Factors associated with contraception and induced abortion among young women in Nepal

This study analyzed data from the 2016 Nepal Demographic and Health Survey (NDHS) supplemented by key indicators from several previous waves of NDHS to identify the sociodemographic profiles of women aged 15 – 24 who were using contraceptive methods and sought an abortion within 5 years before surveys. To augment the abortion analysis, field monitoring data of the Gorkha Safe Abortion (GSA) project were also used. Results from multivariable analyses show that women who ever gave a birth, who were from Province 1, who were economically well-off, and who knew the legal status of abortion, were more likely to practice contraception than their respective counterparts. Results, further, reveal that women practicing traditional methods of family planning were more likely to have an abortion than their counterparts not using any contraception. Women with two or more children ever born were more likely to have an abortion. Women from Karnali Province were most likely to have an abortion. Rich women were most likely to have an abortion than poor women. Regarding safe abortion, it is found that women living in the Tarai area were most likely to have a safe abortion than other geographic areas, due perhaps to family health services being more accessible in the Tarai area. Although the analyses found no relationship between age and abortions, perhaps due to small sample, the GSA project data clearly indicate that women under age 20, and those from Dalit community in particular, were more likely to seek an abortion than women aged 20 – 24. One lesson learned from the GSA project is that the pay-off for increasing access to safe abortion for hard-to-reach populations is high. To understand the complexity of sexual behaviors, contraceptive uses, and abortions among young women, more research using both qualitative and quantitative approaches are needed.


Introduction
Youth aged 15 -24 make up 20% of Nepal's population (United Nations, 2022a). Youth have a high unmet need for contraception, and this has declined very little in the past 20 years in Nepal from 40% in 1996 (Pradhan et al., 1997) to 33% in 201633% in (MOH et al., 2017. Globally, young women are more likely to access unsafe abortion due to stigma (Yokoe et al., 2019). In developing countries, about 40% unsafe abortions occur among women under age 25 and about one in seven women who have unsafe abortions is under High and persistent unmet need for contraception among youth has important implications for reproductive health programming and planning. Youth who have unmet needs for modern contraception are likely to take recourse to traditional methods of contraception such as period or withdrawal methods and their effectiveness is very low. At the same time, as unsafe abortion is common among the youth, it is likely that if unmet need for contraception is not addressed, increasing number of youths will end up in taking recourse to unsafe abortion. Unsafe abortion is a pressing public health concern and is a major contributing factor to high maternal mortality in Nepal (Thapa & Padhye, 2001). The government's aim of reducing maternal mortality ratio (MMR) to <70/100,000 live births by 2030 (NPC, 2017) will not be achieved when unsafe abortion persists among the youth population. As per the past 2016 survey, women aged 15 -24 contribute 55% to age specific fertility (MOH et al., 2017), and therefore, it is important to pay special attention to unmet need for contraception among the youth so that the maximum number of youths are prevented from practicing unsafe abortion. This motivates the present research.
Below in the remaining of this introduction section, some background information about contraceptive use and abortion in Nepal is briefed, followed by reviews on factors associated with contraceptive use and abortion, an introduction of the Gorkha Safe Abortion (GSA) project aiming to help young Nepali women to get safe abortion services, and the objectives of the present research. Subsequently, data sources and methods used for fulfilling the research goals in the data and method section are described, followed by the results, discussion, and conclusion sections.

Brief background on contraceptive use and abortion in Nepal
In Nepal, contraceptive use has increased markedly since 1996. The prevalence rate of using any contraceptive method among currently married women was 32% in 1996 (Pradhan et al., 1997) and increased to 39% in 2001(MOHP et al., 2002), 50% in 2006(MOHP et al., 2007, and 53% in 2016 (MOH et al., 2017). Along with the increase in the use of modern contraceptive methods, increase in the use of traditional methods is also seen in Nepal. In 2006, the proportion of currently married women using traditional methods comprising mainly withdrawal and rhythm methods was 4% (MOHP et al., 2007), which increased to 7% in 2011 (MOHP et al., 2012) and further to 10% by 2016 (MOH et al., 2017).
However, any method can fail, and some are more dependable than others. Research has shown that the chance of an unexpected pregnancy is almost non-existent for couples who rely on sterilization and very low for users of IUD, injectables, or implants. It is moderate for pill and condom users and very high if couples rely on periodic abstinence, withdrawal, or spermicides (The Alan Guttmacher Institute, 1999). This means that increasingly more couples might end up with unwanted pregnancies and thus demand more abortion services. However, despite a steady increase in contraceptive use, the unmet need for family planning is still high in Nepal. For all women aged 15 -49, the total unmet need for family planning was estimated at 18% for 2016, higher for limiting methods (12%) than for spacing methods (6%) (MOH et al., 2017).
Abortion was legalized in Nepal in 2002 (MOH, 2004), and Safe Abortion Policy 2002 and Procedural Process developed and implemented since 2004 (MOH, 2005). The policy called for an expansion of quality comprehensive abortion care services in the country with appropriate numbers of trained and skilled service providers, adequate equipment, and essential drugs. Under this policy, safe abortion services were charged at Nepalese Rupees 1,000/-or US$10, which was out of reach for many women, especially in rural areas, but these services were made free of charge nationwide in 2016 (MOH, 2016).
Before 2002, abortions were totally illegal in Nepal. The Legal Code 1963 (Muluki Ain) of Nepal did not permit the termination of pregnancies even if they were the result of rape or incest or threatened the woman's life. In effect, it equated abortion with infanticide (Ministry of Law and Justice, 1963). Physicians and other medical practitioners were prohibited from recommending abortion or performing abortion without exceptions (Thapa, 2004). In this context, both women sought abortions and providers provided necessary abortion services clandestinely. Most of the abortions that took place were unsafe; only a very small proportion of women, mostly those living in urban or semi-urban areas and able to afford the cost, had access to trained medical practitioners and safe procedures (Thapa & Padhye, 2001).
As a result of the illegal and criminal status of abortion in Nepal before 2002, the conditions under which poor women obtained abortion services were often extremely unsafe. High unintended pregnancy makes women look for means to terminate it and in the absence of readily available safe abortion facility women who are likely to seek unsafe methods of abortion. This situation is likely to contribute to high maternal morbidity and mortality. Many are still dying during or shortly after pregnancy due to unsafe abortion (The Kathmandu Post, March 31, 2022).

