Background
Unsafe abortions are estimated to account for 13% of maternal deaths globally. They give rise to a large number of short- and long-term complications [
1]. An estimated 21.9 million unsafe abortions are performed in the world annually [
2], and 97% of these occur in low-income countries. Africa is responsible for the second largest proportion of unsafe abortions (44%), and the highest rates globally (18–39 per 1,000 women) [
3]. Abortion laws of many countries are restrictive, leaving women no choice other than to procure unsafe abortion [
1]. General access to and use of contraceptive methods are limited, and the low status of women prevents them from making independent decisions on their own sexual and reproductive health (SRH) [
4].
Abortion is defined as the termination of a pregnancy, whether spontaneous, occurring before 22 weeks of gestation [
5], or induced [
6]. The World Health Organization (WHO) defines unsafe abortion as any procedure with the purpose of terminating a pregnancy that is performed by persons lacking the proper skills and/or that is performed in an unhygienic, non-medical setting [
2]. To address the complications related to incomplete, spontaneous or unsafely-induced abortions, post-abortion care (PAC) has been introduced in countries where abortion laws are restrictive. Standard PAC includes emergency care (such as resuscitation using blood transfusions, intravenous lines, antibiotics etc.); contraceptive counselling; treatment of sexually transmitted infections; human immunodeficiency virus (HIV) counselling; and community empowerment [
7].
The recommended method for treatment of an incomplete abortion during the first trimester is manual vacuum aspiration (MVA) [
5] or prostaglandin E1 analogue
misoprostol[
8‐
10]. International studies have compared the two methods (misoprostol and MVA) for the treatment of incomplete abortions, and the results show no significant difference in their effectiveness [
11‐
15]. Recent studies indicate the safety and efficiency of misoprostol use instead of surgical interventions to treat incomplete abortions in low-resource settings [
10,
16,
17]. Misoprostol has therefore been suggested as a first-line therapy in the developing world [
18], although there are still barriers to the accessibility of the drug [
19,
20].
Shortage of human resources in the health care system is common in low-income countries, especially in remote areas where maternal mortality is high [
21]. The strategic use of mid-level providers in the sense of task shifting/sharing – a process of delegating tasks to less specialized health care providers – has been identified as something that increases productivity and efficiency within health systems [
22,
23]. A task shift in the provision of treatment for incomplete abortion will increase women’s access to PAC [
24]. However, along with structural arrangements it is important to consider the providers’ attitudes to the tasks they are to provide. Attitude, beliefs and experience of health care providers have shown to influence the provision of post-abortion care [
25‐
28]. Moreover, training and experience has a positive effect on attitudes and it facilitates the bridging of cultural beliefs with the reality of service provision [
29,
30].
The Ugandan context
Uganda has one of the highest total fertility rates (6.7) in the world [
31], and it is estimated that 56% of all pregnancies are unintended. The contraceptive prevalence is 23% [
32] and the unmet need for contraceptives is 33% among women of reproductive age [
4]. Several reasons for the poor uptake of contraception are reported, for example misconceptions, side effects and poor acceptance [
33,
34]. Induced abortion is restricted and permitted by law only to save the life of a woman. Still, an estimated 297,000 induced abortions are performed annually, resulting in an overall abortion rate of 54 per 1,000 women aged 15–49 [
35]. This is regarded as high compared to the estimated rate for Eastern Africa [
35,
36]. About 77% of abortions treated in the public health system are induced [
35]. Additionally, unsafe abortions account for almost 40% of admissions to emergency obstetric care units, and they are responsible for significant morbidity and mortality among women [
37].
