Introduction
Unsafe abortion is one of the primary causes of maternal mortality around the world, and maternal deaths are more common in areas where abortion is restricted or outlawed (Say et al., Reference Say, Chou, Gemmill and Alkema2014). According to World Health Organization (WHO), 45% of all abortion are unsafe, while almost all these unsafe abortions happen in developing countries (WHO, 2022). An abortion is considered safe if it is done with a method recommended by the World Health Organization and appropriate to the pregnancy duration, and by the person with the necessary skills. If either of these conditions is not met, the abortion is unsafe (MFS, 2022; WHO, 2022). Abortion is illegal in Bangladesh unless it is necessary to preserve a woman’s life. Menstrual regulation (MR) has, on the other hand, been accessible in the country since 1979 (Guttmacher Institute, 2012). According to the WHO, “MR is the process of uterine evacuation without laboratory or ultrasound confirmation of pregnancy for women who report recent delayed menses” (WHO, 2012). Abortion is performed on the basis of a confirmed pregnancy, while MR procedure is performed when pregnancy is suspected based on menstrual delay by inducing the menstrual cycle (Johnston et al., Reference Johnston, Schurmann, Oliveras and Akhter2011). In 1979, national family planning program of the country included MR program in Bangladesh which allows women to securely confirm nonpregnancy following a missing period and avoid recourse to unsafe abortion (Guttmacher Institute, 2012; Population control and family planning division, 1979). Without a confirmed diagnosis of pregnancy, MR entails the use of either manual vacuum aspiration (MVA) or drugs (Misoprostol with or without Mifepristone) to induce menstruation. It can be carried out by doctors at up to 12 weeks after the last menstrual period and at up to 10 weeks by nurses (Guttmacher Institute, 2012). The government has legalized menstrual regulation, and safe MR services are available in both government and private health facilities (Hossain et al., Reference Hossain, Moseson, Raifman, Gerdts, Biswas and Foster2016).
Abortion-related fatalities and complications continue to be a serious maternal health issue in Bangladesh, despite the fact that total maternal mortality has decreased significantly over the last two decades (El Arifeen et al., Reference El Arifeen, Hill, Ahsan, Jamil, Nahar and Streatfield2014). In Bangladesh, the number of abortions and the ratio of abortions to maternal mortality increased from 2010 to 2016 (Singh et al., Reference Singh, Hossain, Maddow-Zimet, Vlassoff, Bhuiyan and Ingerick2017, Reference Singh, Remez, Sedgh, Kwok and Onda2018). In Bangladesh, there were predicted to be 1,194,000 induced abortions in 2014, with many of these likely being performed in hostile conditions or by unskilled personnel, resulting in serious medical complications such as incomplete abortion, hemorrhage, cervical lacerations, sepsis, uterine perforation, bladder injury, and shock (Guttmacher Institute, 2017; Hossain et al., Reference Hossain, Maddow-Zimet, Ingerick, Bhuiyan, Vlassoff and Singh2017). Abortion also has significant socioeconomic and psychological consequences (El Arifeen et al., Reference El Arifeen, Hill, Ahsan, Jamil, Nahar and Streatfield2014). In this setting, MR could be a viable alternative to birth control, as well as a means of addressing the hazards associated with Bangladesh’s high rate of induced abortion (Alam & Sultan, Reference Alam and Sultan2019). In a study of over 100,000 pregnancies in Bangladesh, it was discovered that in comparison to live birth, MR had a decreased risk of maternal mortality, but abortion had a greater incidence of maternal death. (Rahman et al., Reference Rahman, DaVanzo and Razzaque2014).