International Journal of Population Studies
MMR in Nepal was estimated at 790 deaths/100,000 live births in 1990 (WHO, 2014). In a hospital-based study of abortion in Nepal pre-legalization, deaths from abortionrelated complications accounted for over half of all maternal deaths (Thapa et al., 1992). The recognition that illegal abortions were unsafe and contributed to Nepal's high maternal mortality was instrumental in the advocacy efforts to legalize abortion (Shakya et al., 2004). Indeed, MMR had improved and gone down to 239 deaths/100,000 live births for several years before the 2016 survey (MOH et al., 2017). However, MMR of 186/10000 live births still remains among the highest worldwide (WHO, 2022), and unintended pregnancy was around 19% in 2016(MOH et al., 2017. Under the current law, pregnancy can be terminated up to 12 weeks for any reason and up to 18 weeks for pregnancy resulting from rape or incest with the pregnant woman's consent. Recently, a provision has also been made to allow pregnancy termination as late as 28 weeks (The Kathmandu Post, 03 July, 2019), but this has not been operationalized (The Kathmandu Post, 27 August, 2019). However, the legalization of abortion does not seem to have much impact; the proportion of women aged 15 -49 who are aware that abortion is legal in Nepal increased by only 3% points from 38% in 2011 (MOHP, 2012) to 41% in 2016 (MOH, et al., 2017). According to the Nepal Demographic and Health Survey (NDHS) 2016, this knowledge is lower among older women aged 40 -49 (34%), women living in rural areas (36%), women with primary (33%) or no education (28%), and poor (38%) and very poor (30%) women (MOH, et al., 2017). By caste/ ethnic groups, this knowledge is highest among the highest ranked group, namely, Chhetri/Bahun (48%), followed by the second highest ranked group, that is, Janajati ethnic group (40%) and the lowest ranked group -the Dalit (34%). After Nepal became a federal state in 2015, health care has been devolved to each province. Although safe abortion is free and legal, 50% of abortions accessed by young women in 2016 were "unsafe, " that is, performed at unauthorized facilities (MOH et al., 2017). The worrisome situation is that unsafe abortion is reported to be on the rise in the country due to lockdowns during the COVID-19 pandemic (Gorkhapatra, October 07, 2020).
Gender-biased sex selection (GBSS) in favor of boys is a symptom of pervasive social, cultural, political, and economic injustices against girls and women. GBSS can be measured using sex ratio at birth (SRB), a comparison of the number of boys born versus the number of girls born in a given period. According to the WHO, 2011, when many more boys are born than girls, it is a sign that sex selection is taking place (WHO, 2011). In Nepal, high son preference (Brunson, 2010) and prevailing discrimination against girls are factors contributable to sex selection (Nanda et al., 2012). Although sex selective abortion is strictly prohibited by law and punishable, it is also possible that people may take advantage of liberal abortion to fulfill their wishes. In addition, the situation of stagnant contraceptive prevalence but declining fertility in the past 10 years may indicate that women are using abortion as a family planning method (Gorkhapatra, August 3, 2019).
Despite the legalization of abortion and the expansion of maternal health services, such as increased access to safe motherhood services including abortion services, it appears that the sexual and reproductive health needs of women are not being met. Furthermore, the existing accredited health facilities are not functioning well due to the lack of sufficient quantity of drugs and equipment, the frequent transfer of trained abortion service providers, and weak abortion related infrastructure development stand as a serious barrier to women seeking abortion services (Wan-Ju et al., 2017). Nevertheless, Nepal is committed to Sustainable Development Goal 3 of achieving an MMR of 70 by 2030 (NPC, 2017), which calls for more focus on reproductive health and rights including safe abortion.

Factors associated with contraception and abortion
Numerous studies have examined factors associated with the prevalence of contraceptive use and the prevalence of abortion among young women (Bayer et al., 2011;Munakampe et al., 2018). Studies have revealed that lack of or limited knowledge, lack of sexuality education and limited access to services, high risk of misperceptions, and harmful social norms surrounding premarital sexual activity and pregnancy could be major barriers to use contraception and abortion services (Campbell et al. 2006;Siziya et al. 2008). These obstacles apply equally to adolescents and young people of Nepal . While several efforts have been made to understand young people's knowledge, attitudes and practices regarding contraception and safe abortion, systematic research on this matter, and related contexts remain limited. Increase in couple's knowledge that they can plan their family size leads them to search for contraceptive methods.
The relationship between contraceptive use and induced abortion is complex due to interactions of several interrelated factors that range from social, cultural, and economic, from how ideal family size is determined, and from the demand for contraception and abortion, to another set of variables related to the quality of reproductive health services (Phiri et al., 2022;Senlet et al., 2001). Modern contraceptives are considered the safest International Journal of Population Studies methods to help couples realize better family planning. Despite the availability of various contraceptive methods, communities in both the developed and developing countries continue to register high rates of unintended and unwanted pregnancies which contribute to a higher prevalence of abortions (Haub & Kaneda, 2014).
An analysis of data of women aged 15 -49 from 2011 Nepal Demographic and Health Survey found that women of older ages (35 years and above) were less likely to undergo both abortion and unsafe abortion. Educated women were more likely to undergo an abortion along with those who had no knowledge of legal abortion. Being rich was protective against unsafe abortion. Child spacing was the most common reason for abortion (Yogi et al., 2018).
The above brief review of contraceptive use and abortion among women aged 15 -49 shows that, although increases in contraceptive use ultimately led to decreases in induced abortion rates, trends have been following different paths in individual country settings. Abortion rates and contraceptive practice often rise simultaneously because they are affected by the same underlying social and demographic factors (Noble & Potts, 1996). Similarly, women who practice contraception are often those who are more likely than others to undergo abortions, again due to bidirectional causality. In these situations, increases in contraceptive use may reduce reliance on abortion by keeping abortion rates stable or by preventing an even more rapid increase in abortion rates.

