Out of the 536 reproductive-aged street beggar women interviewed, data from 40 participants were excluded from further analysis due to pregnancy. The average age of the women was 29.8 years, with a standard deviation of ± 8.4 years. On average, these women had spent 5.4 years in street life, with a standard deviation of ± 5.5 years. Family sizes varied significantly, ranging from women living alone to those with a maximum of 12 members; the mean family size was 3.6, with a standard deviation of ± 2.1. Furthermore, the daily income of these street women varied widely, with a minimum of 5 birr and a maximum of 500 birr, resulting in an average daily income of 50.4 birr and a standard deviation of ± 53.4 birr in Ethiopian birr as presented in Table 1.

Table 1 Descriptive Statistics for continuous Socio demographic characteristics of street beggar women in sidama region, Ethiopia 2023

The study included 496 sexually active beggar women in the Sidama Region of Ethiopia, with the majority aged 25–34 (45.1%) and 35–49 (31.9%). Most participants identified as Protestant (46.4%) or Orthodox (31.9%) and were predominantly married (35.1%). A significant portion (61.5%) had rural backgrounds before engaging in begging, and over half (55%) had no formal education. Many women had large families, with 43.8% having more than three members. Most had been involved in street life for 1–5 years (58.3%), and 16.9% reported disabilities, mainly affecting their hands or legs. Health issues were prevalent, with 32.5% suffering from chronic illnesses, the most common being HIV (32.3%). In terms of modern contraceptive (MC) utilization, only 3.4% of women with a family size of one used MC, while this figure rose to 15.5% for families of two to three and 20% for families with more than three members. Usage was highest among married women (25.6%), followed by divorced (6.8%) and single women (3%). Educated beggar women had a higher utilization of MC (24.6%) compared to non-educated women (14.3%), indicating that educated women are more likely to use MC. In terms of disability, the utilization of MC among disabled beggar women was low (2.4%) compared to non-disabled women (36.5%), suggesting that women without disabilities are more likely to use MC than those with disabilities (Table 2).

Table 2 Descriptive Statistics for categorical Socio demographic characteristics of street beggar women in Sidama region, Ethiopia 2023Socio economic characteristics of street women in Sidama region, Ethiopia, 2023

The study on the socio-economic characteristics of street women in the Sidama Region of Ethiopia revealed significant patterns in modern contraceptive use related to daily income, living conditions, reasons for begging, and additional jobs. Among women earning 30 birr or less daily, 19.8% utilized modern contraception, while 35.1% did not. For those earning between 31 and 40 birr, only 2% reported using it, and among those making 41–50 birr, usage increased slightly to 6.3%. Notably, among women earning more than 50 birr daily, 10.89% used modern contraception, indicating a gradual increase in usage with higher income levels. Out of 155 women living in rented houses, 11.7% used modern contraception, while 19.6% did not. In mass living rooms, 6.85% used it compared to 10.3% who did not. Of the 21 women living in churches or mosques, only 1% used modern contraception. Conversely, 12.9% of those living on the street utilized it, and 1.4% of women living near someone else did.

Regarding the reasons for begging, chronic illness was cited by 57 women, of whom only 3% used modern contraception. Among 47 women who cited the death of a supporter, 2.4% used it, while 2.2% of those reporting disabilities did. Notably, of the 180 women who identified poverty as the reason for begging, 14.7% used modern contraception. Additionally, out of 88 unemployed women, 9.7% used it. In terms of jobs in addition, among 119 women working as daily laborers, 24.4% used modern contraception, while 40% did not. Among 51 women working as housewives, 10.1% used modern contraception, and 14% did not. For the 320 women solely relying on begging, 4% used modern contraception, and 6.3% did not. The 6 women engaged in other jobs, 0.4% used modern contraception, and 0.8% did not as presented in Table 3.