Bangladesh’s government took several steps to improve the accessibility of MR services such as training of nurses and midwives, increasing the number of MR service providers, making MR services free, expanding the time period for authorizing MR, creating national MR guidelines on provision of services and quality of service, and integrating medicinal MR, etc. (Benson et al., Reference Benson, Andersen and Samandari2011; Guttmacher Institute, 2012; Hossain et al., Reference Hossain, Maddow-Zimet, Ingerick, Bhuiyan, Vlassoff and Singh2017). However, there has been a significant drop in the use of MR services in Bangladesh (Alam & Sultan, Reference Alam and Sultan2019). In 2014, the number of MR procedures performed in public and private facilities in Bangladesh fell by 34%, to an estimated 430,000 procedures (Marie Stopes Bangladesh, 2021). Hence, it is important to identify the potential determinants of MR use so that intervention could be designed focusing on the increasing use of MR, which can curb maternal mortality in Bangladesh.
Although a cross-sectional study (Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019) was undertaken on MR use in Bangladesh using the previous BDHS data, no study has assessed the prevalence and determinants of MR in Bangladesh using the latest BDHS 2017-18 data set. Furthermore, the previous study advised that socioeconomic inequalities in MR services should be reduced; however, to our knowledge, socioeconomic inequalities in MR services have not yet been studied. Moreover, as high usage of MR services is constantly expected for limiting births in populous countries like Bangladesh, a frequent inspection of prevalence and determinants is essential to track its present state. However, there is a dearth of evidence assessing the prevalence and determinants of MR, and its socioeconomic inequalities in Bangladesh. Hence, this study would be an addition to fulfill the research gaps. This study aimed to investigate the prevalence of and factors affecting MR among ever-married women of reproductive age (15 to 49 years) in Bangladesh as well as to identify the socioeconomic inequalities in MR using nationally representative survey data.
Methods
Data sources and study design
This study extracted data from the most recent BDHS 2017-18. The survey was carried out from October 2017 to March 2018 under the National Institute of Population Research and Training (NIPORT), Bangladesh (NIPORT and ICF, 2020). A two-stage stratified sampling design was used in BDHS 2017-18 to select the households from a list of enumeration areas (EAs). Initially, 250 EAs from urban areas and 425 EAs from rural areas were chosen. These EAs were considered as the primary sampling unit (PSU) and had a total of 20,250 households (NIPORT and ICF, 2020). The survey collected data from 20,127 ever-married women aged 15 to 49 years. The detailed sampling procedure is provided in the final report of the BDHS 2017-18 (NIPORT and ICF, 2020). Participants who were currently pregnant (n=739) and had never heard about MR (n=5,781) were excluded from this study. Then after excluding the missing cases, we used a sub-sample of 12,711 ever-married women aged 15-49 years as the final sample of this study. After adjusting for survey sampling weights and survey design characteristics, the weighted sample size was 12,586.
Outcome variable
In the BDHS 2017-18 survey, MR use was assessed by asking the following question to the ever-married women of reproductive age: have you ever used MR? This question was asked to those women who have ever heard of MR (NIPORT and ICF, 2020). The outcome variable was thus dichotomized and coded as “1” for those who ever used MR service and “0” for otherwise (Yes/No).
Explanatory variables
A thorough literature search was conducted in the following database: PubMed, Scopus and Google Scholar, and based on the previous study (Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019), we selected the potential explanatory variables for this study. We categorized religion as Muslim versus non-Muslim since about 90% of the population in Bangladesh are Muslim (NIPORT and ICF, 2020). The household wealth status (wealth quintiles) was constructed using principal component analysis based on the different household assets such as televisions and bicycles; materials used for housing construction; and types of water access and sanitation facilities which had five wealth quintiles from 1 (poorest) to 5 (richest). The details of the assessment of the household wealth index can be found in the survey report (NIPORT and ICF, 2020). The body mass index (BMI) was classified based on the World Health Organization (WHO) as follows: underweight (<18.5 kg/m2), normal (18.5–24.9 kg/m2), overweight (25.0–29.9 kg/m2), and obese (≥30.0 kg/m2) (World Health Organization, 2020).
