The implementation of safe abortion services in Ethiopia

In 2005, a new criminal code was established to align Ethiopia's laws with its new Constitution. Following a period of intense activism and debate, abortion remained criminalized, but several significant exceptions were made, allowing for the expansion and integration of services within the public health system. The passage of the law and the establishment of technical guidelines each served as essential steps in determining the extent to which services were implemented. The integration of safe abortion services expanded the scope of practice for multiple cadres of healthcare providers, including emergency surgical officers, nurses, and health extension workers. The political will of the Ministry of Health, the research produced by the Ethiopian Society of Obstetricians and Gynecologists, and the expertise of nongovernmental organizations were essential to the implementation of services.


METHODOLOGY FOR ALL CASE STUDIES
This case study is one of six comprising a comparative examination of varied countries' approaches to the implementation of national abortion service programs, after changes in laws or policy guidelines that established or expanded access to services. In addition to Ethiopia, case studies were conducted in Colombia, Ghana, Portugal, South Africa, and Uruguay, as they had all either implemented new abortion laws or policies, or changed interpretations of existing laws or policies, within the past 15 years. Each study used the Integrated Promoting Action on Research Implementation in Health Services (i-PARIHS) framework to organize the analyses. i-PARIHS posits successful implementation to be a function of the innovation to be implemented and its intended recipients in their specific context, with facilitation as the "active ingredient" aligning innovation and recipient. 1 For each country case, two types of data sources were used: an in-depth desk review and 8-13 semistructured, indepth interviews with key stakeholders and experts in each country, selected in collaboration with in-country partners. Respondents provided written informed consent and were guaranteed confidentiality. Several respondents from each country served as in-country coauthors, in doing so giving up their anonymity as participants of the study, although no quotations provided as respondents are directly attributed to them. Respondents included healthcare providers, public health and government officials who had been involved in establishing or expanding the service, academics, and members of nongovernmental organizations (NGOs) and legal and feminist advocacy groups; in some countries, interviewees came from the full range listed, in others, from a subset (Table 1). Interviews were conducted in English by a physician member of the team. Quotes presented are from interviews without attribution as we promised confidentiality. Data analysis comprised a multistep iterative thematic analysis, with coding structured to follow the i-PARIHS framework. The WHO's Research Ethics Review Committee approved this study (protocol ID A65920). A full discussion of methodology can be found in Chavkin et al. 2

| CONTEXT
Abortion has been criminalized in Ethiopia since the first penal code of 1930. The law was amended in 1957 to include exceptions for "grave and permanent danger to life or health" with the approval of two physicians, which resulted in limited expansion of services. 3 However, a new constitution was ratified in 1994 enumerating the rights of women, including the establishment of equal rights, access to maternity leave, political participation, and "the right of access to family planning education, information and capacity". The high maternal mortality rate due to complications of unsafe abortion framed the conversation of legal reform. In 2005, Ethiopia's maternal mortality ratio, although declining, was 687 deaths per 100 000 live births, 7 with estimates at the time attributing 32% of maternal mortality to complications of unsafe abortion 8 ; later estimates would range between 19.6% 7 and 31%. 9 In this context, a movement developed across the medical, political, and women's rights communities to reduce maternal mortality and fulfil the rights enumerated by the new Constitution. Increasing access to contraceptive services, postabortion care, and safe abortion care was a critical component of this mission. 3 Ethiopia's healthcare system has expanded significantly following the political reforms of 1991. Healthcare policy is directed by the Ministry of Health, with government services administrated at a regional level and supplemented by nongovernmental and private providers. 10 A series of 5-year programs has been implemented to increase access to the healthcare system through the expansion of primary care and task sharing with new cadres of healthcare workers. 11,12

