How task‐sharing in abortion care became the norm in Sweden: A case study of historic and current determinants and events

Abstract We performed a country case study using thematic analysis of interviews and existing grey and published literature to identify facilitators and barriers to the implementation of midwife‐provided abortion care in Sweden. Identified facilitating factors were: (1) the historical role and high status of Swedish midwives; (2) Swedish research and development of medical abortion that enabled an enlarged clinical role for midwives; (3) collaborations between individual clinicians and researchers within the professional associations, and the autonomy of clinical units to implement changes in clinical practice; (4) a historic precedent of changes in abortion policy occurring without prior official or legal sanction; (5) a context of liberal abortion laws, secularity, gender equality, public support for abortion, trust in public institutions; and (6) an increasing global interest in task‐shifting to increase access and reduce costs. Identified barriers/risks were: (1) the lack of systems for monitoring and evaluation; and (2) a loss of physician competence in abortion care.


| INTRODUCTION
Task-sharing or shifting within abortion care from specialist physicians to nonspecialist physicians or nonphysician staff has, in some contexts, been associated with a reduction in abortion-related deaths. [1][2][3][4] In other contexts it has been used to increase access, cut costs, and increase continuity of care without decreasing quality of care. [5][6][7] The aim of the present review was to understand how the Swedish model of task-sharing in abortion care developed. We specifically wanted to know which factors facilitated or hindered its implementation and what impact the new model has had on access, quality, and perceptions of care.

| The Swedish healthcare system
Healthcare policy change in Sweden can be effected at the state or regional level. The role of the Swedish government and parliament is to establish the overall principles and guidelines of healthcare provision through laws and regulations, with technical support from public health institutions and the professional organizations. The 21 regional administrations are responsible for the provision of healthcare within their constituency and autonomously perform quality controls of the care provided. 8 The 20 regional and 40 subregional hospitals can further independently decide how to organize healthcare service delivery as long as they adhere to the mandate accorded to them by the regional administration. According to the Swedish abortion law, abortion care must be initiated either from a hospital or from a clinic approved for this purpose by the National Board of Health and Welfare (NBoHW). 9 Abortions are offered at approximately 130 gynecological departments and specialized public or private gynecological clinics but are not offered at primary-care level. 10 The taskshifting reform within abortion care in Sweden has primarily occurred at gynecological departments in regional or subregional hospitals.

| The Swedish model of task-sharing in abortion care
The current Swedish abortion legislature from 1975 (SFS 1974:595) allows for abortion until gestational week 18 weeks 0 days at the woman's own request. 9 In 2017, 37 000 induced abortions were performed (20 per 1000 women aged 15-44 years). 11 A high percentage of abortions were performed using medical methods (93%) and before the ninth week of pregnancy (84%). 11 In 2007, the Swedish Society of Obstetrics and Gynecology (SFOG) took on a mandate, suggested by NBoHW, to develop a formal training program for abortion care certification of midwives. The first midwife- The requirements and clinical applications of SFOG-certified abortion care provided by midwives in Sweden are summarized in Table 1.

| Data collection
We performed a country-specific case study of task-shifting in abortion care in Sweden. Data for the case study were assembled in two parts. The first part consisted of a literature review and the second part was a series of six in-depth interviews with key informants.

| Literature review
In Sweden many official documents are accessible to the public as online material. The NBoHW, the Legal Database of the Office of Government (Regeringskansliets Rättsdatabaser), the Swedish Agency for Health Technology Assessment and Assessment of Social Services (SBU), The State Public Reports (SOU), and Statistics Sweden (SCB) are all public institutions that, among other things, provide statistics or analysis related to abortion incidence, abortion legislature, and abortion policy. We searched these online archives for government and nongovernmental organization reports, literature on program strategy and program outcomes, process measures, and/or monitoring and evaluation plans relevant to the development of policies and practices of nonphysician provision of abortion care. We searched PubMed for studies reporting on task-shifting abortion care in Sweden. We performed a Google search to identify nonscientific reports, journalistic reporting, or opinion pieces related to task-shifting abortion care in Sweden.
We organized the content of the identified documents according to the following themes: (1) the present and historic role of midwives in healthcare service delivery in general and in abortion care; (2) the institutional processes that paved the way for midwife-provided abortion care; (3) quantitative and qualitative measures of attitudes to abortion; (4) the historic prevalence of unsafe abortion and consequences in morbidity and mortality; and (5) the history of abortion legislature, access, and practice in Sweden. The content of the assembled documentation was then synthesized in a literature review.

| Interviews
We performed individual in-depth interviews with six key informants who played an important role in the development of midwifery-led abortion care and were involved in implementing task-sharing. The informants were four registered nurse-midwives and two medical doctors with one or more current or past leadership positions in SFOG, SBF, NBoHW, the Swedish Association for Sexuality Education (RFSU), and/or the WHO Collaborating Centre at Karolinska Institutet, who performed the interviews primarily as representatives of these current or past held positions.
The informants were informed about the study and invited to participate by email. They provided consent for the content of their responses to be analyzed in terms of the aim of the study and to be quoted. The interviews were conducted in a private space with one T A B L E 1 Requirements for SFOG-certification in abortion care and clinical applications of midwife-provided care in Sweden.  (4) the relationship between guidelines and practice; (5) resources demands and opportunity costs; (6) healthcare workers' and women's experiences; and (7) monitoring/evaluation and research needs.