The decision-making power was measured from three factors (respondent involvement in decision on her healthcare, decision on large household purchases, and decision on visits to family or relatives) (Ekholuenetale et al., Reference Ekholuenetale, Mistry, Chimoriya, Nash, Doyizode and Arora2021). Using Principal Component Analysis (PCA), the factors were distilled into a more generalized set of weights that score “decision making power” between 0 and 100 The standardized z-scores were used to disentangle the overall assigned scores to low, and high (Ekholuenetale et al., Reference Ekholuenetale, Mistry, Chimoriya, Nash, Doyizode and Arora2021). Wife beating was measured by aggregating responses from women and categorized into low and high. The following items were used: “beating justified if wife goes out without telling husband”, “beating justified if wife neglects the children”, “beating justified if wife argues with husband”, “beating justified if wife refuses to have sex with husband”, and “beating justified if wife burns the food” (Ekholuenetale et al., Reference Ekholuenetale, Mistry, Chimoriya, Nash, Doyizode and Arora2021).
Statistical analysis
Because of the complex nature of the BDHS survey, the data were processed using survey weights prior to analysis. Descriptive statistics including frequencies and percentages were reported. We used the “svy” command for assigning the sample weight to adjust for clustering effect and sample stratification in STATA version 16.0 (StataCorp., College Station, TX, USA). Then, we reported the weighted prevalence of using MR by background characteristics of study participants. Since the BDHS 2017-18 used a two-stage stratified cluster sampling having a hierarchical composition, a single-level analysis model would not be suitable for analyzing such a data set (Khan & Shaw, Reference Khan and Shaw2011). Thus, to reduce the cluster effect that exists in the dataset, a multilevel (mixed-effect) binary logistic regression analysis was used to identify the factors associated with MR use, where clusters were considered as level-2 factor. Both the univariate and multivariate mixed-effect logistic regression models were used to estimate the odds ratios (ORs) and 95% confidence interval (CI). Only the significant variables with a p-value of less than 0.05 in the univariate logistic regression model were included in the multivariate regression model (final model) (Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019). For measures of variation, we also calculated the median odds ratio (MOR) to check the cluster-to-cluster variability in MR use.
Lorenz curve and concentration index (CIX) were used to examine the inequalities in using MR by wealth quintile (Jann, Reference Jann2016; O’Donnell et al., Reference O’Donnell, O’Neill, Van Ourti and Walsh2016). When the concentration index was positive or the Lorenz curve was below the diagonal line (line of equality), it meant that using MR was higher among high wealth-indexed groups (high household wealth groups). Contrary, when the concentration index value was negative or the Lorenz curve lay above the diagonal line of equality, it indicates that using MR was higher among low wealth-indexed groups. In the Lorenz curve, a higher degree of inequality was confirmed by how far away the curves sagged away from the line of equality. Concentration index was used to compute the contrast in using MR (Ekholuenetale et al., Reference Ekholuenetale, Tudeme, Onikan and Ekholuenetale2020). We used Stata version 16.0 (StataCorp., College Station, TX, USA) for analyzing the Lorenz curve and concentration index. The statistical significance was determined at p < 0.05.
Results
Background characteristics
A total of 12,586 ever-married women of reproductive age was included in this study. Of them, more than half (56.61%) were aged ≥ 30 years; about 14% had no formal education, 50% were unemployed, and 90% were Muslim. About 69% of participants were from rural areas and 16% were from families with the highest wealth quintile. Among the study women, 32% women reported that they didn’t use contraceptive methods, and 92% of women had at least one child, while 31.9% of women had a desire for more children. About 36% of women’s decision-making power was low (Table 1).
