Incidence of legal and illegal abortion and menstrual regulation
Three of the countries in our study have laws or policies that permit induced abortion (India and Nepal), or MR (Bangladesh). Procedures that meet each country’s legal criteria for these services (approved providers and facilities) account for a minority of all abortions in India and Nepal. This proportion was higher in Nepal (42%) than in India (an average of 22% among the six states, ranging from 11% in Uttar Pradesh to 32% in Tamil Nadu). In Bangladesh, of the combined total of MRs and induced abortions, MR procedures provided in facilities by trained providers accounted for 26%. Most abortions in Pakistan would be classified as unsafe.
Abortions that are not provided by approved providers vary in safety, depending on the provider and method used. Although comprehensive data are not available on type of provider and method, there is some evidence that increasing proportions of women use medication abortion outside of facilities, purchased from informal sources in India, Pakistan, and Nepal, marking a shift toward safer illegal abortion compared to the past (18–20). Using the new three-category spectrum of abortion safety (21), the majority of illegal abortions in these three countries are likely to fall into the middle category of “less safe” because they use an approved method (medication abortion) although the source of the method is the informal sector, and not legally approved. An unknown proportion of abortions would be considered “least safe,” those that use unsafe methods administered by untrained providers.
Treatment for abortion complications
In order to understand gaps in access to safe abortion services (or MR services in Bangladesh) and post-abortion care (PAC), we examine the extent to which women are being treated for abortion complications. The treatment rate—the number of women treated in facilities for induced abortion complications each year per 1000 women of reproductive age (15–49)—is a useful indicator of the incidence of abortion-related morbidity.b In countries where abortion is severely legally restricted, the treatment rate ranges typically between 5 and 10 per 1000 women (22).
Bangladesh and Nepal had the lowest treatment rates at 6 and 8 respectively, while Pakistan had a rate of 14. In India, the treatment rate varied by state. It was lowest in the states of Gujarat (4), Assam (6) and Tamil Nadu (7); intermediate in Bihar (11), and very high in Uttar Pradesh (21) and Madhya Pradesh (26) (23–28). The rates for the six states of India do, however, need to be interpreted with caution because additional data collected only for these states suggest that a large proportion of post-abortion patients presented with incomplete abortions related to use of medication abortion. Some, possibly a high proportion of these women, may not have needed treatment as their abortion was in progress and would have resolved safely without further intervention. Nevertheless, in all four countries, increased self-use of medication abortion is likely to have reduced the severity of abortion complications and reduced the need for treatment compared to the situation before widespread use of medication abortion. On the other hand, treatment rates usually underestimate abortion-related morbidity because some women who need treatment do not obtain it (2).
Access to abortion services
In order for women to receive legal induced abortion or MR services—relevant in three of the four study countries—access to properly capacitated health facilities is critical. Results show that provision of induced abortion services (or MR in Bangladesh) fall short of what is needed.
In Nepal and five of the six Indian states, between 50% and 71% of all facilities that provided any abortion service provided both induced abortions and PAC; however, in Bangladesh and Uttar Pradesh, the proportion providing both services was notably lower (just under 40%). A small proportion of facilities in all three countries provide only induced abortion (or MR in Bangladesh). Among Indian states, the lowest proportion of facilities providing induced abortion services was reported in Uttar Pradesh, where only 41% of facilities said they provided this service. A substantial proportion of facilities that offer some abortion care provide only PAC: this proportion ranges from 32% in Nepal to 61% in Bangladesh and has a similar range across five of the six states of India (ranging from 30% in Assam to 59% in Uttar Pradesh). In Tamil Nadu, this proportion is very low (11%).
Public sector access: A majority of facilities providing MR services in Bangladesh and induced abortions in Nepal were public-sector facilities (60% and 70% respectively). However, when we examined the proportion of public facilities with the mandate to provide such services that were actually providing these services, we found that only 38% of public facilities in Nepal provided induced abortions. This indicates that in Nepal, there is great potential to expand provision of abortion in the public sector.
Examining the Indian states for which we have data, in Assam, over half of facilities providing induced abortion services were in the public sector; however, in other states only a minority of facilities providing these services were in the public sector, ranging from 12% in Bihar to 23% in Madhya Pradesh. This indicates that the private sector has a dominant role in facility-based abortion service provision across the country in five of the six Indian study states. In a similar pattern to that found in Nepal, the proportion providing abortion services among all public sector facilities that are permitted to provide this service is very low across all six study states in India. Madhya Pradesh has the highest proportion of public sector facilities providing induced services (36%), while Bihar and Uttar Pradesh have the lowest proportions (11%).
