Introduction
After decades of strict anti-abortion laws, abortion was legalized in Nepal in March 2002, marking a watershed in the history of women's rights. The path to this policy change spanned over 30 years. An overview of the history of the reform movement and salient factors that contributed to the reform has been provided elsewhere (Thapa, Reference Thapa2004). In brief, reform efforts began in the 1970s and initially grew out of government concern over population growth. Ultimately, the overriding rationale for legalizing abortion was to reduce maternal morbidity and mortality and to ensure women's rights. Nepal's participation in the global Safe Motherhood initiative and its ratification of a global women's rights declaration helped to set the political stage for changing the law. Research showing high rates of maternal death from unsafe abortions was also influential. The restoration of democracy in 1990 enabled civil society to play an important role in advocating for reform. Greater freedom of the press and deregulation of the media allowed the media to contribute substantially to the reform process through disseminating information on abortion and women's rights. The process of reform was also characterized by an absence of major opposition.
The new law permits termination of an unwanted pregnancy without restriction as to reason, thus placing Nepal, as of 2008, among 17 developing and 33 developed countries where abortion is legal (Singh et al., Reference Singh, Wulf, Hussain, Bankole and Sedgh2009). The 2003 national safe abortion policy (MoH, 2003) emphasized the need to create awareness of the new abortion law and educate women about the specific conditions under which abortion is permitted. It called for various forms of media – electronic, print and interpersonal communication – in both the public and private sectors to be mobilized for these purposes and to inform women of facilities offering safe abortion services.
Beginning in 2004, abortion services were introduced and expanded through public, non-governmental, as well as private clinics. By mid-2011, nearly 500,000 women had obtained abortion services from 487 Ministry of Health and Population-certified abortion clinics in the public, non-governmental and private sectors (www.ipas.org/countries/Nepal). A total of 1276 doctors and nurses have been trained in performing safe abortions. Along with the establishment of service facilities, efforts were made to create awareness of the new law and to provide women with information on where they could access services. These were combined efforts of international organizations and national non-governmental organizations in collaboration with the Ministry of Health (later renamed as the Ministry of Health and Population or MoHP).
As of 2006, 32.3% of women aged 15–44 were aware of the legal status of abortion and 56.5% knew of a place in Nepal where they could obtain an abortion (Thapa & Sharma, Reference Thapa and Sharma2012). The results revealed significant disparities among population sub-groups in both awareness of the legal status of abortion and knowledge of a place for obtaining abortion services in Nepal. The 2006 Nepal Demographic and Health Survey (NDHS) was the first survey to collect information relating to abortion nationally. It thus provides a baseline for assessing progress over time. Five years later, in 2011, a similar survey was conducted that included additional questions relating to abortion. The 2011 round of the survey affords the opportunity to evaluate changes since 2006.
In this paper, the progress made in women's awareness of the legal status of abortion and availability of abortion services is assessed. Three interrelated questions are examined: What are the levels of knowledge of the legal status of abortion and availability of services among various sub-groups of women? Did these two measures improve between 2006 and 2011? Have the disparities between population sub-groups revealed by the 2006 survey diminished over time?
Data and Methods
The data analysed were from the 2011 Nepal Demographic and Health Survey (NDHS), part of the global Demographic and Health Surveys (www.measuredhs.com). The eligible respondents for the survey were defined as women aged 15–49 and the data were collected in the first half of 2011. A total of 12,674 women were successfully interviewed, yielding a response rate of 98.1% (MoHP et al., 2012).
For the present analysis, the study population was defined as all women (regardless of their marital status) aged 15–44. The age group 45–49 was excluded because of relatively few abortion cases and also to make the results comparable with the previous analysis. Two outcome measures were included: (1) whether the respondent is aware that abortion is legal in the country, and (2) whether she knows of a place where services can be obtained. ‘Uncertain or don't know’ responses, if any, were included in the ‘no’ category of the corresponding variables. Thus, each outcome measure had a binary outcome (yes or no). The denominator for the two measures refers to all women aged 15–44 at the time of the survey (N=11,727).