Table 1. Weighted prevalence of menstrual regulation by background characteristics (N = 12,586)
CI: Confidence Interval
Prevalence of MR
The overall prevalence of MR was 7.64% (95% CI: 7.19 - 8.12). This study shows that the prevalence of MR increased with the increasing age, where the highest prevalence (9.94%, 95% CI: 9.27 – 10.66) was found in older age group (≥ 30 years). Similarly, the proportion of MR uses increased with the increase of wealth status among women. The highest prevalence of MR was found among mothers in the highest wealth quintile (9.79%, 95% CI: 8.76 – 10.92). The proportion of MR was higher in women from urban areas (9.14%, 95% CI: 8.27 – 10.10) compared to women from rural areas (6.99%, 95% CI: 6.48 – 7.55). Among the administrative divisions, the highest prevalence was found in Barishal (10.64%, 95% CI: 8.55 – 13.17), while the lowest was in Mymensingh (3.70%, 95% CI: 2.71 – 5.02). The prevalence of having MR was higher among women who had ≥ 3 children (9.56%, 95% CI: 8.75 – 10.43) compared to those having no child (2.63%, 95% CI: 1.76 – 3.91). The proportion of using MR was higher among those who used the traditional contraceptive method (11.92%, 95% CI: 10.31 – 13.74). The prevalence of MR was also higher among women who had knowledge of ovulatory cycles (8.43%, 95% CI: 7.67 – 9.26) compared to those who had not (7.17%, 95% CI: 6.62 – 7.76) (Table 1).
Factors influencing MR
The multilevel (mixed-effect) binary logistic regression showed that women aged 20 to 29 years (adjusted odds r[AOR]: 2.50, 95% confidence interval [CI]: 1.46, 4.30) and ≥ 30 years were more likely to use MR compared to those aged < 20 years (AOR: 4.17, 95% CI: 2.39, 7.26). Urban women were 24% more likely to use MR compared to rural women (AOR: 1.24, 95% CI: 1.04, 1.47). Compared to the women in the middle wealth quintile, women in the lowest wealth quintile were 44% less likely (AOR: 0.66, 95% CI: 0.51, 0.85) and in the highest quintile were 26% more likely to have MR (AOR: 1.26, 95% CI: 1.03, 1.39). Women who had one or two children (AOR: 1.96, 95% CI: 1.25, 3.09) and ≥ 3 children (AOR: 2.26, 95% CI: 1.40, 3.65) had higher odds of having MR compared to those who had no child. Overweight (AOR: 1.33, 95% CI: 1.14, 1.55) and obese (AOR: 1.37, 95% CI: 1.09, 1.73) women had increased odds of having MR compared to women with normal body mass index. The likelihood of having MR was 39% higher among women who used the traditional contraceptive method compared to women who didn’t use any contraceptive method (AOR: 1.39, 95% CI: 1.12, 1.73). Compared to the Dhaka division, women from Chittagong (AOR: 0.62, 95% CI: 0.46, 0.84) and Mymensingh (AOR: 0.51, 95% CI: 0.36, 0.73) were less like to have MR, while women from Barishal (AOR: 1.44, 95% CI: 1.08, 1.93) were more likely to have MR (Table 2).
Table 2. Multilevel logistic regression for determining the factors affecting menstrual regulation of study participants
OR: Odds Ratio, CI: Confidence Interval
*p < 0.05, **p < 0.01, ***p < 0.001
Socioeconomic inequalities in MR
Household wealth-related inequalities in using MR among ever-married women in Bangladesh were also estimated using Lorenz curves and concentration index [CIX] values. MR use was found to be higher among higher socioeconomic status (SES) groups as the CIX value was positive and the Lorenz curve lay below the line of equality (CIX: 0.095, p<0.001) (Figure 1).
Figure 1. Concentration curve for MR use.
Discussion
While abortion procedures have been illegal since 1860 (under colonial rule), MR provides a means for women to resume menstruation without necessitating confirmation of a pregnancy (Guttmacher Institute, 2017). MR includes the use of drugs such as mifepristone and misoprostol to induce uterine bleeding and evacuation. It is possible that some women obtain these drugs clandestinely and use them inappropriately, leading to uterine haemorrhage among other complications (Guttmacher Institute, 2017). This study aimed to investigate the prevalence and determinants of MR use among ever-married women as well as to identify the socioeconomic inequalities in MR in Bangladesh. The finding of this study shows that one in every thirteen women in Bangladesh used MR in their reproductive age. Women aged 20-29 years and ≥ 30 years, from urban areas, having one or two children and ≥ 3 children, who were overweight and obese, and used traditional contraceptive methods were more likely to have MR. MR use was lower among women from Chittagong and Mymensingh divisions. Besides, MR use was higher among higher socioeconomic status (SES) groups.