Objectives of the present study
Research studies on abortion in Nepal are very limited. NDHS is carried out every 5 years and the survey started including abortion section since 2011 which provide basic quantitative indicators. Data on abortion incidence in Nepal are hardly available except for one study by Puri et al. (2016) that show 42/1000 women of reproductive ages 15 -49 in 2014 based on Nepal's National Health Facilities Survey and Health Professionals Survey. Another important limitation in the existing literature is the paucity of research on how women accessed family planning services during the crisis. This is extremely important and closely relevant given the current situation of the ongoing COVID-19 pandemic. A recent estimate shows the contraceptive prevalence rate in Nepal among currently married women declined from 53% in 2016 (MOH et al., 2017) to 33% in 2020 due to disturbance in supply of contraceptives (Global Financing Facility-GFF: https:// www.globalfinancingfacility.org/CoVid19/).
Although no studies on the theme were available by far, it is fortunate that Nepal implemented the GSA project during mid-2018 to June 2020 (including the first period of the COVID-19 pandemic, April-June 2020) in three palika (subdistrict) in Gorkha district that aims to empower women to access safe abortion services (see Map of Gorkha in Figure A1). The data collected in this GSA project enable researchers to investigate how family planning services were delivered to the most vulnerable populations in the area. Research using such data would have added value to the existing literature and have important policy implications on how to improve family planning services on the general as well as in crises, such as in this ongoing pandemic.
Based on the above reviews and consideration, this paper is written with three objectives in mind to fill the research gap. First, using national data from NDHS of different years, it attempts to show different family planning practices by sociodemographic characteristics of adolescents and youth. Second, it examines factors associated with abortion and safe abortion, and third, using the Gorkha project monitoring data on abortion, it investigates how increasing access to functioning safe abortion services could pay off even in difficult times such as in the COVID-19 pandemic lockdown periods.

Data sources
For national level data analysis on family planning and abortion, nationally representative sample survey data from the Nepal Demographic and Health Survey (NDHS) 2016 (MOH et al., 2017) were the main source, but for comparison purposes, NDHS data from 1996 (Pradhan et al., 1997(MOHP et al., 2002(MOHP et al., 2007(MOHP et al., 2012 were also used. In all, 12,862 women of reproductive age 15 -49 were interviewed in 2016 and the corresponding figures were 12,674 in 2011 and 10,793 in 2006. Interviews were conducted with only ever-married women in 2001 and 1996 and the numbers of women interviewed were 8,429 in 2001 and 8,726 in 1996, respectively. However, only the NDHS 2016 collected information about the residence of province because when the earlier surveys were conducted that the country was not divided into federal provinces. There are seven provinces in Nepal (Province No.1, Madhesh Province, Bagmati Province, Gandaki Province, Lumbini Province, Karnali Province, and Sudurpashchim Province; see Appendix). Therefore, for analyses of relationships between contraception, abortion, and sociodemographic variables, only the NDHS 2016 data were used with 4,568 women aged 15 -24.
NDHSs deal with a number of topics, including but not limited to fertility, family planning, infant and child health and mortality, maternal health, and HIV/AIDS-related https://doi.org/10.36922/ijps.v7i1.291 International Journal of Population Studies knowledge and behaviors and mass media exposure. Only NDHS reports of 2011 and 2016 contain separate section on abortion although induced abortion rate as a component of pregnancy outcomes is reported in all NDHS reports. Some details about sampling designs, field work, and data quality of NDHSs can be found at https://dhsprogram.com In addition to five national NDHS datasets, regularly collected abortion monitoring data from the project Empowering women to access safe abortion services in Gorkha implemented in three palika (subdistrict) in Gorkha district of Nepal (see Map in Figure A1) during mid-2018 to June 2020 (including the first few months of the COVID-19 pandemic, April-June 2020) are also used to augment the analysis.
The Gorkha project was funded by the Safe Abortion Action Fund (SAAF) and the International Planned Parenthood Federation (IPPF) (London) and was implemented by the Population, Health, and Development Group (PHD Group -a local NGO). The objectives of the project were to empower women and girls to realize their rights to sexual and reproductive health and for them to be informed and able to access contraception and safe abortion when needed.
Gorkha District was the epicenter of the 2015 devastating earthquake (NPC, 2015); nearly every house was hit by the tremor. People suffered and women and girls were impacted the most as a large number of women and girls who engaged in income-generating activities from their homes, incurred additional losses of homebased economic resources, and assets essential for their livelihood and well-being (NPC, 2015). The 21,000 women of reproductive age (WRA) in the three palika of Gorkha District were the project's main target group. In addition, about 2,000 boys of average age 16.2 years and 2,100 girls of average age 16 years in grades 9 to 12 in the project areas also benefited from basic sexuality education. This paper examines the association between sociodemographic characteristics (including caste/ethnicity) and access to safe abortion in young women aged 15 -24, both nationwide and in the project area.

