohen Kappa statistic to measure the degree of agreement for each ML algorithm. With these metrics, we easily identify which classifier outperformed on the others. Table 1 illustrate the confusion matrix.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;1. Confusion matrix\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"645\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 315px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePredicted Class\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 228px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePositive\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 167px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eNegative\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 228px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"bottom\" style=\"width: 38px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eActual class\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Positive\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003e\u003cem\u003eTrue Positive (TP)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 167px;\"\u003e\n \u003cp\u003e\u003cem\u003eFalse Negative (FN)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 228px;\"\u003e\n \u003cp\u003e\u003cem\u003eSensitivity= TP/(TP+FN)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; Negative\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003e\u003cem\u003eFalse Positive (FP)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 167px;\"\u003e\n \u003cp\u003e\u003cem\u003eTrue Negative (TN)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 228px;\"\u003e\n \u003cp\u003e\u003cem\u003eSpecificity=TN/(TN+FP)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\" style=\"width: 38px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003e\u003cem\u003ePrecision=TP/(TP+FP)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 167px;\"\u003e\n \u003cp\u003e\u003cem\u003eNegative predictive value= TN/(TN+FN)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 228px;\"\u003e\n \u003cp\u003e\u003cem\u003eAccuracy=TP+TN/(TP+TN+FP+FN)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eSociodemographic characteristics of women of reproductive age\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study involves 14,634 women of reproductive age, of whom 9,043 (61.8%) do not use any form of contraception against 5,591 (38.2%) who are currently using some contraceptive methods.\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003eThe descriptive results from Table 2\u0026nbsp;show\u0026nbsp;that among young women aged 15-24 years, 3.9% use contraceptive methods against 96.1% who do not use any method. The women belonging to young adults\u0026rsquo; category (25-34 years) cover 54% of nonusers to 46% of users. Women in rural area have the higher nonuse rate (66.3%) compared to use (33.7%) and for those who live in urban area, the rate of nonuse is higher (60.4%), compared to use (39.6%).\u003c/p\u003e\n\u003cp\u003eThe study further revealed that the proportion of nonusers is a bit higher among women with primary education (56.1%) than among those with no education (55.3%). However, women with secondary and higher education have a significantly higher proportion of contraceptive nonusers (73.8%). As far as wealth index is concerned, women living in rich families display a significantly higher percentage of (65.9%) of women not using contraception compared to those living in the middle wealth class (59 %) or in the poor category (58.4%). The corresponding percentages for contraceptive users were at 34.1%, 41% and at 41.6% for rich women, those in middle class and poor women respectively. For marital status, the rate of contraceptive nonuse among women who have never been in union is higher at 90.9% compared to married women (35.7%) and separated ones (72.4%).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eConcerning religion, women belonging to religions other than Christianity and Islam do not use contraception (62.0 %) more than Christians (61.8%) and Muslims (59.6%). Media exposure follows an interesting trend. The nonuse rate among women who had heard about FP on the radio the last 12 months was found to be lower (58.9 %) than women who did not (64.6 %).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMoreover, contraceptive nonuse is lower (55.3%) among women who visited a health facility in the last 12 months than among those who did not (72%). Women who visited the health facility and were told of family planning show a proportion of nonuse rate of 44.7%, which is lower \u0026nbsp;than among those who were not told of FP (65.8%). The Table 2 illustrates the full distribution of respondents according to contraceptive use status (Use/Nonuse) by independent variables.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTable2. Distribution of respondents according to not using contraception\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;and use by independent variable\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"4\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePredictors\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"4\" valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eCategories\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 352px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eDistribution (Frequency and %)\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 352px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eContraceptive nonuse (Yes, No)\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 172px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eNo (use)\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 180px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eYes (nonuse)\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eCounts\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e%\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ecounts\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e%\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eWealth Index\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003ePoor\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e2310\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e41.61%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e3241\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e58.39%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eMiddle\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1110\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e40.97%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e1599\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e59.03%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eRich\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e2171\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e34..06%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e4203\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e65.94%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eEducation level\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo education\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e604\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e44.67%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e748\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e55.33%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003ePrimary\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e3733\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e43.92%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e4767\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e56.08%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eSecondary\u003csup\u003e+\u003c/sup\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1254\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e26.22%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e3528\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e73.78%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePartner\u0026rsquo;s education level\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo education\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e616\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e61.23%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e390\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e38.77%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003ePrimary\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e4098\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e33.61%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e8094\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e66.39%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eSecondary \u003csup\u003e+\u003c/sup\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e877\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e61.07%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e559\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e38.93%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eEmployment