Table 3 Socio-economic characteristics of women beggar in Sidama, Ethiopia in 2023, along with the use of modern contraception

As presented in Table 4, a significant majority of the street beggar women (83.7%) had given birth, with 55% experiencing pregnancy after joining street life. Among these pregnancies, a striking 70.7% were unplanned, primarily due to a lack of knowledge about modern contraceptive services (44%) and fears regarding side effects (31%). More than half (58.3%) had at least one child, while 38.5% reported a history of rape after joining street life. In the past 12 months, 36% engaged in sexual intercourse, with 20.7% pregnant, 82.2% of these pregnancies were unintended. This history of rape underscores their vulnerability to sexual violence, contributing to reproductive health challenges.

Table 4 Reproductive health-related and sexual exposure factors of street women in sidama region, Ethiopia, 2023

Additionally, 18.1% of participants had a history of induced abortion, with 44.5% using self-administered drugs, The history of induced abortion, with a significant proportion resorting to self-administered methods, public health institutions, or untrained/traditional abortionists, highlights the potential risks and lack of access to safe abortion services for these women. While 84.1% had children, 27.6% expressed a desire for more, and 58.5% did not wish to have additional children, reflecting diverse fertility intentions. Utilization of modern contraception was higher among women with a history of rape (26.2%) compared to those without (12.7%). Women planning to have children in the future were more likely to use modern contraception, 27.7% planned to conceive after five years, 11.7% after 3–4 years, and only 1.5% within two years. The utilization of modern contraception varied based on factors such as history of rape and future childbearing plans, indicating the influence of personal experiences and reproductive intentions on contraceptive practices as presented.

The result presented in Table 5 show that approximately 83.4% of street beggar women were aware of modern contraceptive methods, with the majority having information about injectable (62.9%) and implants (28.4%). Friends are the primary source of information for most street beggar women about modern contraceptive use. Around 68.5% of women express approval of modern contraceptive methods, and 66.7% are culturally accepting of their use. However, 76.8% of beggar women do not receive support from their sexual partners in this regard. Additionally, 58.5% of women do not receive advice from healthcare professionals about contraceptive methods. While 63.7% of women have access to modern contraceptive methods, only 59.5% have ever used them. Currently, 38.9% of beggar women use modern contraceptive methods, and 58.7% plan to use them in the future.

Table 5 Awareness and practice of street beggar women towards modern contraceptive methods in Sidama, Ethiopia, 2023

Among those using modern contraceptives, 60% do so to limit births, while 40% use them to space births. The utilization of modern contraception was 38.1% among women who have ever used any method of modern contraception, compared to 0.8% among those who have never used any modern contraceptive methods. This indicates that women who have previously used any modern contraceptive method are more likely to currently use contraceptive methods than those who have not used any modern contraceptive methods before, as presented in Table 5.

As presented in Fig. 1, the most commonly used contraceptive method among street beggar women in the Sidama region is the implant, with 48.7% of women using this method, followed by injectable at 35.8% (Fig. 1).

Fig. 1

Bar graph for currently used MC methods of street beggar women in Sidama region

Bayesian estimation for logistic regression model

The study employed Bayesian analysis to draw conclusions about the parameters of a logistic regression model for women's contraceptive use in the Sidama region of Ethiopia. The Bayesian method provided parameter estimates by sampling from their posterior distributions using Markov Chain Monte Carlo (MCMC) techniques, specifically the Metropolis–Hastings algorithm implemented in Stata software. The MCMC analysis was conducted with 12,500 iterations in two separate chains, discarding the first 2,500 as burn-in, resulting in 10,000 iterations sampled from the posterior distribution. Diagnostic methods were applied to assess the convergence of the Markov chains to the target posterior distribution.

Model assessment

Before drawing inferences and making predictions about the posterior distribution of the model parameters, the researchers conducted diagnostics to assess whether the Markov chain had reached its stationary or posterior distribution. Using the Metropolis–Hastings algorithm with two simultaneous chains, they generated numeric and graphical summaries of the estimated univariate marginal posterior distributions to check for convergence [24,25,26]. The time series plot showed the two separately generated chains were mixed, indicating convergence. The autocorrelation function for each parameter chain also showed good mixing, with autocorrelation disappearing before 20 lags, suggesting the independence of current values from previous ones and the convergence of the model parameters to their target distributions [27]. Additionally, the Kernel Density plot exhibited unimodal density for the coefficients of all predictors, further confirming that the simulated parameter values had converged to the known target distribution.