| Final data analysis
The text of the literature review and the notes from each interview were read through several times to identify content relating to barriers to and facilitators in the implementation of the task-sharing policy.
We then collated the relevant content from both these sources of data into themes. A theme was defined as an aspect or idea, presented once or repeatedly in the material, that we determined was of potential influence in the success or failure of the implementation of the task-sharing policy in the Swedish context. One researcher (ME) interpreted the data using qualitative content analysis. The content was analyzed on a manifest level (manifest analysis). Notes related to each theme were extracted and presented in a narrative summary.

| Overall assessment of the policy
All informants reported that task-sharing has had a positive impact on abortion care and they believed that implementation had been successful. In support of this, a Swedish clinical study cited in the literature review found that women are very satisfied with midwifeprovided abortion care. 6 Informants stated that the transition to midwife-provided abortion care required resources in terms of training but that this cost was counterweighed because midwife provision had proved more cost-efficient than physician-provided care. The

Midwife, Karolinska University Hospital
Although some concern was expressed that the workload for midwives would increase, several interviewees stressed the benefits of task-shifting for midwives, such as career development, increased salary, and involvement in research and education.

| Facilitators and barriers to implementation of the task-sharing policy
Eight main themes emerged through the data analysis. We identified six facilitators of the policy: the historic and contemporary role of the Swedish midwife; Swedish clinical research and the introduction of medical abortion; autonomy and role of the individual; flexibility of policy and legal interpretation; societal values (equality, secularity, and trust); and cost-saving and service delivery incentives. We identified two potential barriers or opportunity costs to implementation: loss of competence among physicians and lack of monitoring and evaluation.

| The historic and contemporary role of the Swedish midwife
Midwifery has long been a highly autonomous and respected pro-

Midwife, RFSU
Midwives' role in sexual and reproductive health care was first expanded in 1976 when they became certified to counsel and prescribe family planning methods and insert the copper intrauterine device. A second major shift in abortion care started with the introduction of medical abortion in the early 1990s, when midwives were increasingly delegated to perform the components of abortion care that conformed well to their professional role, such as contraceptive counselling, the administration of medication, supervision and pain relief during the procedure, and follow-up. 16

| Clinical research and the introduction of medical abortion
Sweden-based research in the 1980s and 90s was instrumental in developing the current medical abortion regimens for the first and second trimester through the conception of the combined method of a progesterone receptor modulator and a prostaglandin analogue. 17 More recent research has investigated the task-shifting process and simplification of abortion care. For example, several studies linked to Sweden have supported the feasibility of medical home abortion and self-assessment of abortion completion. [18][19][20][21][22][23][24][25][26][27] A clinical trial performed in Sweden concluded that early medical abortion provided by nurse-midwives compared with physicians was equally effective, safe, and acceptable to women. 6 A second study found that early medical abortion provided by nurse-midwives compared with physicians was more cost-effective. 12 The clinical trial showing the noninferiority of midwifeledabortioncarewasofpivotalimportanceinconvincing opponentstothepolicyofitsworthandgainingtheofficial sanctionofprofessionalorganizations.
MD, Linköping University, SFOG, and FARG Medical abortion methods were introduced early and today abortion care has transitioned almost completely to medical and self-administered methods. 11 The interviews revealed that the introduction of medical methods for abortions paved the way for an expanded role for midwives in abortion care. [

Midwife, SBF
The importance of research and the collaboration between research units and professional societies in the implementation process were also described by several interviewed informants as facilitating.  there has been no formal policy for expansion, and no direct financial support either from the government or other institutions.

| Flexible policy and legal interpretation
The historic review of abortion care in Sweden highlighted two important clinical transitions: the transition from illegal to legal abortions, and from physician-to midwife-provided care. Both transitions occurred in practice before they were formally sanctioned.
The abortion law of 1975 gave the legal prerequisites for a radical change in abortion policy that had already occurred in practice. During the 1930s an estimated 20 000 illegal abortions were performed each year, 75 women died, and many more were severely injured. 28 In 1974 however, a year before the more liberal abortion law came into force, the number of legal abortions had risen to 30 000. By that time, fourfifths of abortions were performed before week 12, and no illegal abortions were recorded. 28 What caused this reduction in illegal and unsafe abortions, within the confines of the existing restrictive law, was a consensus around a more liberal interpretation of the grounds on which abortion could be provided.
Task-sharing in abortion care was formally and legally investigated at a time when midwives had already gradually taken over most tasks related to abortion care (except for vaginal ultrasound). An investigation from 2008 suggested that delegating abortion care to midwives was an effective way of increasing access to abortion; it also concluded that in many clinical departments most abortion-related tasks were already being done by midwives. 29, 30 The board of SFOG finally assigned the task of developing a curriculum for midwife certification in abortion care to FARG within SFOG, which then went on to develop the current training curriculum together with the WHO Center at Karolinska Institutet.
The component of the abortion law, which states that a person with a medical degree must be responsible for the abortion procedure, has been questioned by the Swedish Society of Health Professionals (Vårdförbundet), SBF, RFSU, and some political parties. A formal motion to move the medical responsibility from physicians to midwives was put forward to parliament by the Green Party (Miljöpartiet) in 2010 and denied. 31 The response to the motion again concluded that most abortion-related tasks were already being done by midwives within the confines of the law. As the existing law was flexible enough to allow almost complete task-shifting to midwives, it was decided that physicians could retain overall responsibility and supervision of abortion care to not risk reducing the quality of care.

