Assessing facility capacity to provide safe abortion and post-abortion care in Liberia: a 2021 signal function survey across 48 public health facilities

Abstract

Background

Access to safe abortion is legally restricted in Liberia, forcing women to resolve unintended pregnancies through unsafe methods, leading to severe illnesses and deaths. Liberia’s Government has committed to addressing abortion-related maternal mortalities by availing comprehensive post-abortion care. However, limited information exists on the capacity of health facilities to provide quality abortion-related care. This paper assesses the extent to which health facilities in Liberia are capable of delivering safe abortion and post-abortion care services.

Methods

Data for this analysis are drawn from a signal function survey conducted across 48 public facilities in Liberia from September to November 2021. The signal function survey captures several safe abortion and post-abortion care-related services, including staff training, equipment, commodities, and supplies. Data were collected from health providers knowledgeable about abortion-related care, such as safe abortion and post-abortion care, across sampled health facilities using a structured questionnaire. Data analysis involved summarizing proportions of clinics, health centers, and hospitals with the capacity to provide either basic and/or comprehensive safe abortion and post-abortion care.

Results

Out of the 48 facilities, 65% and 28% were classified as capable of providing basic and comprehensive post-abortion care (PAC) services, respectively. Fewer facilities (27%) could provide basic safe abortion care (SAC) and comprehensive SAC (16%). Differences by facility level were statistically significant for comprehensive PAC. The PAC signal functions fulfilled by the fewest facilities included referral capacity, blood transfusion, and surgical or laparotomy capacity.

Conclusion

The study highlights the limitations to providing basic SAC and PAC among our sample of public health facilities in Liberia and the poor capacity of these health facilities to provide comprehensive PAC and SAC services in particular. Full implementation of the 2019 National Comprehensive Abortion Care Guidelines could strengthen critical SAC and PAC services by ensuring adequate resources and training of the healthcare workforce.

Introduction

Under the Liberian Penal Code (1978), abortion is highly restricted. Section 16.3 of the Penal Code provides justifiable grounds for abortion as follows: to save the life of the woman; to preserve the physical or mental health of the woman; for pregnancies resulting from rape or incest or other felonious intercourse; or fetal impairment [1]. Despite the law, induced abortions are commonplace in Liberia. A recent study estimated that approximately 38,400 induced abortions occurred in the country in 2021, translating to an induced abortion rate of 30.7 per 1,000 women of reproductive age [2]. Most of these were estimated to be performed using unsafe methods and procedures [2].

Unsafe abortions are a leading cause of maternal morbidity and mortality, with a sizable proportion of complications severe enough to warrant treatment in a health facility [3,4,5]. Abortion-related complications are most common when safe abortion care (SAC) is restricted, forcing women to resort to unsafe abortion methods or procedures [6]. The study on the incidence of abortion in Liberia reported that half of the women who sought post-abortion care in Liberian health facilities in 2021 presented with either severe (37%) or near miss (11%) complications, such as infections and sepsis (59%), and hemorrhage (31%). Notably, the vast majority of women (67%) seeking post-abortion care did so in public health facilities [2].

Post-abortion care (PAC) is a set of critical life-saving interventions for reducing maternal morbidity and mortality related to abortion [7]. According to the Liberia National Guidelines for Comprehensive Abortion Care, the key elements of PAC are emergency treatment for complications of spontaneous and induced abortions, family planning and provision of other SRH services, community linkages, empowerment, awareness, and mobilization [8]. The Government of Liberia has committed to addressing maternal mortality, including those linked to abortion, through the provision of quality and comprehensive PAC. Among key recent policy-level improvements are the Liberia Reproductive Maternal Newborn Child and Adolescent Health (RMNCAH) Investment Case of 2016–2020 that cites unsafe abortion-related maternal morbidity and mortality as a priority investment area [9], as well as the first-ever National Comprehensive Abortion Care Guidelines that specify programmatic and clinical interventions for safe and legal abortion and PAC to help health providers streamline the provision of quality PAC [8]. Further, in 2019, the Ministry of Health spearheaded the process of revisions to the country’s Public Health Law (Title 33), which includes provisions for expanding access to safe abortion. As of July 2022, the revised Public Health Law had passed the Lower House of Parliament and is being considered by the country’s senate. Successful passage of the proposed Liberia public health law would be remarkable progress in removing barriers to safe abortion care and decriminalization of abortion in Liberia.

Studies have documented several barriers that impede timely access to PAC and SAC services across sub-Saharan Africa (SSA), including legal restrictions on abortion [10, 11], abortion-related stigma and negative attitudes of health providers [12, 13], low levels of awareness and knowledge of the law among women [14], and limited capacity of healthcare systems to provide quality abortion care services [15, 16]. There is wide consensus on the value of expanding access to PAC and SAC on maternal health outcomes [17, 18], and the World Health Organization (WHO) in 2022 released updated abortion care guidelines that recommend improving access to high-quality, person-centered abortion-related services [19]. However, little is known about the capacity of health facilities to provide these services in Liberia. As such, this study aims to examine the capacity of a sample of Liberian health facilities to provide SAC and PAC using the signal functions framework. The signal function framework that was first developed by Healy and colleagues, who adopted an existing United Nations-fronted model for monitoring the availability and use of Emergency Obstetric (EmOC) services. Healy et al. (2006) developed and described the signal function indicators for monitoring the availability and use of abortion care services [20]. The signal function framework has been used to assess the capacity to provide SAC and PAC in several other settings, including Zambia [21], Ghana [22], Burkina Faso, Nigeria, and Kenya [16, 23]. The signal function approach has also been extensively used to assess the availability, readiness, and accessibility of post-abortion care in the DRC [24], Burkina Faso [25], Niger [26], Nigeria and Côte d’Ivoire [27]. Others have used the approach to assess changes in post-abortion and safe abortion care coverage, capacity, and caseloads after the global gag rule policy was imposed in Ethiopia and Uganda [28]. Our analysis is the first of its kind in Liberia. It provides crucial data on facility capacity to provide PAC and SAC services and to understand key gaps and challenges where service provision could be improved.