The National Policy Guidelines and Service Standards for Sexual and Reproductive Health and Rights of 2006 [
38] indicates that PAC is implemented in Uganda for treatment of incomplete abortion caused by spontaneous or induced abortion. In addition, the policy enables a variety of providers (midwife/nurse, clinical officer, medical officer and gynaecologist) to treat incomplete abortions. In spite of this progressive policy, mid-level providers such as nurses and midwives largely lack the proper training to provide PAC. This shortage of qualified providers limits the availability of safe emergency obstetric care, including PAC [
22,
39]. The 2006 SRH policy also permits abortions when the woman’s life is endangered, and in cases of foetal anomaly, rape, incest, cervical cancer, or upon a request from an HIV positive woman [
38]. However, the Ugandan Constitution states that abortion is only allowed if the procedure is authorized by law. Consequently, interpretations of the law are ambiguous, and providers may be reluctant to perform an abortion for fear of the legal consequences [
39]. A study from 2005 suggests that the most common method used to treat post-abortion complications in governmental as well as private facilities in Uganda was curettage for uterine evacuation [
4], a method that is dismissed by the WHO [
5]. It is suggested that optimizing the use of primary health care providers other than physicians, by task shifting, would be cost-effective and decrease the unnecessarily high maternal mortality in Uganda [
22]. However, the utilization of task sharing and shifting is poorly monitored, and little is known of the extent to which mid-level providers are involved. The acceptability and perception of PAC among providers at district level in Uganda has not been adequately explored. Moreover, providers’ needs of training and the challenges they face in the provision of PAC need to be better understood. This may help to overcome the barriers on provider-level, to better implement task sharing and finally to improve quality PAC in Uganda. Attending to the issue of unsafe abortion and PAC is a crucial step towards decreasing the maternal mortality ratio (MMR) in Uganda.
Aim of study
To explore physicians’ and midwives’ perceptions of post-abortion care, with regard to professional competences, methods, contraceptive counselling and task shifting/sharing in PAC.
Methods
Study design
An inductive study approach using an emergent design was employed, utilizing the qualitative method of in-depth interviews (IDI). Thematic analysis was used to structure the data.
Study setting
The study was performed at seven health care facilities situated in five different districts in the Central Region of Uganda. The region was chosen because the abortion rate has been reported to be the second highest in the country, and it is above the national average (62 per 1000 women) [
36]. The seven facilities were purposely selected because of their high caseloads in PAC. The caseloads were mapped through a survey done by one of the Ugandan researchers prior to the initiation of this study (2012). In addition, the employment of both doctors and midwives in a facility was regarded a criterion for inclusion, and we aimed to include facilities from rural, semi-urban and urban settings. The Ugandan health system is made up of a national referral hospital, regional and district hospitals, and health centres II-IV. Health centre IV is the most advanced, employing both doctors and midwives, and health centre II is the least advanced, employing a nurse or a clinical officer. Health centre I, also referred to as village health teams, are the lowest level and have no permanent accommodation [
40].
Study participants
A total of 27 IDI were carried out with health care providers in the health facilities listed above (Table
1). The majority of respondents (all the midwives and one doctor) were female (70%). The midwives had all worked for eight or more years, while the work experience of the doctors ranged between one and twenty years. The health facilities selected were staffed by doctors and midwives, and were equipped to provide basic and comprehensive emergency obstetric care in rural, semi-urban and urban areas. National, regional, district and sub-district levels were included (Table
1) to explore any discrepancies in service provision among the levels as well as differences between the cadres.
Table 1
Facilities included in data collection
Mpigi | Rural | Health Centre IV | 1 | 1 |
Luweero | Rural | Health Centre IV | 1 | 3 |
Nakaseke | Semi-urban | District Hospital | 1 | 4 |
Gombe | Semi-urban | District Hospital | 1 | 2 |
Entebbe | Urban | District Hospital | 2 | 2 |
Masaka | Urban | Regional Hospital | 2 | 3 |
Mulago | Urban | National Hospital | 2 | 2 (pilot) |
Total
| | |
10
|
17
|
Inclusion criteria
1.
Being employed in one of the hospitals/Health centres listed above (Table
1)
2.
Being a nurse, a midwife, a clinical officer or a doctor
3.