status\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eEmployed\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e4374\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e45.15%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e5313\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e54.85%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eUnemployed\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1217\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e24.60%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e3730\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e75.40%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eHeard about FP on radio in last few months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e2633\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e35.43%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e4798\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e64.57%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eYes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e2958\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e41.07%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e4245\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e58.93%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eHeard about FP on TV in last few months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e5004\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e38.33%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e8050\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e61.67%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eYes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e587\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e37.15%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e993\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e62.85%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eRead about FP newspaper in last few months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e5202\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e38.89%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e8173\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e61.11%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eYes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e389\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e30.90%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e870\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e69.10%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eVisited_health_facility_in_last_12_months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1597\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e28.02%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e4102\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e71.98%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eYes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e3994\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e44.70%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e4941\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e55.30%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eAt health center, told of FP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e4043\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e34.17%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e7790\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e65.83%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eYes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1548\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e55.27%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e1253\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e44.73%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePartner\u0026rsquo;s occupation\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNo profession\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e327\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e56.30%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e254\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e43.7%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eHaving profession\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e5264\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e37.50%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e8789\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e62.50%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eAge\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eAdolescent\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e130\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e3.93%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e3178\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e96.07%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eYoung adult\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e3018\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e45.96%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e3548\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e54.04%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eAdults\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e2443\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e51.32%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e2317\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e48.68%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eTotal children ever born\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003e0-2\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e2338\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e25.18%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e6947\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e74.82%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003e3-5\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e2485\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e63.98%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e1399\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e36.02%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003e6 \u003csup\u003e+\u003c/sup\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e768\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e52.42%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e697\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e47.58%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eMarital status\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNever in union\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e549\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e9.06%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e5511\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e90.94%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eMarried\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e4688\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e64.31%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e2602\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e35.69%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eSeparated\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e354\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e27.57%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e930\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e72.43%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eType of residence\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eRural\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1197\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e33.71%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e2354\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e66.23%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eUrban\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e4394\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e39.65%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e6689\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e60.35%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"5\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eResidence