We found that the prevalence of MR among women in Bangladesh was 7.64%, while in 2014, it was reported as 12.3% indicating a decline in using MR services in Bangladesh (Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019). A rising trend of contraception use, for instance, increased from 62% in 2014 (Hossain et al., Reference Hossain, Khan, Ababneh and Shaw2018) to 68% in 2018 (Khan et al., Reference Khan, Islam and Islam2022), which might be a plausible reason for why Bangladeshi women are less likely to seek MR services (Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019). Furthermore, evidence shows that social or religious issues, conservative health attitudes, aversion to performing MR, a lack of competent training, inadequate personnel or equipment, and space may all contribute to the significant drop in using MR services in Bangladesh, especially in the Union Health and Family Welfare Centers (UH & FWCs) that mainly provides MR services in the rural regions (Hossain et al., Reference Hossain, Maddow-Zimet, Ingerick, Bhuiyan, Vlassoff and Singh2017; Vlassoff et al., Reference Vlassoff, Hossain, Maddow-Zimet, Singh and Bhuiyan2012). A lack of training in a group of recently employed providers who were hired to replace a large group of providers who have reached retirement age might be contributing to the reduction in the proportion of MR service providers UH & FWCs (Guttmacher Institute, 2017).
According to the findings of this study, women with a higher socioeconomic status were more likely to use MR than their counterparts. In addition, socioeconomic inequalities in MR use were also discovered. Previous studies conducted in developing countries consistently found a significant association between SES and MR use (Ankara, Reference Ankara2017; Bose & Trent, Reference Bose and Trent2006; Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019). Women with higher socioeconomic status have more possession over their reproductive conduct, have a better quality of life, and have better access to health care from both public and private services (Bose & Trent, Reference Bose and Trent2006). Another explanation might be linked to the accessibility of family planning services. Women with lower SES may have less access to family planning services, resulting in lower rates of MR use (Adato et al., Reference Adato, Roopnaraine and Becker2011). In contrast to our findings, previous study has found that higher SES in women is associated with a reduced incidence of MR (Gil-Lacruz et al., Reference Gil-Lacruz, Gil-Lacruz and Bernal-Cuenca2012). However, the study was undertaken in a developed country, which has quite different structural dynamics than developing countries (DaVanzo et al., Reference DaVanzo, Rahman, Ahmed and Razzaque2013).
The MR use was more prevalent among older women compared to their younger counterparts which is consistent with the finding of a prior study (Tan, Reference Tan1983). Earlier literature reported that older women are more likely to utilize safer methods of pregnancy termination than younger women who use less secure methods (DaVanzo & Rahman, Reference DaVanzo and Rahman2014). This could be due to the fact that older women are more likely than younger women to be aware of safe ways of termination and to be aware of the consequences connected with less-safe procedures (DaVanzo & Rahman, Reference DaVanzo and Rahman2014), MR is an effective reproductive health service to terminate pregnancies (Population Council Bangladesh, 1999). Another perception-based study reported that pregnancies in older women are often argued as shameful in Bangladesh, while MR could be an acceptable alternative among them (Tan, Reference Tan1983). These are some of the likely explanations for why older women use MR more than younger women. The number of children or parity was found to be positively linked with MR use. This finding corroborates an earlier study that found a link between the number of living children or parity with the likelihood of obtaining MR services in Bangladesh (Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019). The use of MR services to restrict or postpone births in order to reduce their family size, which is common in South Asian countries (DaVanzo & Rahman, Reference DaVanzo and Rahman2014), might be a plausible explanation of our finding.