Six independent variables were included in the analysis: ecological and development sub-region, type of residence, education, household wealth quintile, age and number of living children. These variables represent ecological- (structural) as well as individual-level characteristics of the respondents. Ecologically, the country is divided into three horizontal belts – Mountain, Hill and the Terai (the southern, sub-tropical belt adjoining India) – and five vertically divided development zones. These are further divided into 75 districts based largely on population density. Because of relatively low population density, the Mountain belt has fewer districts than the other two belts. In the NDHS sampling, the less densely populated regions were combined and a total of thirteen (instead of fifteen) ecological development regions were created.
Age and parity represent a woman's life-stage. The variable ‘household wealth’ is a composite measure of the cumulative living standard of a household (including asset items such as water and sanitation facilities, televisions and type of material used for flooring). The construction of the index, based on Principal Component Analysis, is described in detail elsewhere (Rutstein & Johnson, Reference Rutstein and Johnson2004). Wealth status is a proxy measure of an economic enabling factor. Women's education is a measure of their acquisition of information and knowledge (Martin & Juarez, Reference Martin and Juarez1995).
Bivariate analysis was carried out first, followed by multivariate analysis. For the latter, logistic regression was used since both the outcome measures are binary (Retherford & Choe, Reference Retherford and Choe1993). In order to avoid bias towards the over-sampled sub-populations, sample weights were applied for all estimates (means, percentages and regression coefficients). The standard errors were adjusted for design effect due to multi-stage cluster sampling.
Results
Awareness of the legal status of abortion
Overall, 38.7% (95% CI: 37.8, 39.6) of women were aware of the legal status of abortion (Table 1). At the sub-regional level, the proportion varied from 29% in Central Terai to nearly 60% in Far-Western Terai. Awareness was considerably higher among those living in the urban areas. Along economic lines, awareness increased from 22% of women in the poorest stratum to nearly 55% in the richest stratum. Similarly, it increased from a low of 20% among those with no education to a high of 69% among those with high school or higher educational attainment. Awareness was higher among those in the 20–29 age group, and it was slightly higher among women of parity 1 and 2.
Percentage of women (aged 15–44) who are aware of the legality of abortion and odds ratio (OR) of awareness based on logistic regression, Nepal, 2011
***p<0.001; **p<0.01; *p<0.05.
Note: p-values referring to percentages are based on chi-squared test. CI=confidence interval.
Table 1 shows odds ratios (ORs) based on multivariate logistic regression. Three of the thirteen sub-regions – Western Mountain, Far-Western Hill and Far-Western Terai – showed an OR of over 2 compared with the reference sub-region, Eastern Mountain. Women in the fourth and fifth wealth quintiles were significantly more likely to be aware than those in the poorest quintile. Similarly, women with secondary or higher level of educational attainment were significantly more likely to have a higher level of awareness than those with no education. Women aged 30 and higher were more likely to be aware than younger women. Women of parity 1–3 had higher awareness than those of 0 or 4+ parity.
The survey also asked women an open-ended question (allowing for multiple responses) about conditions under which a woman could have a legal abortion. Of all the women who were aware of the legality of abortion, about two-fifths (38.3%) did not know of specific conditions; 35.5% thought it was legal for any woman with an unintended pregnancy of up to 12 weeks; 20.8% thought it was legal for up to 18 weeks if the pregnancy resulted from rape or incest; 8.6% thought it was legal if a pregnant woman's life was in danger; 9.1% thought it was legal if a mother's physical or mental health were at risk; 7% understood fetus abnormality to be a condition under which abortion was legal; and 8.8% gave other answers.
Knowledge of a place for services
Women's knowledge of a place to obtain an abortion was much higher than their awareness of the legality of abortion. Overall, 59.8% (95% CI: 58.9, 60.7) reported having knowledge of a place (Table 2). The correlation between awareness of the legality and knowledge of a place was 0.3802 (p<0.001). Across the sub-regions the percentage ranged from a low of 40% to as high as 74%. Only 41% of women in the poorest stratum had knowledge of a place, in contrast to 72% of women in the richest group. The disparity was even higher by education, ranging from 49% to 82%. Knowledge of a place was higher among women aged 25–34 and those of parity 1 and 2.