| INNOVATION
With concerns around the high rates of maternal mortality linked to unsafe abortion, changes in abortion law were primarily driven by public health arguments. One interviewee stated: Rights-based arguments also resonated and were used to complement the health-based justification. However, several interviewees felt that the rights-based argument carried less weight in the national debate and could become too politically charged to result in meaningful legislation. One interviewee described this decision: I know people will take it, will take this an extra mile and talk about the rights perspective. Yes, there is the rights perspective, but in a conservative society like ours…people will ask whose right are you talking about, you know?
With several components of reproductive health and women's rights regulated in the criminal code, revised legislation was needed to address multiple elements of reproductive health, including the provision of abortion services. 5 With new norms established in the Constitution, the Criminal Code of 1957 was no longer applicable, creating a policy window for significant legal reform. Some interviewees described the push to reform the criminal code as partly due to a sense that the old code was foreign, copied from European criminal codes, and unsuited for the Ethiopian context.
Interviewees reported that initial opposition to the revised criminal code was well organized by domestic and international advocacy groups, joined by religious organizations. Concerned about a potential backlash to the law, the drafting committee retained abortion in the criminal code with exemptions for specific indications. In exchange, it permitted pro-choice advocacy groups to have a strong negotiating presence in drafting the exemptions. One interviewee who advocated for legal reform described the process as follows: This opportunity allowed for collaboration among different sectors of the healthcare system. Another interviewee identified that "the coordination between a number of civil societies, professional associations, and then elements of the government that have really understood the issue we are trying to address has really helped us." This alliance created sufficient momentum to advance the issue in Parliament.
The revised criminal code came into effect in 2005. Abortion remained criminalized with the following exceptions: rape or incest, a T A B L E 1 Professional domains of interviewees in Ethiopia.

Medical practitioners 3
Government officials 1
a "Other" comprises academics, or individuals from feminist or legal advocacy groups, or UN agencies. risk to the life or health of the mother, fetal malformation, and maternal disability or age younger than 18 years. Two features of the law allowed for the creation of guidelines that expanded access to safe abortion services. First, the law was crafted to shift liability away from providers, who were permitted to accept that "the mere statement by the woman is adequate" when determining if the patient met the criteria for an abortion in the case of rape or incest. 6 Formalizing legal deference to the patient's statements created a regulatory framework in which implementing organizations could apply the exemption broadly. Second, the law empowered the Ministry to interpret the law and determine how broadly the exemptions could be applied by developing technical guidelines. No limits on gestational age or level of provider training were initially established in the statute. Instead, they were deferred to the Ministry for clarification. One interviewee stated that through the drafting process "we wanted to push for a more liberal law without specifying in it the specific language." Interviewees reported that after the process moved to the Ministry of Health, opposition groups did not participate in developing the technical guidelines.

| RECIPIENTS
Population-based mortality data and the publication of provider experiences shaped health providers' attitudes toward expanding access to safe abortion care. Initially, the medical community did not focus on abortion, as there was a misperception that abortion affected a small proportion of the population. One interviewee recalled that