The role of the public sector in the provision of PAC services is similar to that for provision of induced abortion services in Nepal and the six Indian study states. Public facilities were strongly represented among all PAC-providing facilities in Nepal, indicating the smaller role played by the private sector. However, as with induced abortion service provision, less than half of all public facilities in Nepal that have the potential to provide PAC, actually did so. Among five of the six Indian study states, public sector facilities were the minority among facilities that provide PAC (14% to 29%), and only in Assam were they a majority (60%). Of public facilities which are potentially able to provide PAC, the proportion that did so was a minority in five of the six study states, the exception being Madhya Pradesh, where the majority did so (62%). The situation in Bangladesh was somewhat different with respect to provision of PAC services: the public sector had a much smaller role in providing PAC services (32% of all PAC-providing facilities were public sector) while almost all public sector facilities that are permitted to provide PAC do provide this service.
Even in Pakistan, where the public sector is mandated to provide PAC, public sector facilities make up a minority (37%) of all facilities providing this service and only about a quarter of public-sector facilities that have the capacity to do so (26%) provided PAC services. Similar to the situation in India, this indicates that the private sector has an important role in providing both abortion and PAC services in Pakistan.
Rural access: The distribution of facilities providing induced abortion and PAC services between rural and urban areas is important for gauging access to abortion services in South Asia, since the majority of the population, including the majority of the population of reproductive-age women in all four South Asian countries, reside in rural areas (7–9,29).
In Bangladesh, 60% of MR-providing facilities are located in rural areas, while 72% of reproductive age women live in rural areas. Similarly, in Nepal, a majority of facilities providing induced abortions were located in rural areas (57%). In India, however, a minority of facilities providing induced abortion services were in rural areas, while the majority of women in five of the six states live in rural areas (7). In Tamil Nadu, although 49% of women live in rural areas, only 5% of facilities that provide abortion services are located in rural areas. These sharp imbalances in favor of urban areas indicate that in India, rural women have much poorer access to abortion services, compared to urban women.
The concentration of services in urban areas also exists in access to PAC services in the six Indian states, with a majority of PAC-providing facilities located in urban areas. The situation is similar in Pakistan and is even more extreme in the case of Bangladesh where 7% of facilities that provide PAC are located in rural areas but 72% of women are rural residents. In contrast, in Nepal, a higher proportion of PAC-providing facilities are in rural areas (60%) than the proportion of women living in rural areas (37%).
Turn away of women seeking abortion or MR services: This issue is relevant for three of the four study countries—with respect to abortion services in India and Nepal and MR services in Bangladesh. Even when women are able to find a facility approved for induced abortion service provision or for MR, access to induced abortion services is often hampered by facilities turning women away. Similar levels of women seeking an induced abortion were turned away in Nepal, Bangladesh, and four Indian states. The exceptions were the Indian states of Assam and Gujarat, where much lower proportions of women were turned away (6% and 8%, respectively).
Some of the reasons women are turned away are understandable. In Bihar, Uttar Pradesh, Madhya Pradesh and Nepal, between 28% and 56% of facilities turned away one or more women for suspicion of wanting sex-selective abortion, a practice that is illegal in India, and about 63% facilities in Nepal turned some women away because the pregnancy was above nine weeks gestation, which suggests that these are facilities that only offer medication abortion (data not shown). In Bangladesh, 97% of facility respondents reported that some women were turned away because they exceeded permitted LMP limits (data not shown), which is consistent with the relatively low LMP limits permitted for MR procedures. In Nepal, Bangladesh and Assam, 65% or more facilities cited unspecified medical reasons for turning women away, which may include the facilities lack of capacity to provide second trimester services; in other Indian states, this proportion was lower but still substantial, ranging from 19% in Madhya Pradesh to 28% in Bihar.
However, a substantial proportion of facilities reported turning away some women in the six Indian states who were otherwise eligible to receive abortion services, for reasons such as being too young, unmarried, or not having had any children. Among these states, this proportion ranged from a high of 54% in Assam to 22% in Tamil Nadu. About 35% of facilities in Bangladesh reported turning away women seeking MR for these types of reasons, while in Nepal over half of facilities reported turning away some women seeking abortion services for being too young.