Dependent variables
This study has three dependent variables that are all dichotomous variables. One dependent variable is whether a participant was using any of contraception at the time of survey; the second is whether a participant had an abortion within 5 years preceding the survey, and the third whether such an abortion is a safe abortion. Any respondent using any of the birth control methods, such as Pill, Intra Uterine Device (IUD), injections (popularly known as Depo Provera), male condom, female sterilization, male sterilization, periodic abstinence or rhythm method, withdrawal, and implants/Norplant, is categorized as currently using a contraceptive method. A respondent was classified as having had an abortion if she obtained such abortion at a government accredited medical facility or at any other health facility, although as a rule, only government accredited health facilities can provide abortion services. In principle, a government facility that has not been accredited for safe abortion service by the Ministry of Health and Population cannot perform any abortion service. Any health facility that is private or non-governmental that has been accredited by the government can provide safe abortion services. In practice, however, NGO or private health facilities not accredited by the government also provide abortion services, but such abortions are regarded as "unsafe. " Women performing abortion at home using modern drugs such as Mifepristone and Misoprostol bought from a pharmacy and abortions performed using crude methods are all regraded as "unsafe. "

Explanatory variables
Having reviewed the literature earlier on reproductive health including contraception and abortion in the context of Nepal and elsewhere, a number of factors have been identified that are associated with women's access to contraception and abortion. These variables are grouped within three categories: demographic characteristics (age, marital status, caste/ethnic groups, and number of children ever born), socioeconomic characteristics (residence, ecological region, province, wealth index, and education), and knowledge (knowledge of legal status of abortion). For abortion models, the type of contraception (modern methods or traditional methods) currently used is also included as an independent variable. The traditional methods used by young women in 2016 included rhythm and withdrawal methods (MOH et al., 2017). Therefore, ten independent variables were used for contraception models and 11 independent variables for abortion models.
Education is referred to as responsible for turning knowledge into practice (Martin & Juarez, 1995). The household wealth index is a composite measure of the cumulative living standard of a household (including assets such as the type of material used for flooring, water, and sanitation facilities and the possession of televisions). The methodology for construction of the index, based on the principal component analysis, is described in detail elsewhere (Rutstein & Johnson, 2004). The variable was classified into five categories using quintile (poorest, poor, middle, rich, and richest).

International Journal of Population Studies
Demographic variables included respondent's age (15 -19 years and 20 -24 years), marital status, and caste/ethnic groups categorized as Chhetri/Bahun, Janajati, Tarai, and Dalit. In the 2011 census about 81% of the Nepalese reported their religion as Hindu and thus locating themselves within the caste system (CBS, 2012). In caste hierarchy, Chhetri/ Bahun are ranked the highest followed by Tarai (those aboriginally living in the Tarai area excluding Dalit), Janajati, and Dalit -the lowest ranked group (Bennet et al., 2008). Socioeconomic variables included urban-rural residence (urban vs. rural), ecological region (Mountain, Hill, and Tarai area), residence of province (seven provinces), wealth index (poorest, poorer, middle, richer, and richest), and educational attainment (no education, primary school, secondary school, and school leaving certificate or higher).

Analytical strategies
To obtain the profile of young women who were practicing contraception and those who obtained an abortion service in the past 5 years preceding the survey of 2016, basic statistical techniques such as frequencies, cross tabulations, and Chisquire tests were conducted. In this paper, multivariable logistic regression analyses were employed using SPSS v.21 to cross-check the results regarding who practiced contraception who obtained abortion and examined who opted for safe abortion. In addition, bivariate analyses and Chi-square tests were also conducted using project monitoring data on women 15 -24 only. These young women obtained safe abortion services during the stringent lockdown of COVID-19 pandemic period January -June 2020 and during the 1-year period just before the lockdown.