region\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eKigali\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e622\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e32.38%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e1299\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e67.62%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eSouth\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1356\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e38.94%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e2126\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e61.06%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eWest\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1220\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e36.84%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e2092\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e63.16%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eNorth\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e976\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e42.55%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e1318\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e57.45%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eEst\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e1417\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e39.09%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e2208\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e60.91%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eReligion\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eChristian\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e5382\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e38.16%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e8720\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e61.84%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eMuslim\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e116\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e40.42%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e171\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e59.58%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003eOthers\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e93\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e37.96%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e152\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e62.04%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 195px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eIdeal number of children\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003e0-2\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e986\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e33.98%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e1916\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e66.02%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003e3-5\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e4122\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e38.96%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e6458\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e61.04%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cem\u003e6 \u003csup\u003e+\u003c/sup\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 82px;\"\u003e\n \u003cp\u003e\u003cem\u003e483\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 90px;\"\u003e\n \u003cp\u003e\u003cem\u003e41.93%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 96px;\"\u003e\n \u003cp\u003e\u003cem\u003e669\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e\u003cem\u003e58.07%\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIdentification of key determinants of contraceptive nonuse\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe table 3 reveals that marital status is the most significant determinant of contraceptive nonuse. The never in union women show the greatest likelihood of not using contraception. This category displays an odds ratio of 10.60 (CI: 9.3251-12.0513, p-value \u0026lt; 0.001) compared to married (ref. category). This means that women who have never been in union are more than ten times more likely to not use contraception than those in union. Separated women also have an increased risk of not using contraception (OR: 5.21, CI: 4.51 - 6.01, p \u0026lt; 0.001).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Young women (15-24 years) taken as a reference category are at a relatively higher risk of contraceptive nonuse when compared to both adult women and aged ones. The odd ratios are 0.301 (95% CI: 0.26 - 0.35, p-value \u0026lt;0.01) for adult women, and 0.59 (95% CI: 0.49 - 0.70, p-value \u0026lt;0.01) for aged women. The OR less than 1 indicate a lower risk than the reference category. This would mean that adult and old women are less likely to not use contraceptives than young women.\u003c/p\u003e\n\u003cp\u003eIn addition, women who desire many children as six or above are more likely to experience contraceptive nonuse (OR: 2.10, CI: 1.77 -2.49 p-value \u0026lt; 0.001) than those who want fewer children (ref. category). Likely, women who want to have 3 to5 children also show more significant association with nonuse of contraception (OR: 1.39, CI of 1.24- 1.54, p-value \u0026lt; 0.001) than those wanting fewer, but less than those wanting more. This reflects how higher fertility preferences, determines the likelihood of not using contraception.\u003c/p\u003e\n\u003cp\u003eThe region of residence was also found to be one of the key predictors of contraceptive nonuse. Women residing \u0026nbsp;in the West region are \u0026nbsp;1.32 times more likely not using contraception (95% CI, 1.17-1.48, p-value \u0026lt; 0.001) than those residing in the Eastern province (ref. category) and for Kigali, we have OR:1.33, CI: 1.13 - 1.55, p-value \u0026lt; 0.001). This suggests that regional factors significantly affect access to or attitudes toward contraception, hence leading to high rates of nonuse. Moreover, women in urban areas have an odd ratio of 1.25 (CI: 1.12 -1.40, p-value \u0026lt; 0.001). This means that they are about 25% more likely not using contraception compared to their rural counterparts.\u003c/p\u003e\n\u003cp\u003eSomewhat surprisingly, rich women have an OR of 1.31 (CI: 1.16 -1.48, p-value \u0026lt; 0.001), when compared to middle class indicating that they are more likely to be nonusers. The study further found out that women with \u0026nbsp; a primary education have an OR of 0.81 (CI: 0.71-0.92 p-value = 0.001), indicating lower odds of not using contraception compared to those with no education. Conversely, women with education higher than the primary level behavior as those without education. \u0026nbsp;This result looks surprising since existing literature highlights high contraceptive nonuse among lees educated women. This is because most women with secondary education or higher are still students and so have not yet entered in childbearing, while women with no education are mostly among old women, married, sexually active, and therefore much more willing to control for their fertility. Lastly, women who desire more children (6 or more) are at increased risk of not using contraception.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eReligion also showed no significant impact, with Muslim women (OR = 0.77, p = 0.084) and women of other religions (OR = 1.09, p = 0.59) not differing significantly from Christians.\u003c/p\u003e\n\u003cp\u003eIn terms of wealth, the poorest women (OR = 1.06, p = 0.34) do not differ from those in the middle wealth group, and exposure to family planning messages via radio (OR = 0.98, p = 0.63), TV (OR = 0.97, p = 0.72), or newspapers/magazines (OR = 1.03, p = 0.75) showed no significant impact on nonuse. Health facility visits (OR = 0.94, p = 0.19) and being informed about family planning at a health center (OR = 0.97, p = 0.59) were also not significant. Additionally, the education level of a partner, whether primary (OR = 1.004, p = 0.96) or secondary and higher (OR = 1.03, p = 0.79), showed no significant impact.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTherefore, these factors were not key determinants of contraceptive nonuse, as their p-values indicate no statistically significant associations in this study. On the other hand, the key predictors of contraceptive nonuse are marital status, age, region, residence type, the number of children ever born, wealth status, employment status, partner\u0026apos;s occupation, and the desired number of children.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTable 3. Multivariable binary logistic regression analysis results of nonuse/use of contraception\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"529\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eVariable\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eORs\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e95% CI\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP-value\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eAge\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Young adults (25-34)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.301\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.257 - 0.352\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Adults (35-49)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.585\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.486 - 0.704\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* Adolescents (15-24) (\u003cstrong\u003eRef\u003c/strong\u003e)\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eRegion\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Kigali\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.327\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.135 - 1.553\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- North\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.988\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.866 - 1.127\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.857\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- South\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.107\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.984 - 1.246\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.092\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- West\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.317\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.169 - 1.483\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* East \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eResidence type\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Urban\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.252\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.120 - 1.399\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* Rural \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eEducation level\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Primary\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.81\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.713 - 0.922\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Secondary and Higher\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.984\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.834 - 1.161\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.848\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* No education \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eReligion\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Muslim\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.769\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.571 - 1.036\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.084\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Others\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.091\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.798 - 1.492\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.585\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* Christian \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eWealth index\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Poor\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.056\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.943 - 1.183\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.342\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Rich\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.31\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.157 - 1.483\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* Middle \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eTotal children ever born\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e3-5\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.445\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.396 - 0.501\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- 6 and more\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.611\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.518 - 0.720\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* 0-2 \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eHeard about FP on radio last few months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Yes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.979\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.897 - 1.068\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.629\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* No \u0026nbsp;\u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eHeard about FP on TV last few months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Yes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.973\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.840 - 1.127\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.717\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* No \u0026nbsp;\u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eHeard about FP in Newspaper/Magazine last few months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Yes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.027\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.870 - 1.212\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.754\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* No \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eVisited_health_facility_in_last_12_months\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Yes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.937\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.851 - 1.032\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.19\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* No \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eAt health center, told of FP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Yes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.972\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.874 - 1.081\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.598\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* No \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eMarital status\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Never in union\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e10.601\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e9.325 - 12.05\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Separated\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e5.205\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e4.511 - 6.006\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* Married \u0026nbsp;\u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eIdeal number of children\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e3-5\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.387\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.248 - 1.541\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- 6 and more\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e2.103\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.774 - 2.494\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* 0-2 \u0026nbsp; \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePartner\u0026rsquo;s education level\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Primary\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.004\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.872 - 1.157\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.955\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Secondary and Higher\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.026\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.853 - 1.234\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.788\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* No education \u0026nbsp;\u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePartner\u0026rsquo;s occupation\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Not working\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.329\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.109 - 1.593\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e* Having a profession \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;Employment status\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 54.4423%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e- Unemployed\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.379\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e1.252 - 1.520\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e\u0026lt;0.01 **\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;Employed \u0026nbsp; \u0026nbsp; \u003cstrong\u003e(Ref)\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e1\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45.5577%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.3913%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.9036%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.1474%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eModels performance analysis and comparison\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFrom results in Table 4, Logistic Regression model constantly does poorly with 65.4% accuracy, with a misclassification rate standing high at 34.6%. The model\u0026apos;s precision is 64.7%, recall is 65.4%, and it has an F1 of 64.9%. It also scores a Cohen\u0026apos;s Kappa of 0.254, showing only fair agreement to the actual outcome. This model slightly underperforms in comparison to the others when evaluated for the job of classification, having very little efficiency. The performance of the logistic regression model reflects limitations. Its accuracy of 65.4% implies a struggle in the classification of the instances effectively, which in turn means that nearly one-third of the instances would be misclassified. This can be a problem in applications where accurate predictions are of utmost importance.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePrecision and recall of the logistic regression, both balancing at 65%, reflects that it has a moderate ability to identify positive cases, although leaving a high degree of probability of false positives and missing true positives. Gaussian Naive Bayes performs very well in the task of classification. It resulted in 71.2% accuracy, with a very small misclassification rate of 28.8% compared to DT and logistic regression. Its precision is 71.8%, with recall of 71.2% and F1 score of 71.4%.