Capacity of facilities to provide surgical and medication procedures
In Nepal and the six Indian study states, the large majority of facilities that provide induced abortion (including public facilities) provided medication abortion. In Bangladesh, however, provision of MRM (MR done using medications) was very low—21% of MR-providing facilities offered MRM, and this proportion was 14% among public facilities. This is likely due to the recent timing of adoption of MRM, which was authorized in late 2013, just months before the HFS was conducted.
We also assessed the capacity of facilities to provide surgical procedures. In Bangladesh, Nepal and four of the six Indian study states, more than three-quarters of all facilities that provide abortion services (MR in the case of Bangladesh) had the equipment and trained staff to provide vacuum aspiration (between 74% and 93%). In Tamil Nadu, however, fewer than half of all such facilities (public and private combined) had this capacity; nevertheless, it is notable that 83% of public facilities that offer abortion care in this state had trained staff and equipment to provide vacuum aspiration, suggesting that public facilities are more likely to provide abortions using vacuum aspiration than private facilities. In Uttar Pradesh, only 44% of public facilities that provide abortion services reported having the equipment and trained staff to provide vacuum aspiration. However, given that 54% of all facilities in the state that offer abortion services reported having equipment and trained staff to provide vacuum aspiration, it would appear that the private sector is better equipped and prepared than the public sector to provide vacuum aspiration procedures. Another factor that likely contributes to this differential is that private facilities are more motivated to provide surgical abortion than medical abortion, given that the high cost of MA pills.
Use of appropriate procedures: Quality of abortion care
The procedures used for induced abortion and post-abortion care are useful indicators of the quality of clinical care. WHO guidelines recommend the use of medication abortion or vacuum aspiration for first trimester abortions, while D&E is recommended only for second trimester procedures (30,31). D&C is considered to be an outdated and invasive technique and is no longer recommended by the WHO as an abortion procedure.
In order to assess the quality of induced abortion services, we compared data from India and Nepal only. This is because in Bangladesh MR is permitted only in the first trimester and the use of medication abortion was very new at the time of survey; and the topic is not applicable for Pakistan, where abortion is highly legally restricted. The data show that of all induced abortion procedures provided across the six Indian states and Nepal, over half were done by surgical methods. In Nepal, the overwhelming majority of surgical abortions used vacuum aspiration (92%; calculated from data in Table 7).c In contrast, the proportion of abortions using vacuum aspiration in the six Indian states was smaller, and use of invasive D&C or D&E procedures was much higher compared to Nepal.
In Tamil Nadu and Uttar Pradesh, about 64% of all surgical abortions used D&E or D&C,d and the proportion was also high in Bihar, Gujarat, and Madhya Pradesh (42–47%; calculated from data in Table 7) (23–28). The proportion of all facility-based abortions performed by D&E or D&C procedures range from 25% to 37% in the six Indian study states, providing a strong contrast with the much lower proportion of facility-based abortions that were second trimester (4%–13%, data not shown): This finding strongly suggests that D&E or D&C procedures were often used for first trimester abortions. Inappropriate use of invasive surgical methods appears to be common in the six states of India (20).
Between 41% and 45% of abortion procedures in health facilities in four of the six Indian study states and Nepal, used medication abortion. However, in Assam, only 13% of procedures were done by medication abortion, even though 84% of facilities that offer abortion services report that they provide medication abortion.
The procedures used for provision of PAC are harder to assess across countries, since the data are not entirely comparable. In Bangladesh and Nepal, the survey asked specifically about procedures used to treat patients with incomplete abortions, whereas in India and Pakistan, the data were obtained on procedures used to treat patients with the full range of post-abortion complications. Further, in Pakistan, information was obtained on the main procedure used (summing to 100%), and in India, data were obtained on all procedures and treatments that patients received, and some more than one type of procedure or treatment. Because of these data constraints, we have restricted our comparison of PAC care across countries to the proportion of procedures that use D&C or D&E. It is important to compare countries on this dimension since these are invasive techniques, and our examination of induced abortion procedures used in Nepal and India shows that these invasive methods are overused. In Nepal, D&C or D&E procedures were used in a small minority of cases to treat incomplete abortions (about 8%). In Bangladesh, 33% of incomplete abortions were treated using one of these two procedures.e D&C or D&E procedures were used in treating over half of women with post-abortion complications in Pakistan and the Indian states of Assam and Bihar. In Gujarat, Tamil Nadu and Uttar Pradesh, close to a third of patients with post-abortion complications were treated with D&C or D&E.f