Actively participating in PAC
Data collection
The IDI were conducted in February and March 2012, and continued until no new data were encountered. Purposive sampling was employed. The person in charge at each health facility, such as the medical officer or head nurse/midwife, was used as a gatekeeper, facilitating the identification of eligible study subjects. All participants signed a written consent prior to the IDI, and the interviews were conducted in the hospital, meaning that interruptions could sometimes occur. Two pilot interviews were performed at Mulago hospital to test the interview guidelines, resulting in a revision to impart a better in-depth character to the questions. The final guideline was semi-structured, open-ended and utilized probes. Topics covered included (i) attitudes towards abortion, PAC and family planning, (ii) perceptions of methods for uterine evacuation, (iii) skills and competences needed and (iv) task sharing in PAC. The IDI were performed in English, lasted for 30–60 minutes and were tape-recorded. The research team consisted of researchers with different cultural and professional backgrounds. The first author and main researcher, a Swedish woman with a background in global health, and a Ugandan assistant, a woman with a degree in public health, conducted the data collection. All the researchers were involved in the analysis and interpretation of the data.
Data analysis
The recorded data were transcribed verbatim, read through several times and carefully coded manually. Since little is known about the views and perceptions on PAC in Uganda, thematic analysis was used [
41]. Furthermore, a thick description to reflect the content of the data was employed, and codes were organized into sub-themes using an inductive approach. Sub-themes were arranged into semantic themes focusing on identified patterns as well as on the broader meanings and implications of the data. The analysis was done with a realistic or epistemological approach, looking at experience and meaning in a straightforward way. The interpretation of the data was continuously discussed and re-evaluated by all researchers in the team. Additionally, the data were reviewed to identify extracts for illumination of the themes.
Ethical considerations and confidentiality
Ethical approval was obtained from the Makerere University College of Health Sciences, School of Biomedical Sciences Research and Ethics Committee, Kampala. National approval was given by the Uganda National Council for Science and Technology. The study was identified as a minimal risk study since it addressed health care providers’ perspective on a work task, and their opinions on abortion. A written consent form stated the rights of the participants, and confidentiality and the anonymity of the participants were guaranteed. The data and informed consent forms are stored safely under lock and key at Karolinska Institutet and are used only by the researchers involved in the study.
Discussion
This study highlights the lack of skills and up-to-date guidelines for the provision of adequate quality PAC at district level in Central Uganda. Moreover, our findings identify midwives as the main providers of PAC. Providers identified barriers to provision of quality care as the lack of in-service training, poor supervision, and too heavy a workload to manage PAC in addition to other services. Moreover, our study identified a need for training in the use of MVA and misoprostol for uterine evacuation and training in contraceptive counselling in the context of PAC.
Experience and training seems to be crucial factors determining the willingness to provide safe abortions. The providers expressed conflicting opinions when discussing legalization and provision of abortion. Providers with experience or in-service training, or both were the ones that preferred performing an abortion rather than rejecting a woman in need. Those with a negative attitude towards abortion would describe it as sinful and as nothing they could support. These providers were more likely not to have much experience or training. This is in line with findings from an Ethiopian study where providers with experience of safe abortion were 2.5 times more likely to have a positive attitude towards abortion than those without any experience [
29]. Several studies highlight the importance of education and in-service training to improve attitudes and furthermore the quality of abortion care [
25‐
28]. Similar to our findings, a study conducted in Vietnam sheds light on the contradiction between cultural norms and the reality facing health care providers in the context of adolescent SRH [
30]. It is recommended that education and training addresses this controversy specifically [
42].