Sample description and odds ratios of use of contraception
Nationwide, contraceptive prevalence rate among married young women steadily increased from a mere 10% in 1996 to 17% in 2001 and further to 23% in 2011 ( Figure 1). However, the proportion of young women using contraception stagnated at 21% between 2011 and 2016. Among the adolescents aged 15 -19, the contraceptive use was much lower than their youth counterparts, exhibiting an increasing gap of about 10% points since 2001 between the two groups. Table 1 presents the descriptive analysis of the young women aged 15 -24 who were currently using contraceptive methods at the time of the survey in 2016 and also presents chi-squire tests of the significance of the association between the various background characteristics and the use of contraception. In 2016, the sample of young women aged 15 -24 was 568 (Muslim women not included in the analysis). Of these 4,568 women, 15.1% were currently using any method of contraception, while the corresponding figure for all women aged 15 -49 was 40.8% (MOH et al., 2017).
Among the demographic variables, as shown in Table 1, age was significantly associated with the use of contraception, with more older women (24.9%) practicing contraception compared to their younger counterparts (6.4%). The prevalence of contraceptive use among married young women was 30% and it was virtually non-existent among currently not married. Among the four broad caste/ethnic groups of women, the highest proportion of women practicing contraception was seen among Janajati women (17.7%) in comparison with the lowest proportion among Dalit women (12.8%). Equal proportions (13.6% each) of women of the Tarai caste and the Chhetri/ Bahun caste practiced contraception. Contraceptive use significantly increased with increasing number of children ever born: only 3.9% women practiced contraception without having any children ever born in comparison with the corresponding figures 34.8% for women with one child and 40.8% for women with two or more children.
The association between contraceptive use and socioeconomic variables was found significant for provinces and education. Among the seven provinces, the highest proportion (19.3%) of young women in Province 1 practiced contraception followed by Gandaki (15.9%), Karnali (15.7%) provinces, and so on while the contraceptive use was the lowest in Madhesh province (11.9%). Urbanrural residence, ecological region, and the wealth index were not significantly associated with contraceptive use ( Table 1). Knowledge of the legal status of abortion was not significantly associated with contraceptive use either. Table 1 shows that higher contraceptive use among less educated women compared to their educated counterparts, although it was not statistically significant. Table 2 presents the odds ratios of practicing contraception among young women from multivariable  All estimates were weighted. Per cents may not sum to 100 due to rounding. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001. a Currently not married also includes 6 women aged 15 -19 and 14 women aged 20 -24 formerly in union. SLC: School leaving certificate. International Journal of Population Studies logistic regression models. The results regarding the demographic characteristics of women who practiced contraception were supported in the multivariable analyses. Among all the samples, women aged 15 -19 was less likely to use contraception, which is mostly because they were not married and were involved in sexual activity less frequently. NDHS 2016 shows that 72% of women aged 15 -19 never had sex in the last 1 year and among those who were sexually active only 15% had sexual intercourse in the past 1 month compared to 40% among their counterparts aged 20 -24 (MOH et al., 2017). The results among married women verified the no differences for use of contraception between these two age groups.
The number of children ever born stood out prominently: young women with two or more children ever born were about as much as 25 times of and women with one child were about 15 times of the odds of practicing contraception among the entire sample, compared with women who had no children (the higher odds ratios were reduced to 6 times and 3 times among the married sample). Women of the Janajati ethnic group were more likely to use family planning methods than adolescents and women from other caste/ethnic groups. Young women from Madhesh province were significantly less likely (65%) (OR = 0.35) (it was 69% in the married sample) than those from Province 1 to practice contraception. Young women from Gandaki and Lumbini provinces were also less likely than those young women from Province 1 to practice contraception. The odds of using contraception were the highest in Province 1 compared to all other provinces. This is also supported by simple contraceptive use prevalence among adolescents and youth by provinces, as shown in Table 1 above. Odds ratios were weighted. SLC: School leaving certificate. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.

International Journal of Population Studies
The multivariable analysis shows that the economic status of respondents was positively related to the practice of contraception. Compared to women from the poorest wealth quintile, women belonging to the rich quintile were significantly 1.7 times more likely to use contraception (it was the same in the married sample) and it is even higher (2.7 times) for women from the richest wealth quintile (it was 3.4 times among the married sample). The multivariable logistic analyses largely confirm the results for the socioeconomic characteristics except for the association between educations. Knowing about the legal status of abortion was associated with higher odds of contraceptive use among married women, although such higher odds were not significant among the entire sample. Table 3 presents descriptive profiles of young women aged 15 -24 who obtained an abortion in the past 5 years preceding the survey of 2016. It also presents the results of Chi-square tests of the significance of the association between the various background characteristics and having obtained an abortion. Overall, among all young women aged 15 -24, about 1.8% had an abortion during the 5 years preceding the survey. There was a significant distributional difference in the prevalence of abortion over age, marital status, children ever born, contraceptive use, and wealth index.

Descriptive profile of women having an abortion
The results in Table 4 show that women who ever gave one birth were nearly 6 times (OR=5.72) more than to have an abortion than women without any birth. The odds ratio was even higher among women who ever gave two or more births (OR= 7.14). The higher ratios for these two subgroups were reduced to 2 times and about 3 times in the married sample. Women who were using modern traditional contraceptive methods were 6 times (OR = 6.17) (4.4 times in the married sample) more likely to have an abortion than their counterparts using no contraception, implicating either a low effectiveness of traditional methods of family planning or lower sexual activities among those who were not using any methods. No difference in prevalence of abortion was found between women who were not using any contraceptive method and women who were using modern contraceptive method. The richer a woman, the greater the odds they had an abortion. Women from Karnali Province had nearly 6 times higher odds of having an abortion compared to women from Province 1, but there was no provincial difference for other provinces compared to Province 1. The results are very similar among the married sample with some reduced odds ratios for few factors as we also noted above.

Descriptive profile of women obtaining safe abortion services
In NDHS 2016, a total of 492 women aged 15 -49 reported having had an abortion 5 years preceding the survey (MOH et al., 2017), and 51% of them were performed at government authorized health facilities. Among the youth aged 15 -24, the total number of recent abortions in the past 5 years was 83 with 42 abortions that were "safe" or performed at government authorized health facilities and the remaining 41 abortions were "unsafe" (see Table 5). Caste/ethnic group and the number of children ever born were the only demographic variables associated with abortion safety. It was only ecological region that was significantly associated with abortion safety among the five socioeconomic variables. Knowledge of the legal status of abortion was also associated with abortion safety.
Of the total recent 17 abortions reported by young women in the Tarai area in the past 5 years, 88% were performed at "safe" facilities, while the corresponding figures for other caste/ethnic groups were 38% for Chhetri/ Bahun high castes, 41% for Janajati or indigenous ethnic groups, and 44% for Dalit castes (Table 5). The number of children ever born was also significantly associated with abortion safety. A large proportion (73%) of young women without having live birth were putting them at risk of going to "unsafe" health facilities for an abortion. More young women (51%) with one child ever born also took recourse to "unsafe" health facilities for abortion, while nearly onethird (65%) young women with two or more children ever born visited "safe" abortion health facilities for abortion (Table 5).
Proportionately, more women not currently using contraception were accessing safe abortion service than those who were using any contraception, although the relationship is not statistically significant. Women not using contraception were accessing "unsafe" abortion service; perhaps, the pregnancy was unplanned or was due to forced sex. For wealth quintiles and education, it appears that accessing safe abortion services decrease with rising economic status and education, although the association is not statistically significant. Higher proportion (60%) of women having no knowledge of the legal status of abortion were reporting accessing safe abortion service compared to their counterparts (42%) aware of the legal status of abortion (Table 5).