Percentage of women (aged 15–44) who know of a place to obtain an abortion and odds ratio (OR) of knowledge of a place, Nepal, 2011
***p<0.001; **p<0.01; *p<0.05.
Note: p-values referring to the percentages are based on chi-squared test. CI=confidence interval.
The ORs, based on multivariate logistic regression, indicate that sub-region and urban–rural residence are relatively less important than other variables. The odds of having knowledge of a place were significantly higher for the richer groups (fourth and fifth quintiles) than the poorer groups. It was also very strong among those with a secondary or higher level of education. Women aged 25–39 had significantly higher ORs compared with those in the 15–19 age group. Women of parity 1 and higher were significantly more likely to have knowledge of a place than those with 0 parity.
The survey also asked and probed women about specific types of places where abortion services may be available. Of the women who knew of a place, 71% mentioned various types of governmental facilities; 30% mentioned non-governmental organizations; and 58% mentioned the private/commercial sector (because this was a multiple response question, the total does not add up to 100%).
In the multivariate analysis, the higher influence of urban residence attenuated for both the outcome measures. The direction even reversed yet remained significant, particularly for the awareness of legality outcome. More detailed analysis showed that this seemingly paradoxical result was due primarily to the confounding relationship between urban residence and the wealth status variable. That is, those in the higher quintiles came disproportionately from urban areas and the analysis stratified by wealth quintile suggested that there was no urban–rural difference in knowledge of the legal status of abortion or of a place to obtain services. However, in order to make comparisons with the 2006 survey results, it was decided to retain the urban–rural variable in the model. Fairly strong collinearity between age and parity was also found. This relationship was the reason for attenuated effects of parity on awareness (in Table 1). Once the age variable was omitted in the analysis, the influence of parity remained strong and highly significant.
In a further analysis, the region variable was reconstituted into six categories: (i) Eastern Hill, Central Hill and Eastern Mountain, (ii) Far-Western Hill, Mid-West Hill and Western Mountain, (iii) Western Hill, (iv) Eastern Terai and Central Terai, (v) Western Terai, and (vi) Far-Western Terai and Mid-Western Terai. Category (i) was used as the reference category and the regression model was re-run. The results showed that OR values for awareness were significantly higher (1.6 or higher) for the reconstituted (ii) and (iv) sub-regions. The OR values were lower (in the order of 0.6 or 0.7) for the other sub-regions. Similarly, for the outcome measure of knowledge of a place to obtain an abortion, (iii) was the only sub-region that had lower OR (0.5) than the reference category. The rest of the sub-regions had OR of over 2.0 for each.
In a separate analysis, the respondent's current occupation was included, categorized into (a) household work only, (b) agricultural or other manual work, and (c) office or other service-related work in the two regressions shown in Tables 1 and 2. The results showed that the respondents who worked in office or service-related occupation had a significantly higher OR (1.3) in both the outcome measures. The inclusion of this variable did not, however, substantially change the results for other variables as shown in the two tables. (The results obtain from the additional analyses are not shown in a table.)
Changes between 2006 and 2011
Table 3 shows change, in percentage points, between 2006 and 2011 in the two outcome measures. In the majority (nine of thirteen) of the sub-regions there was considerable increase in women's awareness of the legality of abortion. The Far-Western Terai and Central Mountain regions showed the highest increase. Gains in urban and rural areas were about equal. Women in higher economic strata and women with higher levels of education registered large increases. Conversely, those in lower economic strata and those with lower educational attainment had either no or only modest gains. Gains in the various age groups were somewhat mixed. Increases were especially higher for women with parity 1 and 2 than other groups.
Change (percentage points) between 2006 and 2011 in awareness of legality of abortion and place to obtain an abortion among women aged 15–44, Nepal, 2011
* Indicates that the mean difference in percentages between the two time periods is significant (based on 95% CI).
Note: Percentage values for 2006 are given in Thapa & Sharma (2012).
Regarding changes in women's knowledge of a place for obtaining an abortion, only five sub-regions showed large increases. Western Terai had the highest (25 percentage points) increase. In contrast, Western Hill showed a 17 percentage point decline, although the change was not statistically significant. Rural women showed an increase whereas urban women did not. Among women in different wealth quintiles, those in the third and fourth showed large increases between the two time periods. No significant changes took place among women in the various educational sub-groups. The percentage increased substantially among women aged 25–39. Some gains were also observed among women by specific parities; however, the changes were not statistically significant.