| Societal values: equality, secularity, trust
According Together with Finland and Iceland, Sweden in practice does not allow conscientious objection to abortion provision. A study of the implications of this legal framework indicated that disallowing conscientious objection facilitates good access to reproductive healthcare services by reducing barriers and delays to care and entails no negative impacts on providers. 33 According to studies that have evaluated perceptions of abortion care among Swedish providers, both gynecologists and midwives in general support the existing abortion legislature and do not hesitate to take part in abortion care even though they sometimes experience complex and difficult situations providing this care. The character of the work is experienced as contradictory but also gratifying. 16,34 Informants consistently answered that there have been no formal advocacy campaigns, no formal dissemination of information to civil society, and no media coverage of the new policy. To the question: "Has there been any backlash or negative response? From whom?
And what has been done to mitigate this?", most informants simply answered "No." Therehasbeennoneedtomitigateanynegativeresponse.

| Cost-saving and service delivery incentives
Several informants described trying to expand the role of midwives in abortion care during the late 1990s and early 2000s and being met with disinterest, as exhibited by the following quotes.

| Loss of competence among physicians
Several informants described that there was initial concern on the part of SFOG that junior doctors would lose competence in the area of abortion care and that this might become a cost to the program over

| Lack of monitoring and evaluation
Interviews showed that the quality or effectiveness of the task-

| DISCUSSION
The present case study identified six facilitating factors to the tasksharing reform in abortion care in Sweden and two potential opportunity costs. We assessed that task-sharing between physicians and midwives was a gradual and informal process, which was implemented without significant opposition owing to a combined effect of historical, cultural, individual, and structural factors.

| Interpretation
Our study shows that midwives have played a pivotal historic role in improving maternal health in Sweden, which may have facilitated their transition into abortion care. The consensus among interviewed informants that women perceived the policy as completely natural suggests that there were no barriers, from the public's perspective, to midwife-provided abortion care. Previous multicountry research on task-shifting in abortion care has shown that successful task-shifting is influenced by providers' willingness to provide care and their perceptions of their professional roles. In general, however, women's preferences seem to be guided more by trust, privacy, and ease of access to care than by the category of medical staff providing the care. 35 We found that the introduction of medical abortion into clinical abortion practice in Sweden occurred early through research and that this paved the way for an enlarged role for midwives in abortion care.
Research supporting the safety, effectiveness, and cost-efficiency of midwife-provided care was later important in creating a consensus among providers in favor of the policy. A study from the USA supports the role of research as a driving force behind successful policy reform in the area of task-shifting in abortion care. 36 We Our study indicates that earlier initiatives at policy reform, taken by midwives, were unsuccessful. This may suggest that the senior gynecologists who initiated the successful reform had greater leverage, but it is also probable that their initiative occurred at a time when more research existed in support of the reform, task-shifting had moved onto the global agenda, and service delivery concerns related to abortion were increasing. The implementation of task-sharing in abortion care also occurred in a context of a liberal and supportive abortion law, high public support for free access to abortion, low religiosity, high gender equality, and high trust in public institutions. All these factors may have facilitated the transition to midwife-provided abortion care.
Some of our informants expressed concern that task-shifting in

| Strengths and limitations
Task-sharing in abortion care in Sweden occurred upon the initiative of individual clinical departments and has expanded without systematic controls, monitoring, and evaluation. This weakens the ability of this study to systematically identify facilitators and barriers to implementation. Some clinical units may have chosen not to implement the policy because of barriers that would not have been captured in this review.
A strength of this study is the existing extensive background data owing to the easy access to current and historical public policy and legal documents in Sweden, as well as the long tradition of keeping registries on birth and death rates, abortion statistics, and causes of mortality. Furthermore, the process of task-sharing in abortion care was driven by a small number of individuals, many of whom are captured in this review. Interviews were not audio recorded but transcribed in note form and quotes verified retrospectively. One informant is also a coauthor of the article, but she did not take part in the data analysis.
In conclusion, task-shifting abortion care to midwives in Sweden was neither a regulatory nor a legal reform but occurred at facility level within the existing health and legal framework upon the initiative of clinicians and researchers within the medical professional societies and research units. Although the process was facilitated by factors specific to the Swedish context, this model for policy change may be applicable to other settings.

AUTHOR CONTRIBUTIONS
All authors contributed to the study design. ME, KG-D, and IS per-