Safe abortion in Gorkha project area
The implementation of the PHD Group women empowerment and the safe abortion project in Gorkha has some good experience to share. The project was supplementing the government's efforts of expanding safe International Journal of Population Studies All estimates are weighted. Per cents may not sum to 100 due to rounding. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001. a Among women currently not in union there was one woman reporting abortion but due to rounding it is reported as "zero;" SLC: School leaving certificate. Provision of safe abortion services calls for, among other things, training of health personnel, supply of drugs, and equipment, record keeping of clients receiving service and follow-up of abortion clients for 2 weeks and taking care of post abortion complications (PAC). From January 2019 to end of June 2021, 348 women received safe abortion service in the project areas. Of them only four women had PAC and they were well taken care of following the procedure.
Monitoring data from PHD Group's project in the Gorkha district indicate that despite difficult times due to COVID-19 lockdown which caused cancellation/ postponement of several proposed project activities a large number of women and girls utilized safe abortion service. During the first 3 months of this period, the number of women and girls accessing safe abortion service was 67 (23 among women aged 15 -24 and 44 among women aged 25+), while the corresponding figure increased to 106 (28 among women aged 15 -24 and 78 among women aged 25+) in the following 3 months, representing a growth of 58.2% during the lockdown (Figure 2).
A number of reasons have been put forward for the increase in client flow during the pandemic. First, due to mothers' group meetings which are organized by Female Community Health Volunteers (FCHVs) attended by women of reproductive age (WRA), more girls and women in villages are now aware that abortion is legal. Second, project employed Female Field Facilitators (FFFs) and FCHVs do compassionate counseling to local women and girls about abortion and sometimes help them visit safe abortion facility. Third, free and quality safe abortion service is available. Fourth, local safe abortion facilities are never stock out with necessary drugs and the service is available 24 hours and 7 days a week. Fifth, service facility is readily accessible and nearby. Sixth, due to lockdown, women and girls are unable to visit health facility for contraception, and in several cases, the facilities are facing shortage of contraceptives resulting in more girls and women getting pregnant. Seventh, due to lockdown, frequency of meetings of family members increased giving rise to more unplanned pregnancies compared to other normal times making girls and women to opt for abortion. Eighth, due to introduction of sexuality education in schools by the project, adolescent girls and boys are aware of consequences of unwanted pregnancies and what to do about them. Finally, due to project and other development activities, women and girls are more empowered, have learned about reproductive health and rights, and make their own decision about sex and sexuality.
The monitoring data of the GSA project indicate that adolescents and youth aged 15 -24 from Dalit community were significantly more likely to utilize safe abortion service before the COVID-19 pandemic. For instance, during January 2019 -June 2020, among the 15 -19 and 20 -24-year olds combined, the proportion of Dalit youth utilizing safe abortion service in the project areas was 38.5%, while among their high caste or non-Dalit counterparts, the corresponding proportion was 26.8% (Table 6A). During lockdown period from March 24, 2020, to July 21, 2020 too, the proportion of Dalit youth utilizing safe abortion service stood at 41.8% which was https://doi.org/10.36922/ijps.v7i1.291 significantly higher than the proportion of 23.7% utilizing safe abortion service among the non-Dalit or Chhetri/ Bahun and Janajati youth (Table 6B). Although as the national data show proportionately that fewer Dalit women were aware that abortion is legal in Nepal than their non-Dalit counterparts, they utilized safe abortion service more in PHD Group project catchment areas. Field enquiry revealed that over two-thirds of the under 20 women and girls were unmarried, although marital codes were not included in the data set due to sensitivity.