Frustration provoked among providers in the provision of PAC was mainly explained by the lack of skills and resources. A major concern among informants was the limited training in the use of safe methods for uterine evacuation and contraceptive counselling in a post-abortion context. Sources of frustration among the midwives were mainly derived from their lack of skills and knowledge of how to care for women with incomplete abortions, in the absence of a doctor. Other barriers to quality PAC, as identified by the informants, were the lack of space and privacy, and the shortage of supplies, equipment and contraceptives. Apart from the issue of the lack of human resources pervading the majority of health services in low-resource settings [
21], similar barriers to PAC have been identified in other studies. Barriers such as sustainable access to MVA instruments [
43], as well as poor contraceptive counselling and lack of available contraceptives [
44‐
46]. Moreover, barriers to misoprostol use in PAC refer to restrictive drug policies as well as poor access and affordability of the drug [
19,
20,
22,
47]. Misoprostol has recently (2013) been approved for post-partum haemorrhage and incomplete abortion in Uganda. It is therefore important to develop a strategy for effective implementation for the use of misoprostol in PAC. The use of misoprostol for uterine evacuation is believed to result in more accessible and improved quality PAC [
8,
10].
Our study highlights midwives as the main providers of PAC in terms of assessment, care and uterine evacuation in the first trimester. Midwives reported to mainly use MVA for uterine evacuations, however they also mentioned use of D&C and misoprostol. The rare use of misoprostol in PAC at district level was explained by providers’ limited skills in the medical treatment of incomplete abortion. Additionally, lack of stock and absence of hospital guidelines in the use of misoprostol were mentioned. Similar factors have previously been identified as limiting the provision of other SRH services in Uganda [
33]. Interestingly, the national medical store in Uganda has reported an overstocking of misoprostol since 2011, and non-existent demand from health facilities [
48]. Misoprostol is effective for the treatment of incomplete abortions [
8,
18,
49] and feasible in resource-limited settings where surgical treatment is largely unavailable [
16,
50]. In addition, a study from Uganda shows no difference in effectiveness between the use of misoprostol and MVA in the treatment of incomplete abortion [
15]. In line with recommendations and evidence from previous studies, our study emphasizes the importance of specific in-service training to scale up the use of misoprostol to improve PAC in Uganda [
18,
22,
24].
According to the WHO guidelines of PAC, contraceptive counselling ought to be included [
5]. The providers in our study understood the advantages of providing family planning in PAC. However, they felt that they were unable to provide it due to the lack of human resources and in-service training. Further reasons that were given are the absence of hospital guidelines and a lack of structure (the facility set-up and resource allocation) to provide contraceptive counselling in the context of PAC. Studies show that accessible family planning can prevent repeated unintended pregnancies and abortions. Still, contraceptive counselling is the component of PAC that is commonly overlooked [
44‐
46]. Integrating the provision of family planning methods into PAC has been suggested as a way to decrease the number of unintended pregnancies [
51‐
54]. Hence, it is of great importance to support PAC providers in their provision of contraceptive counselling.
A key finding of our study was that midwives felt that PAC was a main contributor to stress in general. This was particularly in terms of additional workload and the difficulties involved in the consequences of unsafe abortions. Doctors however, did not seem to perceive PAC as major contributor to their workload. This emphasizes the extent of the midwives’ role in the provision of PAC. In terms of roles and responsibilities doctors and midwives had different perceptions. Doctors believed that midwives did not conduct MVA without supervision, whereas midwives reported that they did conduct MVA generally when doctors were not available. Doctors would appreciate when midwives took responsibility and perceived that efficiency would increase if tasks were to be shared. However, doctors emphasized the importance of training midwives prior to task sharing. Other studies from Africa show that task shifting contributes to more sustainable, cost-effective and accessible health care [
23,
24,
55‐
58]. Furthermore, the provision of first-trimester abortion utilizing MVA or misoprostol has been technically feasible for mid-level providers for years. When policies and laws are liberal, task shifting can increase women’s access to safe services [
24]. Due to the restrictive abortion laws in Uganda, doctors in hospital settings are the main providers trained to perform safe abortion. This also limits access to safe treatment after an incomplete abortion, especially in remote areas [
39]. A study conducted in Uganda reports an informal amount of task shifting as a response to shortages of staff. This has resulted in dissatisfaction among mid-level providers due to the increased workload and the persisting limited resources [
22]. The SRH policy (2006) in Uganda enables midwives and clinical officers to provide PAC [
38]. This study suggests that task sharing occurs in the clinics, but not in an organized way with clear guidelines, training and infrastructure. Moreover, the present sharing of tasks is partial, since midwives seldom administer misoprostol themselves. This, unlike MVA, was perceived as the doctors’ responsibility.