Results indicate that women from urban areas were more prone to use MR services than rural women. Differences in the prevalence of MR between rural and urban women might be due to differences in access to health facilities, healthcare-related information, and access to the media. This could also be due to that although MR services are provided by the UH & FWCs and NGOs, mainly in suburban and rural regions, barely half of the UH & FWCs in Bangladesh are able to provide MR services (Hossain et al., Reference Hossain, Maddow-Zimet, Ingerick, Bhuiyan, Vlassoff and Singh2017; Singh et al., Reference Singh, Hossain, Maddow-Zimet, Vlassoff, Bhuiyan and Ingerick2017).
The prevalence of MR use was inconsistent with the geographical variation, which revealed that women in the Mymensingh and Chittagong divisions were less likely than women in the Dhaka division to have MR. According to earlier studies (Kant et al., Reference Kant, Srivastava, Rai, Misra, Charlette and Pandav2015; Kapil Ahmed et al., Reference Kapil Ahmed, Van Ginneken and Razzaque2005), women in Mymensingh and Chittagong may have less awareness and access to MR services, as well as experience more constraints and social stigma compared to women in the Dhaka division. In addition, due to the larger density of governmental and private clinics, hospitals, and NGOs in the capital city (Dhaka) of Bangladesh than in other areas, MR services are largely provided there (Rana et al., Reference Rana, Sen, Sultana, Hossain and Islam2019). However, women from Barishal division were more likely to have MR compared to those from Dhaka division. This finding could be explained by the division-wise adult literacy rate in Bangladesh, where Barishal is ahead with 84.1% among all the divisions (Dhar, Reference Dhar2021). Further research is needed to find out the causal reasons of this geographical disparities in MR use in Bangladesh.
This study’s strength is the generalizability of the findings for Bangladesh, since this survey included nationally representative data from all administrative divisions of the country. In addition, appropriate statistical procedures were used to assess the sample’s weighted prevalence of MR use as well as to identify the associated factors. Besides, this is the first study that measured the socioeconomic inequalities in MR services in Bangladesh, however, the decomposition of the concentration index value may aid in determining the proportion of inequality caused by different explanatory variables. Thus, further study on decomposing the concentration index for socio-economic inequalities in MR use in Bangladesh is recommended. The study had some limitations also. No causality could be established due to the cross-sectional design of the study. The information on utilization of MR services was self-reported by ever-married women, thereby putting at risk of recall bias. Regardless, these limitations do not over-ride the validity and reliability of the findings of this study.
Conclusion
This study found that the usage of MR services was independently associated with age, residential status, geographical location, SES, parity, and contraceptive method use of the women. Also, socioeconomic inequalities in MR use were found, where MR use was higher among higher SES groups. There was also a need to address the socioeconomic inequalities in MR use. As a result, the findings of this study should be assessed in light of existing and planned policymaking. By making sure accessibility and availability of MR services, health policymakers and intervention designers should focus on limiting socioeconomic inequalities in regard to MR services in Bangladesh.
List of abbreviations
MR: Menstrual Regulation; SES: Socioeconomic Status; BDHS: Bangladesh Demographic and Health Survey; NIPORT: National Institute for Population Research and Training EAs: Enumeration Areas; PSU: Primary Sampling Unit; AOR: Adjusted odds ratio; CI: Confidence Interval; MOR: Median Odds Ratio; CIX: Concentration Index; NGO: Non-Government Organization; UH & FWC: Union Health and Family Welfare Center.
Acknowledgements
Author want to thank Demographic Health Surveys (DHS) for providing the datasets with no cost and permit us for using the data for independent research.
Funding
The author received no specific funding for this work.
Conflicts of interest
The author has declared that no competing interests exist.
Ethical approval
Secondary data set was used from the Demographic and Health Surveys (DHS) Programme for this study which is publicly available upon suitable request; therefore, further ethical approval was not required. Details of the ethical procedures followed by the DHS Program can be found in the BDHS report. All the procedures were performed in accordance with the relevant guidelines and regulations.