Discussion
Using both the Nepal national DHS and some project specific monitoring data, this paper reviewed the trend of proportion of contraceptive use over time from 1996 to 2016 from several DHSs, and the prevalence of abortion among young women aged 15 -24 from the 2016 DHS. The study also investigated the factors associated with the prevalence of contraceptive use and the prevalence of abortion based on multivariable regression models and examined the distributional difference in abortion safety for some major sociodemographic characteristics.
This study revealed several findings worthy of reporting. Contraceptive use data of various years reveal that since 2006 among the currently married adolescents use of modern contraceptive methods has remained nearly constant at 14% and among all adolescents, it declined from 4.5% in International Journal of Population Studies Percentages were weighted. Percent may not sum to 100 due to rounding. SLC: School leaving certificate. The statistical tests were based on the Chi-square tests. *P ≤ 0.05; **P ≤ 0.01  Among the women aged 20 -24, the use of contraceptive methods declined from 22.4% in 2006 to 17.9% in 2016 and the corresponding figures for the use of traditional methods increased from 2.3% to 6.1%. These findings indicate that either the modern methods of contraception were increasingly becoming unpopular among both the adolescent and young women, or there was increased inaccessibility over the period. These imply that family planning programs in Nepal need to be strengthened and the young women need education and information on the usefulness of modern contraceptive methods instead of relying on traditional methods. Perhaps, there are misconceptions about modern contraceptives methods which need to be removed.
Regression analyses show that women who were aged 20 -24, from Janajati ethnic group, and in rich/richest wealth index groups, and ever gave a birth, were more like to practice contraception than their counterparts who are adolescents, have fewer children ever born, poorer wealth index groups, and from Chhetri/Bahun or Dalit castes. The finding that women in their 20s are more likely to obtain an abortion than adolescents could be because the women in the former group are sexually more active (40% sexually active within the past 4 weeks) than adolescents (Cont' d...) Contraception and abortion in Nepalese young women International Journal of Population Studies One girl aged 14 is added. Pearson Chi-square test significant at *P < 0.10; **P < 0.05. (15% sexually active within the past 4 weeks) (MOH et al., 2017) and have higher chance of becoming pregnant and, therefore, have a higher rate of unplanned pregnancies. Among the seven provinces contraceptive use was the lowest among young women from Madhesh Province and the highest in Province 1, which calls for special focus of family planning programs in Madhesh.
Despite limitation of sample size, our regression analyses of abortion among young women aged 15 -24 also provide a glimpse of young women who are likely to take recourse to abortion in case that the pregnancy was unwanted. The results showed that while women who ever gave a birth were more likely to seek an abortion compared to women who did not give a birth, they were more likely to seek unsafe abortions. Results, further, show that women who were well-off were more likely to seek an abortion, and that if they sought abortions, they were more likely to seek safe abortions. Women who were using the family planning services, especially those who were using traditional methods, were likely to seek an abortion service and obtain the abortion services at "unsafe" facilities. By contrast, those who were not using any contraceptive methods were less likely to seek abortion services and if they obtained services, they were more likely to obtain it from authorized facilities. Women in Madhesh Province were least likely to seek an abortion, while women in Karnali were likely to seek abortion services. Among young women seeking an abortion, it was found that women from the Tarai caste or the Tarai region were likely to have a safe abortion.
The likelihood of seeking an abortion among women who had been practicing traditional methods of family planning may reflect their mistrust for modern contraceptive methods or because they are unable to obtain modern methods that they wished. Increasing use of traditional methods indicates not just unavailability of contraceptives but it also implies the entrenched belief that contraceptives have side effects and may even make woman infecund. As fecundity is highly valued, it is challenging for family planning program to persuade young women to practice modern contraceptive methods. Women with greater wealth were more likely to obtain abortions services simply, because they have better access to financial resource and to better health care. Women who gave one or more births were more likely to seek abortion may be because they did not wish to have more children and wish to have space for more personal development (Kirkman et al., 2009;World Bank, 1993). Total fertility rate in Nepal was substantially decreased down to 2.0 in 2021 from 3.9 in 2000and 2.5 in 2010(United Nations, 2022a. Apparently, factors other than contraceptive use and abortion, such as socioeconomic developmental, are attributable to such fertility decline (Anderson & Kohler 2015;Caldwell et al., 1992;and Jain & Ross, 2012). Women in Madhesh province were least likely to seek an abortion reflecting lowest human development in the province (NPC & UNDP, 2020).
In the context of Nepal, these results may imply improvement and empowerment of women which enables them to decide when to seek an abortion. The finding that women with one or more children ever born are likely to obtain an abortion could be because the women who have been giving births to increasing number of children ever born are more likely ever to have been pregnant than women who have had no birth. In the abortion model, women who had one or more children ever born were more likely to have an abortion than those who did not give any birth. Furthermore, it could be because women's desired family size has become 2 or less, many women opt https://doi.org/10.36922/ijps.v7i1.291 International Journal of Population Studies for permanent method of family planning and if accessing such service is difficult, they choose to go for medical abortion which is free and relatively easy to access. Indeed, according to the Ministry of Health of Nepal, of the total abortions in the 5 years preceding the 2016 survey, 72% were medical abortions (MOH et al., 2017).
Nevertheless, the NDHS 2016 data show nearly half of all abortions among women aged 15 -24 were taken place in unauthorized facilities such as at home, which is likely higher than the world average (Shah & Ahman, 2012). This is because Nepalese women buy medical abortion drugs from private pharmacies and use them at home and some women get abortion service at private pharmacies not authorized by government to perform abortion (Gorkhapatra, September 5, 2019;The Himalayan Times, March 7, 2022). This proportion is alarmingly high. With the prime objective of reducing maternal mortality, the government of Nepal legalized abortion in 2002. In 2004, the government announced a policy of providing safe abortion services through the extensive health infrastructure existing in the country. In principle, such private facilities are not legally allowed to sell abortion drugs, but still this takes place. Journalists report about illegal sale of abortion drugs and as per the government law illegal sellers can be subject to 3 years imprisonment and/or charged with fines but the situation has not improved (Gorkhapatra, October 9, 2020). However, until mid-2020, only about 1,500 health facilities in the country had provided safe abortion service out of about 4,400 facilities (Gorkhapatra, October 7, 2020). It is also argued that distance to a health facility is a major problem of accessing health care including abortion service in Nepal (MOH et al., 2017). Besides, the facilities that are authorized to perform safe abortion service are not functional all the time, because they frequently face the absence of service providers and/or the shortage of supply of necessary drugs, and it is common that post-abortion contraceptives and equipment are interrupted. In addition, there are a number of social, normative, economic, and distance barriers for women to access safe abortion services. Nepalese is a patriarchal society and the NDHS 2016 data show that only 23% currently married women make decisions on their own about their own health care (MOH et al., 2017). Last, but not least, due to cultural values and norms of son preference prevailing in Nepal (Karki, 1988), there is also evidence that sex selective abortion is on the rise after it was legalized (Frost et al., 2013), although there are strict laws that prohibit such a practice (MoH, 2016). All these factors could make women to go to private unauthorized health-care facilities which are usually close by. The situation was likely worsened due to the COVID-19 lockdown. Relatively more Nepalese women ended up with unwanted pregnancies during the lockdown, because contraceptives were in short supply, and many service centers were closed during the lockdown period, because abortion service providers were overwhelmed by COVID-19 pandemic, compelling women to seek abortion service from unauthorized private facilities (The Himalayan Times, 15 July 2021).
Fortunately, the analysis of the project data shows that access to safe abortion services has been improved even during the COVID-19 pandemic and lockdown. This is because no woman in the project was prevented from accessing safe abortion service, while, in other parts of the country, more women were utilizing unsafe abortion and putting their life at high risk. In the Gorkha project catchment area, there are 18 health facilities that provide safe medical abortion service. This indicates one facility per about 1,000 women of reproductive age, which is advantageous compared to the national rate of about 5,000 women per facility (MOH, 2021). The beauty of the project is that women in highly disadvantaged Dalit, in relative terms, were utilizing the safe abortion service made available in close-by facilities more than their advantaged counterparts. Furthermore, post-abortion family planning is much higher (87%) in the Gorkha project area than the national rate of 25% (MOH et al., 2017). Admittedly, the findings in the Gorkha field monitoring data, and those from Dalit community in particular, were not sufficiently captured in the NDHS of 2016. To understand the complexity of sexual behavior and abortion among the young age population, more research aiming to capture large sample using both qualitative and quantitative approaches are needed. Furthermore, more investment is needed in that as evidenced from the data that even during difficult times of COVID-19 pandemic many women could access the safe abortion service available nearby.
In sum, findings of this study indicate that the sexual and reproductive health and rights (SRHR) programs including family planning and safe abortion programs in Nepal have not been satisfactory. For this, not only the family planning and safe abortion programs need to be strengthened but also the local Female Community Health Volunteers (FCHVs) have to be encouraged to do their jobs in communities better. The FCHVs should also be asked to hold community meetings with local women and young girls to spread the knowledge about effective use of modern contraceptive methods, unreliability of traditional contraceptive methods, legal status of abortion, and availability of free abortion service. The health infrastructure such as Health Posts, Primary Health Care Centers, local, district, and other hospitals must ensure that the contraceptives and medical abortion drugs are never in short supply. Family planning program was https://doi.org/10.36922/ijps.v7i1.291 International Journal of Population Studies introduced in the country more than 50 years ago, and yet, it appears that the program needs an overhaul to increase its efficacy. Furthermore, abortion was legalized and safe abortion service was introduced for over 12 years, and yet, it is still far from providing extensive safe and quality abortion service for common women and girls to access it.