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe value of 0.405 for Cohen\u0026apos;s Kappa implies moderate agreement to actual results; this position is better than Logistic Regression but not the best. Random Forest improves accuracy even further to 73.3%, with a 26.7% misclassification rate, 73.4% precision, 73.3% recall, and an F1 score of 73.4%, along with an impressive Cohen\u0026apos;s Kappa of 0.441. Decision Tree slightly did less on the metrics, with an accuracy of 70.1% and a misclassification rate of 29.9%. \u0026nbsp;Na\u0026iuml;ve Bayes with precision is at 70.4%, recall at 70.1%, and an F1 is at 70.2%, with a Cohen\u0026apos;s Kappa of 0.377 does better than Logistic Regression but does not beat Random Forest.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe SVM model makes the best classifier, with an accuracy of 75.2%, yielding a misclassification as low as 24.8%. The precision is 78.3%, the recall is 75.2%, and the F1 score is 75.5%. At the top is a Cohen\u0026apos;s Kappa of 0.509, indicating substantial agreement. The SVM model is strong in all measures. K-Nearest Neighbors (KNN) similarly performs well, having an accuracy of 72.9% and a misclassification level of only 27.1%. The classifier presents 73.4% precision, 72.9% recall, and F1 score 73.1%, with Cohen\u0026apos;s Kappa 0.440. However, KNN performs like Random Forest but slightly lags in values of precision and recall. Therefore, considering these metrics, the best model was SVM because it outperformed the others in terms of all the measures evaluated. The table below highlights the model performance metrics.\u003c/p\u003e\n\u003cp\u003eTable 4. Model evaluation metrics for six classifiers.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"655\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 19.0259%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eModel\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.9589%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ePrecision\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.45814%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eRecall\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.1979%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eF1_Score\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.1339%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eMisclassification\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11.7199%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eROC_AUC\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.0457%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eAccuracy\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14.4597%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eCohen\u0026rsquo;s Kappa\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 19.0259%;\"\u003e\n \u003cp\u003e\u003cem\u003eLogistic Regression\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.9589%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.65\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.45814%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.65\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.1979%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.65\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.1339%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.35\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11.7199%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.74\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.0457%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.65\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14.4597%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.25\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 19.0259%;\"\u003e\n \u003cp\u003e\u003cem\u003eNa\u0026iuml;ve Bayes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.9589%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.72\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.45814%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.71\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.1979%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.71\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.1339%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.29\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11.7199%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.78\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.0457%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.71\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14.4597%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.4\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 19.0259%;\"\u003e\n \u003cp\u003e\u003cem\u003eRandom Forest\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.9589%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.45814%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.1979%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.1339%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.27\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11.7199%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.80\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.0457%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14.4597%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.43\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 19.0259%;\"\u003e\n \u003cp\u003e\u003cem\u003eDecision Tree\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.9589%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.71\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.45814%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.71\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.1979%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.71\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.1339%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.29\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11.7199%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.70\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.0457%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.71\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14.4597%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.39\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 19.0259%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eSVM\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.9589%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.78\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.45814%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.75\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.1979%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.76\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.1339%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.25\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11.7199%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.82\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.0457%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.75\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14.4597%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.51\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 19.0259%;\"\u003e\n \u003cp\u003e\u003cem\u003eKNN\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.9589%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 7.45814%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.1979%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 16.1339%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.27\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11.7199%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.78\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 10.0457%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.73\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 14.4597%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.44\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eDiscriminative performance analysis: ROC curve-AUC\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBuilding upon our previous discussion of performance metrics derived from confusion matrices, the ROC-AUC analysis offers a different perspective on model performance, focusing on the trade-offs between sensitivity (true positive rate) and specificity (false positive rate). The ROC curve illustrates the model\u0026apos;s ability to discriminate between positive and negative classes across various threshold settings, while the AUC quantifies this ability as a single value representing the model\u0026rsquo;s performance. We presented this on both the AUC results table and the ROC curves visualization.\u003c/p\u003e\n\u003cp\u003eTable 4 shows various classification algorithms, ranked using their respective ROC AUC scores, showing a comparative assessment of models for predicting contraceptive nonuse among sexually active women in Rwanda.\u0026nbsp;The Fig.2 visualizes the ROC curves and AUC scores for Logistic Regression, Random Forest, Support Vector Machine, Decision Tree,\u0026nbsp;Na\u0026iuml;ve Bayes and K-Nearest Neighbor.