Discussion
This analysis used simple binary measures to assess women's knowledge of the new abortion law and of a place to obtain abortion services in Nepal. An affirmative response to the question on knowledge of legalization was followed by a question on knowledge of specific conditions under which the law permits women to have an abortion. The results showed that about two out of five women did not know of specific conditions. Clearly, a considerable proportion of women still lacked in-depth knowledge of the abortion law. Future programme interventions need to focus on improving both basic awareness of the legality of abortion and knowledge of the circumstances under which it is allowed.
Regarding the survey question on women's knowledge of a place for obtaining an abortion, a respondent who gave an affirmative answer was then asked if she knew of a place where a woman can get a safe abortion. She was then asked to name the place. Some women may not have known the actual name of the place. Moreover, simply naming the place did not indicate whether it was indeed a facility certified by the MoHP as a provider of safe abortion services. Many of the women may not have known or may have had inaccurate information about a facility's certification status. The MoHP has introduced a logo to indicate certification so that women can be readily informed as to whether or not a facility offers safe services. However, the logo is not always displayed in a way that it can be easily recognized. Furthermore, not all certified facilities have displayed it.
For these reasons, the responses most likely included any type of place that the women had knowledge of, whether or not MoHP-certified. Further, while the legality of abortion is relatively new in Nepal, abortion services are not, since it was possible to obtain an abortion even when doing so was illegal (e.g. Thapa et al., Reference Thapa, Thapa and Shrestha1992; Thapa, Reference Thapa2004). Therefore, the finding that knowledge of where to obtain an abortion was more widespread (1.5 times higher) than knowledge of the legal status of abortion should not be taken as inconsistent or implausible. The measure of knowledge of a place warrants further refinement and the measure, as used in this analysis, should be considered a crude indicator at best. Future efforts to educate women about the logo and its importance may prove effective not only in promoting safe delivery sites, but also in reducing ambiguity about which facilities are considered safe.
It should also be noted that both of the knowledge indicators showed some negative changes between 2006 and 2011. The negative changes were proportionately more for the knowledge-of-a-place indicator than the knowledge-of-legalization indicator. Most of the negative changes in specific categories of a variable are within the error term and therefore do not indicate significant differences between the two surveys. However, in one category – the first quintile – of the wealth index variable the negative changes in both the outcome measures were significant. The reasons for this are unknown.
Awareness of the new abortion law increased across many sub-groups of women of reproductive age. Many of the districts, particularly in the Mountain and Hill regions of the Mid- and Far-Western regions of the country, had long been bastions of neglect and poverty, resulting in extreme disparities (Thapa, Reference Thapa1995). It is only in the last decade or so that these regions have begun to receive attention (United Nations Development Programme, 2004, 2009). While the data suggest that the attention is yielding positive results, more intensive efforts are needed to raise the absolute level of awareness and knowledge.
It is also clear that the improvements over the last several years were limited mainly to women in higher educational sub-groups and higher wealth quintiles. Awareness does not appear to have increased in the less educated and poorer sub-groups of women, implying that the disparity has not decreased over time. Assessing progress made in reducing this disparity remains an important agenda item for future research. In addition, since the present analysis is limited to knowledge measures, it will be important to examine whether or not the strong correlation between the economic enabling factor (wealth status) and knowledge of legality and availability of services found in this analysis also holds for women's actual use of services.
Overall, while the results indicate that there has been some progress since 2006, large numbers of women still do not know that abortion is now legal, the specific contents of the law, or where they can access safe abortion services. Consequently there continues to be a strong need for educational efforts in all three of these areas, especially in relation to the most educationally and economically disadvantaged women. Nepal has witnessed a sea change in communication modalities in the last two decades (Thapa & Mishra, Reference Thapa and Mishra2003), and the expansion offers several new possibilities for delivering locally or regionally tailored messages to women about the abortion law and safe services. Since democratization, the state-controlled mass media has been liberalized and the private/commercial sectors are playing an increasing role in broadening access to information. Newspapers and magazines have mushroomed, although concentrated in cities and large towns, and FM radio and TV stations have grown rapidly. Internet service is proliferating, and mobile phone service is increasingly penetrating remote areas.