Limitations
This study has several limitations. The NDHS data that have been used here are cross-sectional in nature which does not permit to establish causal relationships between contraceptive use and abortions. The NDHS data that are regularly collected in Nepal since 1996 are meant for monitoring progress made in health and demographic dynamics in every 5 years. The issues explored in the surveys do not probe for in-depth understanding of such a causal relationship, although they are good for program managers and policy makers. For instance, we have learned from the data that a certain proportion of women had abortion in the recent past, but we do not have information with respect to what made her to go for abortion nor we know how she felt about the service or how she felt herself after the service.
The monitoring data used in this analysis have limitations too. Because the government categorizes abortion facility data as highly sensitive and only a few characteristics of abortion clients are collected, it is not possible to understand deeply about the way a woman decides to opt for abortion service at a facility. Nevertheless, project monitoring data contain a few basic background characteristics in numerical forms which are helpful to make quantitative analysis.
In addition, the number of young women who had an abortion in the 5 years preceding the survey in this study was only 83 and when they were further divided into two groups the analyzed sample number became even smaller. It is thus important that the interpretations presented here must therefore be taken cautiously. Abortion is highly stigmatized in Nepalese society and people still believe that abortion is a sin (Karki & Magar, 2018). In such a society, collecting information on abortion from a faceto-face interview in a questionnaire survey is subject to underreporting. There is evidence that many women just do not report that they have had an abortion. On the basis of the total sample of 12,862 women aged 15 -49 in 2016 (MOH et al., 2017) and annual abortion rate of 42/1000 women of reproductive age 15 -49 in 2014 (Puri et al., 2016) of other data sources, approximately 2,700 abortion cases should be captured for the 5-year period preceding the 2016 NDHS survey. This suggests that only about 20% of abortions were reported in NDHS. As adolescents face severe social sanctions against premarital pregnancy , it is likely that many premarital pregnancies that ended in abortions were not captured in the survey. The analysis is based on cross-sectional data -abortions that took place in the past 5 years preceding the survey, which is subject to several limitations. In the absence of some important variables such as education, residence, and socioeconomic status at the time of abortion, the analysis was based on data collected at the time of the survey. As education and economic status change over time, they can impact the results that are presented here.

Conclusions
The 2016 Nepal Demographic and Health Survey (NDHS) supplemented by key indicators from several previous waves of NDHS shows that the prevalence of contraceptive use among women aged 15 -24 has been stagnant, and their prevalence of unsafe abortion is high in the last two decades in Nepal. The study also investigated the factors associated with the prevalence of contraceptive use and the prevalence of abortion based on multivariable regression models for some major sociodemographic characteristics. It, furthers, analyzes the distributional difference in unsafe abortion by major factors, which shows that different sociodemographic factors played varied roles in affecting Nepalese young women's choice of contraceptive use and abortion behaviors. Overall, the sexual and reproductive health and rights (SRHR) programs including family planning and safe abortion programs in Nepal have not been very effective in the past two decades. Family planning and safe abortion programs need to be strengthened to achieve sustainable development goals. The Gorkha project demonstrated provided a good example that access to contraceptive and abortion services among disadvantaged populations could be largely improved when available services are provided to them. Post-abortion family planning services could also increase safe abortion. To understand the complexity of sexual behavior and abortion among the young age women and to achieve the sustainable development goals for reduction in maternal mortality risk in Nepal, more research aiming to capture large sample using both qualitative and quantitative approaches are needed.