This study is focusing on PAC, however its implications may be applied within other areas of health care in Uganda. Understanding the sources of frustration among providers and their perceptions of barriers to provide quality care are crucial factors that need to be considered in the development of new policies as well as the implementation of present policies. Improving PAC is an important intervention to decrease maternal morbidity and mortality in Uganda. It is also a step towards a more equitable health care with focus on women’s reproductive health and their own choices.
Methodological considerations
The emergent nature of the study allowed each IDI to have its own individual character. Focus Group Discussions (FGD) were considered, however due to logistical barriers such as long distances between the facilities and little availability of PAC providers, FGDs would not have been feasible in the context. Interviews were conducted until no new data were encountered and saturation was reached. The data from each IDI was rich, explained by the good and safe environment created on each occasion. The interviews were carried out in English, the working language of health workers in Uganda but a second language for most people. It was therefore important to consider potential misunderstandings, even though no obvious problems were encountered. The fact that the interviewer was a young Swedish woman could also have influenced the responses, but this problem is thought to have been circumvented with the presence of the Ugandan research assistant. The role of the research assistant was to take notes and to help with any language difficulties. Her involvement appeared to create trust between the interviewees and the interview team and thus is reflected in the rich data. The interview guideline was pre-tested, revised and validated through pilot interviews and peer review. The credibility was further strengthened by quotes highlighting the findings. The data were continuously analysed by a group of individuals with different cultural and professional backgrounds, to strengthen dependability. Additionally, it is important to consider the researcher as an interpreter in terms of the construction of meaning in the research process. To ensure reflexivity, the transcripts were read several times, the data were discussed and supervisors were consulted. The study was conducted in governmental hospitals in the Central Region of Uganda, but transferability of findings to other regions of Uganda as well as context-similar settings is likely to be relevant.
MP conducted the study as a Master’s thesis in Global Health at Karolinska Institutet in Stockholm, Sweden. She has a background in biomedicine (BSc) and has previously worked with SRH in Mozambique and Nigeria. She is currently pursuing her PhD studies in abortion-related research in India. KGD is a professor and senior physician at the Department of Women’s and Children’s Health at Karolinska Institutet in Stockholm, Sweden, with previous experience in Uganda and several other African countries. CK is a medical doctor with an MSc in Obstetrics and Gynaecology at Mulago National Referral Hospital in Kampala, Uganda. RN has a BSc in Public Health from Makerere University, Kampala, Uganda. MKA is a midwife and a research scientist in the Department of Women’s and Children’s Health at Karolinska Institutet in Stockholm, Sweden, as well as a lecturer at Dalarna University in Falun, Sweden. She has previous experience in Uganda as well as in other African and Asian countries.
Competing interests
There are no competing financial or non-financial interests involved in this study.
Authors’ contributions
MP contributed to the conception and design of the study, applied for ethical approval, collected data and is the main interpreter of the data, drafted and approved the submitted manuscript. KGD revised the manuscript for intellectual data and approved the submitted manuscript. CK contributed to the conception and design of the study and its revision for ethical approval, the data collection and the interpretation and analysis of the data, and approved the submitted manuscript. RN contributed to data collection, peer review and interpretation of data, and approved the submitted manuscript. MKA contributed to the conception and design of the study, the data interpretation and analysis and the revision of the intellectual data, and approved the submitted manuscript. All authors read and approved the final manuscript.