| FACILITATION
Interviewees reported strong support within the Ministry of Health, which "owned a desire to expand access to safe abortion" and was the natural locus for implementing and normalizing a service framed in public health terms. Working within the one-system health framework, the commitment of the Ministry to the expansion of services was essential to increasing access. The government included NGOs in implementation, enabling them to work openly with government agencies, advocate for services, and address policy issues. When reflecting on the relationship between NGOs and the government, one interviewee stated that "we can talk very openly with the government, and the government actually really owns this and supports it." NGOs, in turn, participated in drafting the technical guidelines and advising the government on the regulatory framework. Interviewees representing NGOs perceived that timely government action had been contingent on continued NGO participation.
Interviewees reported that the drafting process for the technical guidelines was heavily influenced by international and peer-country models. 8 The WHO guidelines strongly influenced the drafting process and provided legitimacy for advocacy groups, as international standards were well received by the Ministry of Health. 18  Conservative choices were also made in drafting the guidelines to minimize adverse outcomes and internal conflict in the early stages of implementation. First, the recommended use of medical abortion was limited to less than 9 weeks' gestational age, reduced from the WHO recommendation of 12 weeks owing to insufficient in-country safety data. Second, there was insufficient data to develop a standardized protocol for second-trimester abortions in the first edition of the guidelines. One interviewee recalled how this process: By deferring the issue, the drafters were able to maintain support for the guidelines.
The technical guidelines were revised in 2014 following a revision to the WHO's guidelines as additional data were available both locally and internationally. 20 The use of medical abortion was expanded beyond 9 weeks, as additional data demonstrated the safety of the procedure at later gestational ages. Second-trimester abortions were Providers are screened prior to training, which is only offered to those who agreed to provide services. Conscientious objection is not permissible by regulation, as referrals are often impossible 21 : "We know that service provision centers are widely scattered. Now if a provider from a locality has to refer to another, then how would a woman go to a referred facility? It means we lose her." However, de facto objection does exist among providers who choose to defer training or complete training but decline to offer services. In keeping with an integrated healthcare system, no financial incentives were offered for those who provide abortion services. While some interviewees reported stigma in both the providers' professional and social communities, several interviewees reported that stigma has decreased as the role of abortion has been better understood in the years since that law was passed.
One interviewee reflected that: "I think people understood that having such interventions can save the lives of [women] so I think it's much better." One organization developed a system of meetings where providers could share and discuss their experiences.
Interviewees emphasized the role of values clarification at the facility level to introduce abortion services to the healthcare workforce. For second-trimester services, they underscored that including all members of the clinical and administrative staff was essential to ensuring that services were provided, as any employee of the clinic, including the security guard and the receptionist, could serve as a gatekeeper to care.
Surveillance in the public health system occurs through the national Health Management Information System. 22 The collection and digitization of data is a resource-intensive process that limits the number of health metrics that can be monitored. Fewer are followed than had been recommended in the technical guidelines. Interviewees reported concern that the centralized data are overly aggregated and inadequate to effectively monitor the provision of abortion services. Parallel monitoring systems exist in the private and NGO clinic networks; periodic studies combine retrospective and prospective methodologies and use both public and private data to fill the gaps in the public surveillance system. 16,17,[23][24][25][26] Interviewees stressed the value of publishing periodic assessments of national data to document the impact of the law.

| LESSONS LEARNED
Increased access to abortion services in Ethiopia is a consequence of a carefully crafted law and the persistent efforts of multiple actors across the healthcare system. A health-based justification to reduce maternal mortality galvanized support among medical providers and within the healthcare system. A stepwise approach to advocacy began with a strong research program to document the harms of unsafe abortion; this was necessary to build and maintain support. However, further amplifying a rights-based approach could have expanded and solidified ownership of political reform for women, girls, and human-rights advocates. The government's consideration of the concerns of opposition groups led to maintaining abortion within the criminal code. While the new law expanded the indications for legal abortion, the service remains unavailable for many women. The law explicitly placed responsibility for policy development in the Ministry of Health, and the development of technical guidelines was a key step in determining the scope of the law. The political will of the Ministry and regional administrators was essential in all phases of implementation, as was the strong and lasting alliance of supporting actors including NGOs, advocacy groups, and ESOG. The implementation of safe abortion services in Ethiopia has significantly increased access, advancing the obligation to uphold the rights and health of women.

AUTHOR CONTRIBUTIONS
DB-P: Contributed toward the initial proposal and interview instrument, conducted desk reviews for all case study countries, conducted Ethiopia interviews, coded transcripts, drafted the manuscript, and collaborated in reviewing and editing. SK: Advised on the interview instrument, served as in-country coordinator, interviewee, collaborated in writing, reviewing, and editing the manuscript.

CONFLICTS OF INTEREST
SK functioned as the in-country partner, was interviewed, and is coauthor of this case study. The authors have no conflicts of interest to declare.