Formulating more effective strategies for reaching women will require a full appraisal of the communication landscape as to what currently exists, what means are most accessible to and preferred by women, and what methods of sharing information and knowledge are most effective in a given district or community. Evidence suggests that while radio and TV are important pathways for delivering health messages, reaching all groups of women will require using other avenues of communication as well. As of 2011, in rural Nepal, where 85% of the population lives, only about two in five women of ages 15–49 reported watching TV at least once a week and slightly more listened to the radio at least once a week (MoHP, 2012). In urban Nepal, an overwhelming majority (80%) viewed TV, while only about 50% of women reported listening to the radio at least once a week. FM radio was preferred by women more than twice over the national radio, and TV was preferred over other media sources. It will be important for those strategies that do focus on radio and TV to take into account such differences in usage and preferences.
In a recent process evaluation undertaken in two relatively remote districts of Nepal it was found that married women generally preferred street drama or other locally organized activities (such as documentaries and video showings), which are widely accessible to community members and enable them to talk with one another about the theme(s) of the event (Phuyal, Reference Phuyal2013). There exist many other platforms, such as women's groups, micro-credit programmes, non-formal education groups and community paralegal and citizen forums, where new policies and programmes can be collectively discussed. Undoubtedly, these are more intensive and expensive interventions than radio or TV programmes. However, much more can and should be done through inter-personal communication approaches to more effectively inform women about the new abortion law and the availability of safe services.
Evidence also points to the significance of marital status in relation to media access. Research has shown that in urban Nepal married women are at a significantly greater disadvantage with respect to exposure to media, including radio, TV and newspapers, compared with unmarried women (Thapa & Mishra, Reference Thapa and Mishra2003). This is surmised to be due to differences in socialization between the married and unmarried and to time availability, particularly for married women. The gulf between married and unmarried can be much worse in rural Nepal. It is equally important to understand the ways in which women (and men) get information in their communities. Friends, peers, neighbours and relatives are often the primary sources of information. Due to the availability of multiple forms of media, including social media, the challenge of accessing information is most likely less in urban areas than in rural communities.
It should also be emphasized that the nationally formulated strategies provide only a macro framework within which to design and implement programmes. It will be important to supplement the national strategies, preferably with a region- or district-specific, bottom-up approach to designing, implementing and evaluating communication strategies around abortion law and services (as in other health policy and service areas). Reaching the most disadvantaged segments of the population will most likely warrant intensive and targeted interventions. Supplementing national strategies in this way would also be important in light of Nepal's goal of achieving a decentralized system of governance. Recently, and for the first time, Nepal's MoHP has included abortion as a component of the national communication strategy in the overarching areas of maternal, newborn, and child health (MoHP, 2011). This consolidated and integrated approach within the larger maternal health context is expected to provide further impetus to expand the educational efforts already underway.
Conclusion
As of mid-2011, about two in five women of reproductive age (15–44) in Nepal were aware of the legal status of abortion and three in five knew of a place to have an abortion. Of those who were aware, a considerable proportion did not know of the specific conditions under which a woman has the right to seek termination of an unintended pregnancy. The percentages of women who were aware of the law and who knew of a place to obtain services varied considerably by population sub-group. Awareness of legality and knowledge of a place increased by 6.4 and 3.3 percentage points, respectively, during the 5-year period. However, the increases in both measures were largely limited to women in higher wealth quintiles and those with higher educational attainment. The results suggest the need to intensify efforts to educate women, particularly the most disadvantaged women, about the abortion law, including the conditions under which abortion is permitted, and where to access safe abortion services.
Acknowledgments
Ipas/Nepal provided partial support for this research. The authors thank Dr Indira Basnett and Leah Retherford for helpful comments and suggestions on an earlier version of the paper. The views expressed are those of the authors only and do not necessarily reflect the views of Ipas/Nepal or other organizations with which the authors are affiliated. The study results were presented and programmatic implications discussed at a symposium held in Kathmandu in December 2012. The authors thank the Ministry of Health and Population and Ipas/Nepal for organizing the symposium and the participants for their comments and insights.