\u003c/p\u003e\n\u003cp\u003eThe results on ROC-AUC scores presented in Table 5 show that among all models, Support Vector Machine proved very effective in the prediction of the target variable, returning a score of 0.83 on the ROC AUC. The Random Forest model followed very closely with a score of 0.80, which is very strong in terms of discriminatory power, while the K-Nearest Neighbor algorithm performed admirably well with a score of 0.79, showing its robustness for this classification task. On the other hand, the Naive Bayes model returned a respectable ROC AUC score of 0.78, slightly behind KNN but still effective in these circumstances of contraceptive nonuse prediction. Logistic Regression had some accuracy with a score of 0.74 but way off from the accuracies of SVM, RF, KNN and Naive Bayes. DT had the worst model performance, coming in at an ROC AUC score of 0.68, indicating it was very poor in class distinction. Given these findings, the SVM model indicated the best choice for this classification task, offering the best tradeoff between accuracy and discriminatory power, though Random Forest and KNN might be good alternatives in light of other practical considerations.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe study was carried out with the aim to build a ML model that can predict contraceptive nonuse among women reproductive age in Rwanda and to identify the key predictors of the nonuse of contraception among the previously mentioned target group.\u003c/p\u003e \u003cp\u003eIt is in that regard that we trained and tested six ML classifiers, which are Logistic Regression, Na\u0026iuml;ve Bayes, Decision Tree, Random Forest, Support Vector Machine and K-Nearest Neighbor to compare their classification performance through different evaluation metrics. Using multivariable binary logistic regression, the study revealed the key predictors of the nonuse of contraceptives among women of reproductive age in Rwanda namely younger age, living in an urban area, having no education, higher status of wealth, never married or separated, higher desired number of children, not working status and having a not working partner. These results are consistent with prior research that shows regional disparities in access and utilization of the contraceptive method [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSimilarly, the odds of nonuse were significantly higher among women living in urban environments, likely due to the very same pattern of urbanization, cultural factors, and the preference for certain methods of using contraceptives [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Education level also emerged as another significant determinant, as nonuse odds were lower among the women who had primary education compared to women who were not educated. This agrees with other studies where education has been shown to enhance a woman's ability to make active choices concerning family planning [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. However, the lack of significance for women with secondary or higher education suggests that other factors may moderate the association between higher education and contraceptive use, such as cultural norms or spousal influence [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRegarding the model performance, Support Vector Machine outperforms the other models in metrics that we employed in evaluation with mainly the best accuracy of all the others at 75%, precision of 78%, Recall of 75%, F1 score of 76%, Cohen\u0026rsquo;s Kappa statistic of 0.51, the highest ROC AUC score (0.83) and the misclassification rate at 25%.\u003c/p\u003e \u003cp\u003eAlthough SVM stands out the best performing, the two last metrics suggest that the model need further improvement and refinement to lessen errors, trying to achieve more agreement rate with the actual outcomes and to reach the least misclassification rate in classifying instances into positives and negatives more accurately. However, these results takes us to achieving the first objective of this research, which was to build a highly performing ML model that accurately predicts contraceptive nonuse in sexually active women in Rwanda.\u003c/p\u003e \u003cp\u003eThese results confirming outperformance of SVM were supported by the findings of the study that found out this model to be the most accurate in predicting the intention to use FP among reproductive \u0026ndash;age woman in Ethiopia(with imbalanced data), while for the balanced dataset, they had found Random forest to be the best model [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Note that, on this turn, they had deployed five same as the current research\u0026rsquo;s models namely LoR, SVM, RF, KNN and NB with the other three models, which are Artificial Neural Network, XGBoost, and AdaBoost.\u003c/p\u003e \u003cp\u003eIn the current study, the use of ML models, particularly Support Vector Machines demonstrates the potential of these methods to enhance the predictive accuracy in health-related research especially contraceptive behavior.\u003c/p\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eThis study was limited to a cross-sectional analysis using six ML algorithms only and could neither provide insights into dynamic interplay of factors influencing contraceptive behavior nor understand causal relationships and identify the trends over time. Therefore, future researches could incorporate longitudinal data and qualitative insights to deepen the understanding of these determinants and further refine predictive models as well as using other algorithms than these six we used in this research.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn the current study, we identified the key predictors of contraceptive nonuse among sexually active women of reproductive age in Rwanda. These are woman age, residence region, education, wealth status, marital status, residence type, total children ever born, working status, partner's occupation, and the desire for more children. Younger women (particularly those aged 15\u0026ndash;24), urban residents, wealthier women, and those desiring more children are at a higher risk of not using contraceptives. Furthermore, Support Vector Machine model performed better than other five classifiers in predicting nonuse of contraceptives with an accuracy of 75%, giving an ROC-AUC score of 82.96%, making it the best model to predict contraceptive behavior in Rwanda. This model would be deployed to predict contraceptive behavior with unseen data for future researches. The findings of this research will promote more focused and customized intervention strategies, particularly targeting younger women, women in urban areas, and women of high socioeconomic status, and this may ramp up contraceptive uptake to improve health and reproductive outcomes.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003e\u003cstrong\u003eSDGs:\u0026nbsp;\u003c/strong\u003eSustainable development goals\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRDHS:\u0026nbsp;\u003c/strong\u003eRwanda Demographic Health Survey\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNISR:\u0026nbsp;\u003c/strong\u003eNational Institute of Statistics of Rwanda\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRPHC:\u0026nbsp;\u003c/strong\u003eRwanda Population and Housing Census\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEAs:\u0026nbsp;\u003c/strong\u003eEnumeration Areas\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eML:\u0026nbsp;\u003c/strong\u003eMachine Learning\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFP:\u0026nbsp;\u003c/strong\u003eFamily Planning\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKNN:\u0026nbsp;\u003c/strong\u003eK-Nearest Neighbor\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSVM:\u0026nbsp;\u003c/strong\u003eSupport Vector Machine\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRF:\u0026nbsp;\u003c/strong\u003eRandom Forest\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDT:\u0026nbsp;\u003c/strong\u003eDecision Tree\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNB:\u0026nbsp;\u003c/strong\u003eNa\u0026iuml;ve Bayes\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLoR:\u0026nbsp;\u003c/strong\u003eLogistic Regression\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eROC:\u0026nbsp;\u003c/strong\u003eReceiver Operating Characteristic\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAUC:\u0026nbsp;\u003c/strong\u003eArea Under Curve\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSMOTE:\u0026nbsp;\u003c/strong\u003eSynthetic Minority Oversampling Technique\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCI\u003c/strong\u003e: Confidence Interval\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eORs:\u0026nbsp;\u003c/strong\u003eOdds Ratios\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eEthics approval and consent to participate\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEthical considerations focused on using anonymized, publicly available data from the Rwanda Demographic and Health Survey (RDHS) 2019/2020, for registered users. The secondary data use did involve directly respondents.