As depicted in Figs. 2 and 3, the convergence of the Markov chain was initially assessed visually using various plots. In addition to the graphical method, the convergence of the chain to its posterior distribution can be verified using numeric summaries of the estimated univariate marginal posterior distributions of the specified model parameters [26].

Fig. 2

Convergence of Time Series, density and autocorrelation plots for the coefficients of disability

Fig. 3

Convergence of Time Series, density and autocorrelation plots for the coefficients of women approval of MC

The simulation in the study was continued until the Monte Carlo error for each parameter of interest was less than 5% of its corresponding posterior standard error, which validates the convergence and accuracy of posterior estimates [28]. The numeric summary estimates of the MCMC algorithm include Monte Carlo (MC) error, posterior mean, standard error, and a 95% confidence interval for the posterior mean. For ease of presenting the analysis results, the parameter estimates were shown in Table 6 in terms of odds ratio, which is the exponential of the estimates (Exp (Mean)).

Table 6 Posterior distribution parameter estimates of modern contraception utilization of among street beggar women in Sidama region by Bayesian logistic regression model

The Bayesian analysis results presented in Table 6 indicate that several explanatory variables were significantly associated with modern contraceptive utilization among street beggar women in the Sidama region. These variables include age, marital status, education level, disability, chronic illness, employment (in addition to begging), current living situation, history of sexual assault or rape, history of pregnancy after street life, discussions about modern contraception with sexual partners, advice from health professionals, sexual activity in the last six months, time taken to arrive at a health facility on foot, approval of modern contraception, and knowledge about contraceptive methods.

Interpretation of Bayesian logistic regression analysis results

The results of the Bayesian logistic regression analysis indicate that the baseline odds of modern contraception utilization among street beggar women were approximately 0.12. This means that, when holding all other predictors constant, the odds of using modern contraception are about 0.12 times the odds of not using it. For the age group, the odds of modern contraception utilization were approximately 2.69 times higher for women aged 25–34 compared to the reference group of 15–24 years old. In contrast, the odds were approximately 0.59 times lower for women over the age of 34 compared to those in the 15–24 age groups.

Marital status was a significant predictor of modern contraceptive use among street beggar women. The odds ratio for marital status was 4.9, indicating that married women are 4.9 times more likely to use modern contraceptives compared to single women, who serve as the reference category. Similarly, the odds ratio for widowed women was 2.8, suggesting that widowed women are 2.8 times more likely to use modern contraceptives compared to their single counterparts. The odds ratio for the "educated" category was 2.16, meaning that educated women have approximately 2.16 times higher odds of using modern contraceptive services compared to uneducated women, holding all other factors constant.

The odds ratio for the "absence of disability" category was 5.18, indicating that street beggar women without a disability are approximately 5.18 times more likely to use modern contraceptive services compared to those with a disability. Furthermore, a history of sexual assault or rape also emerged as a significant predictor. The odds ratio for the "no history of sexual assault/rape" category was 0.12, suggesting that the odds of using modern contraceptives are approximately 87.8% significantly lower for women with such a history. Sexual activity in the last six months was a significant predictor of modern contraceptive use. The odds ratio for the "not sexually active in the last six months" category was 0.11, indicating that the odds of utilizing modern contraceptives are approximately 89.18% lower for women who were not sexually active compared to those who were, holding other factors constant. This finding suggests that street beggar women who were not sexually active in the last six months are significantly less likely to use modern contraceptive services.

Women who did not receive advice from health professionals about modern contraceptives were 0.25 times less likely to use them compared to those who did receive advice. Additionally, women who took more than 30 min to reach a nearby health facility on foot were significantly less likely to use modern contraceptives compared to those who took less than 30 min 0.10. The odds of modern contraceptive use among women who took more than 30 min to arrive were 0.10, indicating a 90% decrease compared to those who arrived within 30 min. Furthermore, women who did not approve of modern contraceptives were 0.15 times less likely to use them compared to those who did approve. This means the odds of modern contraceptive use were 15% lower among women who did not approve compared to those who did.