Modeling Socio-demographic Factors as Covariates for Choice of Modern Family Planning Method Among Women of Reproductive Age 15-49 Years a Multinomial Logistic Regression Approach

Abstract

Introduction: Family planning refers to information, ways and methods to help people decide whether to have children and also do it at the right time. It comprises of a varied array of contraceptives and they include pills, implants, intrauterine devices, surgical procedures that limit fertility, and barrier devices such as condoms – as well as non-invasive approaches such as the calendar technique and self-restraint from sexual intercourse. Contraception also involves education on how to conceive, desirable time, as well as remedies for failing to get pregnant (UNFPA, 2019). The World Bank (2019) estimated that the global population stood at 7.5 billion people with fast growth being recorded in low-income countries. The UN (2020) estimates the world population to reach 9.6 billion people by the year 2050 and that Africa is the fastest growing in terms of population and will account for half of this speculated increase. If population growth is not controlled, it can lead to economic stretch on both natural and social resources e.g. health and this can lead to immense suffering and even increase world mortality levels. The study aims to identify the cause of these variations by analyzing the effect socio-demographic factors on modern family planning methods. If the cause of variation among the methods of choice can be identified, then it will be possible to come up with specific messages to target certain groups to encourage the use of contraceptives. Objective: This study aimed at modeling socio-demographic factors as determinants of the choice for modern family planning methods using KDHS 2014 data. Methods: This was a quantitative cross-sectional study with both descriptive and analytical statistics. This study covered the whole republic of Kenya since it employed data records from the Kenya Demographic Health Survey of 2014 collected throughout the country. The study population comprised of eligible women in the reproductive age category (15-49) years and that were sampled randomly from KDHS 2014 data. A sample size of 375 participants was obtained using a difference in proportion formula. Observations that met the inclusion criteria were sampled using simple random sampling method. R software version 4.0.2 was used for sampling. Both descriptive statistics e.g. means and frequencies and inferential statistics e.g. Pearson's chi-square and regression were carried out on the data. The analysis of data was done using R version 4.0.2 and STATA version 13.0. The expected outcomes for this study were that provision of family planning services will be tailor-made to meet the needs of the various socio-demographic groups rather than being general, that there will be an increase in contraceptives consumption as this study will help address gaps in family planning service provision, and the insights gained will help the government in the procurement of contraceptives as per the various socio-demographic groups. xiii Results: The study had 375 participants with a mean age of 30.08 years. The youngest study participant was 15 years while the oldest was 49 years old. In terms of age categories, the largest number of participants was between ages 25-29 years while the least was between ages 15-19 years. A Pearson's chi-square test of association showed that all the socio-demographic factors in this study were significantly associated with the choice of contraception except age; place of residence (p-value < 0.01), religious affiliation (0.01), level of education (p-value < 0.01), socioeconomic status (p-value < 0.01), age of a woman (0.10). All the p-values below 0.05 indicate a significant statistical association between the factors and the contraceptive method of choice. A multinomial logistic regression carried out to compare the choice of contraception among the levels of the covariates produced the following odds ratios, residence (urban) OR = 1.193 (0.941, 1.512) for the use of pill compared to the condom, OR = 1.227 (1.009, 1.492) for implants when compared to the use of condoms. The use of pills, injections and IUCDs was not significantly different as opposed to using condoms when rural women were compared to urban women. The various levels of education also showed differences in some while others were not. Using no education and condom as reference covariate and outcome respectively; for the pill method, primary education OR =0.936 (0.695, 1.260), secondary OR = 0.557(0.237, 1.408), higher education OR = 1.664 (1.282, 2.160). For the injections; primary education OR =0.364 (0.290, 0.455), secondary OR = 1.580(0.982, 2.542), higher education OR = 1.980 (1.650, 2.375). For the use of IUCDs, primary education OR =0.621 (0.485, 0.795), secondary OR = 1.756(1.052, 2.930), higher education OR = 1.898 (1.551, 2.322). Another variable of interest is the socio-economic levels where upper and middle were compared with lower class as a reference covariate and condom as a reference method. The odds ratios were as follows; for the pill, middle class OR = 0.634 (0.442, 0.910), upper class OR = 1.517 (1.117, 2.060). On injections, middle class OR = 1.161 (0.925, 1.456), upper class OR = 0.785 (0.633, 0.972). For the implants, middle class OR = 0.876 (0.668, 1.149) and upper class OR = (1.002, 1.649). The IUCDs was the last method under which socio-economic classes were compared and had the following odds ratios, middle class OR = 1.096 (0.856, 1.404), upper class OR = 0.866 (0.684, 1.097).