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eConsent for publication:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003eNon applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eData availability\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data were accessed at:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;https://dhsprogram.com/data/dataset/Rwanda_Standard-DHS_2019.cfm?flag=1 after the official request and approval by DHS program.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eFunding\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study did not obtain any funding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eConflict of interests\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Authors declares no competing interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eAuthors\u0026rsquo; contribution\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNorbert Nawe\u003c/strong\u003e designed the research, analyzed the results and drafted the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDieudonn\u0026eacute; N.\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eMuhoza\u003c/strong\u003e provided guidance and participated to the design of\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ethe research, reviewed the analysis and the manuscript and approved the full paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eAcknowledgement\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe acknowledge\u0026nbsp;DHS program for providing the approval with data access authorization, which made this study possible.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eAuthor details:\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNorbert Nawe\u003c/strong\u003e is a graduate from African Centre of Excellence in data Science, College of Business and Economics, University of Rwanda, Kigali, Rwanda.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDieudonn\u0026eacute; N. Muhoza\u003c/strong\u003e is an Associate Professor in the Department of Applied Statistics, School of Economics, College of Business and Economics, University of Rwanda, Kigali, Rwanda.\u003cem\u003e\u0026nbsp; \u0026nbsp;\u003c/em\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eJain R, Muralidhar S. Contraceptive Methods: Needs, Options and Utilization. J Obstet Gynaecol India. Dec. 2011;61(6):626\u0026ndash;34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s13224-011-0107-7\u003c/span\u003e\u003cspan address=\"10.1007/s13224-011-0107-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchwandt H, et al. Family planning providers and contraceptive users in Rwanda employ strategies to prevent discontinuation. BMC Womens Health. Oct. 2021;21(1):361. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12905-021-01503-1\u003c/span\u003e\u003cspan address=\"10.1186/s12905-021-01503-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNISR. The 2019-20 Rwanda Demographic and Health Survey (RDHS) | National Institute of Statistics Rwanda. Accessed: Oct. 20, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://statistics.gov.rw/publication/1724\u003c/span\u003e\u003cspan address=\"https://statistics.gov.rw/publication/1724\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSebuhoro D, Ndengo M, Mung\u0026rsquo;atu J, (PDF) MODELLING DETERMINANTS OF CHOICE OF CONTRACEPTIVE METHODS IN RWANDA.. Accessed: Sep. 15, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.researchgate.net/publication/310422798_MODELLING_DETERMINANTS_OF_CHOICE_OF_CONTRACEPTIVE_METHODS_IN_RWANDA\u003c/span\u003e\u003cspan address=\"https://www.researchgate.net/publication/310422798_MODELLING_DETERMINANTS_OF_CHOICE_OF_CONTRACEPTIVE_METHODS_IN_RWANDA\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFrederiksen BN, Ahrens K. Understanding the extent of contraceptive non-use among women at risk of unintended pregnancy, National Survey of Family Growth 2011\u0026ndash;2017. Contracept X. Jan. 2020;2:100033. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.conx.2020.100033\u003c/span\u003e\u003cspan address=\"10.1016/j.conx.2020.100033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFlam S. Machine Learning in Healthcare: Guide to Applications \u0026amp; Benefits, ForeSee Medical. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.foreseemed.com/blog/machine-learning-in-healthcare\u003c/span\u003e\u003cspan address=\"https://www.foreseemed.com/blog/machine-learning-in-healthcare\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWHO. Family planning/contraception methods. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/news-room/fact-sheets/detail/family-planning-contraception\u003c/span\u003e\u003cspan address=\"https://www.who.int/news-room/fact-sheets/detail/family-planning-contraception\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUN. World Contraceptive Use | Population Division. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.un.org/development/desa/pd/data/world-contraceptive-use\u003c/span\u003e\u003cspan address=\"https://www.un.org/development/desa/pd/data/world-contraceptive-use\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNISR. Rwanda\u0026rsquo;s population reaches 13,2 million in 2022 | National Institute of Statistics Rwanda. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://statistics.gov.rw/publication/Rwanda_population_2022\u003c/span\u003e\u003cspan address=\"https://statistics.gov.rw/publication/Rwanda_population_2022\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWB. Rwanda - Fertility Rate, Total (births Per Woman) \u0026ndash;\u0026thinsp;2024 Data 2025 Forecast 1960\u0026ndash;2022 Historical. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tradingeconomics.com/rwanda/fertility-rate-total-births-per-woman-wb-data.html\u003c/span\u003e\u003cspan address=\"https://tradingeconomics.com/rwanda/fertility-rate-total-births-per-woman-wb-data.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePal S. A Comparative Analysis of Machine Learning Algorithms for Predictive Analytics in Healthcare, vol. 72, pp. 10\u0026ndash;25, Mar. 2024.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartin L-C. Machine Learning vs Traditional Forecasting Methods: An Application to South African GDP, \u003cem\u003eWork. Pap.\u003c/em\u003e, Art. no. 12/2019, 2019, Accessed: Oct. 11, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://ideas.repec.org//p/sza/wpaper/wpapers326.html\u003c/span\u003e\u003cspan address=\"https://ideas.repec.org//p/sza/wpaper/wpapers326.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNISR. The 2019-20 Rwanda Demographic and Health Survey (RDHS) | National Institute of Statistics Rwanda. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.statistics.gov.rw/publication/1724\u003c/span\u003e\u003cspan address=\"https://www.statistics.gov.rw/publication/1724\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDambo N, Jeremiah I, Wallymahmed A. Determinants of contraceptive use by women in the central senatorial zone of Bayelsa State, Nigeria: A cross-sectional survey. Niger Med J. Jan. 2017;58:26. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4103/0300-1652.218409\u003c/span\u003e\u003cspan address=\"10.4103/0300-1652.218409\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTesema ZT, Tesema GA, Boke MM, Akalu TY. Determinants of modern contraceptive utilization among married women in sub-Saharan Africa: multilevel analysis using recent demographic and health survey. BMC Womens Health. May 2022;22(1):181. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12905-022-01769-z\u003c/span\u003e\u003cspan address=\"10.1186/s12905-022-01769-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJean de Dieu H, Lambert M. Predictors of contraceptive discontinuation in Rwanda: evidence from demographic and health survey 2019\u0026ndash;2020. Contracept Reprod Med. Apr. 2024;9(1):19. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s40834-024-00282-y\u003c/span\u003e\u003cspan address=\"10.1186/s40834-024-00282-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOtim J. Contraceptive nonuse among women in Uganda: a comparative assessment of predictors across regions. BMC Womens Health. Dec. 2020;20(1):275. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12905-020-01148-6\u003c/span\u003e\u003cspan address=\"10.1186/s12905-020-01148-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKebede SD, Sebastian Y, Yeneneh A, Chanie AF, Melaku MS, Walle AD. Prediction of contraceptive discontinuation among reproductive-age women in Ethiopia using Ethiopian Demographic and Health Survey 2016 Dataset: A Machine Learning Approach. BMC Med Inf Decis Mak. Jan. 2023;23(1):9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12911-023-02102-w\u003c/span\u003e\u003cspan address=\"10.1186/s12911-023-02102-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHaq I et al. Machine Learning Algorithm-Based Contraceptive Practice among Ever-Married Women in Bangladesh: A Hierarchical Machine Learning Classification Approach, in \u003cem\u003eArtificial Intelligence Annual Volume 2022\u003c/em\u003e, IntechOpen, 2022. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5772/intechopen.103187\u003c/span\u003e\u003cspan address=\"10.5772/intechopen.103187\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMulugeta SS, Muluneh MW, Belay AT, Yalew MM, Agegn SB. Reason and Associated Factors for Nonuse of Contraceptives Among Ethiopian Rural Married Women: A Multilevel Mixed Effect Analysis. SAGE Open Nurs. Jan. 2023;9:23779608221150599. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/23779608221150599\u003c/span\u003e\u003cspan address=\"10.1177/23779608221150599\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHabyarimana F, Ramroop S. The Analysis of Socio-Economic and Demographic Factors Associated with Contraceptive Use Among Married Women of Reproductive Age in Rwanda. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.openpublichealthjournal.com/VOLUME/11/PAGE/348/FULLTEXT/\u003c/span\u003e\u003cspan address=\"https://www.openpublichealthjournal.com/VOLUME/11/PAGE/348/FULLTEXT/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZeleke GT, Zemedu TG. Modern contraception utilization and associated factors among all women aged 15\u0026ndash;49 in Ethiopia: evidence from the 2019 Ethiopian Mini Demographic and Health Survey. BMC Womens Health. Feb. 2023;23(1):51. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12905-023-02203-8\u003c/span\u003e\u003cspan address=\"10.1186/s12905-023-02203-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemeke H, Legese N, Nigussie S. Modern contraceptive utilization and its associated factors in East Africa: Findings from multi-country demographic and health surveys. PLoS ONE. 2024;19(1):e0297018. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0297018\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0297018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKotsiantis SB, Kanellopoulos D, Pintelas PE. Data Preprocessing for Supervised Leaning, 2007, Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://publications.waset.org/14136/data-preprocessing-for-supervised-leaning\u003c/span\u003e\u003cspan address=\"https://publications.waset.org/14136/data-preprocessing-for-supervised-leaning\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWes M. Python for Data Analysis, 2nd Edition[Book]. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.oreilly.com/library/view/python-for-data/9781491957653/\u003c/span\u003e\u003cspan address=\"https://www.oreilly.com/library/view/python-for-data/9781491957653/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAWE OO. IASC - Computational Strategies for Tackling Imbalanced Data in Machine Learning | ISI. Accessed: Sep. 01, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.isi-web.org/webinar/iasc-computational-strategies-tackling-imbalanced-data-machine-learning\u003c/span\u003e\u003cspan address=\"https://www.isi-web.org/webinar/iasc-computational-strategies-tackling-imbalanced-data-machine-learning\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTajdini F, Kheiri M-J. (PDF) Recent advancement in Disease Diagnostic using machine learning: Systematic survey of decades, comparisons, and challenges. Accessed: Sep. 02, 2024. [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.researchgate.net/publication/366086539_Recent_advancement_in_Disease_Diagnostic_using_machine_learning_Systematic_survey_of_decades_comparisons_and_challenges\u003c/span\u003e\u003cspan address=\"https://www.researchgate.net/publication/366086539_Recent_advancement_in_Disease_Diagnostic_using_machine_learning_Systematic_survey_of_decades_comparisons_and_challenges\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePedregosa F et al. Jan., Scikit-learn: Machine Learning in Python, \u003cem\u003eJ. Mach. Learn. Res.\u003c/em\u003e, vol. 12, 2012.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWestoff CF. Unmet need for modern contraceptive methods., \u003cem\u003eICF Macro\u003c/em\u003e, Oct. 2012, [Online]. Available: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.measuredhs.com/publications/publication-AS28-Analytical-Studies.cfm\u003c/span\u003e\u003cspan address=\"http://www.measuredhs.com/publications/publication-AS28-Analytical-Studies.cfm\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCleland J, Harbison S, Shah IH. Unmet need for contraception: issues and challenges. Stud Fam Plann. Jun. 2014;45(2):105\u0026ndash;22. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1728-4465.2014.00380.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1728-4465.2014.00380.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBongaarts J. The impact of family planning programs on unmet need and demand for contraception. Stud Fam Plann. Jun. 2014;45(2):247\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1728-4465.2014.00387.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1728-4465.2014.00387.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStephenson R, Baschieri A, Clements S, Hennink M, Madise N. Contextual Influences on Modern Contraceptive Use in Sub-Saharan Africa. Am J Public Health. 2007;97(7):1233\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2105/AJPH.2005.071522\u003c/span\u003e\u003cspan address=\"10.2105/AJPH.2005.071522\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdem JB, et al. Explainable machine learning algorithm to identify predictors of intention to use family planning among reproductive-age women in Ethiopia: Evidence from the performance monitoring and accountability (PMA) survey 2021 dataset. Jan. 2024;25. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.21203/rs.3.rs-3848375/v1\u003c/span\u003e\u003cspan address=\"10.21203/rs.3.rs-3848375/v1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Contraceptives nonuse, Women, Rwanda, Machine Learning, Support Vector Machine","lastPublishedDoi":"10.21203/rs.3.rs-5300030/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5300030/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e: Despite efforts and commitments put in place by the Rwandan Government and collaborating great health organizations, contraceptive prevalence rate (CPR) remains low in Rwanda (64% in 2019 for married women using any method and 58.% for only modern methods. CPR has however been increasing from 45% in 2010 and 48% in 2015. Consequently, unmet need for family planning dropped from 34% to 14% between 2010 and 2020. \u0026nbsp;This study aims to leverage Machine Learning to predict contraceptive nonuse among women of reproductive age in Rwanda.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eA cross-sectional analysis of secondary data was conducted on 2020 Rwanda Demographic and Health Survey. We used six Machine Learning algorithms on the sample of 14,634 women of reproductive age, which were trained and evaluated using various metrics to know the best model. Moreover, multivariable binary logistic regression was used to determine key factors of contraceptive nonuse through Python software and to identify women at higher risk of not using contraceptives, providing valuable insights for targeted interventions and policy enhancements to improve access to reproductive health and family planning services for underserved populations in Rwanda.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Findings revealed that woman age, residence region, education, wealth status, marital status, urban-rural residence, total children ever born, working status, partner's occupation, and the desire for more children are key determinants of not using contraceptives. Younger women, particularly those aged 15-24, urban residents, wealthier women, and those desiring more children are at a higher risk of not using contraceptives. Furthermore, Support Vector Machine model performed better than other five classifiers in predicting nonuse of contraceptives status with an accuracy of 75%, giving a ROC-AUC score of 83%, and making it the best model to predict contraceptive behavior in Rwanda.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e The results from this study suggest the use of Machine Learning to predict contraceptive outcomes accurately. Additionally, by tailoring focused interventions for identified women at higher risk of not using contraceptive could contribute to contraceptive use uptake in Rwanda.\u003c/p\u003e","manuscriptTitle":"Leveraging Machine Learning models to predict contraceptive nonuse among women of reproductive age in Rwanda: A machine learning approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-11 06:38:45","doi":"10.21203/rs.3.rs-5300030/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"aa65b5fd-a904-40b8-b39e-d165c0cb7454","owner":[],"postedDate":"November 11th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-10-09T15:38:22+00:00","versionOfRecord":[],"versionCreatedAt":"2024-11-11 06:38:45","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5300030","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5300030","identity":"rs-5300030","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}