Materials and methods

Study context

The signal function survey targeted a sample of health facilities, including clinics, health centers, and hospitals across all regions in Liberia. Liberia is a West African country with about 5.3 million people. It is geographically divided into 15 counties, grouped into five regions. Liberia’s healthcare system is organized into three tiers: primary, secondary, and tertiary level facilities. Primary-level facilities include clinics that provide basic health services and emergency care, while secondary and tertiary facilities can provide more complex care. Secondary-level facilities consist of health centers and county hospitals; tertiary-level facilities include regional and national referral hospitals (John F. Kennedy Medical Center). According to the 2019 National Guidelines on Comprehensive Abortion Care [8], all health facilities in the country are required to provide PAC regardless of the suspected cause of abortion. However, based on the same guidelines, induced abortion can only be performed in health centers and hospitals by a licensed doctor, following provisions of the Liberia Penal Law (Sect. 16.3).

Study design, sampling, and recruitment

The signal function survey was part of a larger multi-component study aimed at estimating the incidence of abortion and the severity of abortion-related complications in Liberia in 2021. As such, the signal function survey (a cross-sectional health facility-based survey), was embedded within the broader study to assess the capacity of health facilities to deliver PAC and SAC services in Liberia.

The larger study included a Health Facility Survey (HFS) administered among a nationally representative sample of health facilities theoretically capable of providing PAC services. A previous publication (under review when writing this paper) has a detailed description of the sampling strategy [2]. In brief, the research team sampled a total of 132 facilities for the HFS, which represents 100% of hospitals able to provide PAC in Liberia in 2021, 84% of all health centers, and 7.5% of all clinics. To get to the 132 health facilities, a two-stage stratified random sampling technique was used to draw the list of facilities to be included in this survey, that is, (a) the highest sub-national administrative units (i.e., regions, namely North Central, South Central, North Western, South Eastern A and South Eastern B regions), and (b) the levels of health facilities. The researchers then obtained an updated master list of all health facilities within the different regional units in the Country from the Ministry of Health (updated October 8, 2020), and a sample of facilities was drawn for the study.

A total of 128 facilities participated in the HFS, representing a response rate of 97% of health facilities. All the 128 health facilities in the HFS were considered for participation in the signal function survey, from which the data for this analysis are derived. Facilities eligible for participation in the signal function survey were public health facilities that accepted participation in the study and had PAC/SAC providers who had been at the health facilities for at least six months and who were to be interviewed.

The focus on public health facilities in Liberia was intentional. While both private and public health facilities can provide PAC, we only focused on public facilities because government policies and investments mainly target public facilities. Further, private facilities tend to be business-oriented and often set their standards and quality above and beyond what is stipulated in national policies and guidelines. These private facilities usually set higher standards as an investment decision, not public policy. Since the study is intended to have direct policy relevance to the Ministry of Health, we excluded private facilities. Ultimately, about 48 health facilities completed the signal function survey questionnaire and had complete data needed to describe the capacity of health facilities to provide PAC and SAC services. The remaining 80 health facilities only had data for the HFS and the Prospective Morbidity Survey (a clinical component in the larger study) but lacked complete data for the signal function analysis.

Data collection and study variables

Trained field workers administered a structured signal function questionnaire adapted from previous studies conducted in Burkina Faso, Kenya, and Nigeria [15]. The tool was initially developed by Healy et al. (2006), who drew from the UN Guidelines on Emergency Obstetric Care (EmOC) for the care of women and newborns during and after pregnancy and delivery, to propose a set of indicators for monitoring the implementation of safe abortion care (SAC) interventions [20]. The questionnaire was administered to one health provider per facility, identified by hospital management as knowledgeable on PAC services. Within large health facilities, such as national referral hospitals and regional hospitals, respondents included heads of the obstetrics and gynecology department or a key obstetrician-gynecologist. In lower-level facilities, such as clinics and health centers, a nurse, a midwife, or another health provider knowledgeable on abortion services provided in the facility was interviewed.

The signal function questionnaire captured details on the availability of key sexual and reproductive health services, where providers responded to the question - Which of the below EmONC procedures are performed in this facility? Respondents were asked to respond with either a Yes/No to a list of services, including parenteral antibiotics, administration of uterotonics/oxytocics, administration of IV fluids, provision of contraceptives, medical and surgical PAC, and medical and surgical SAC. Other services included laboratory tests such as HIV, Hepatitis, and blood grouping, among others. Whenever a service was indicated as unavailable, the subsequent questions probed why the service was not provided. Choices included no health providers available (meaning the relevant providers of the appropriate cadre were unavailable), absence of trained health providers (meaning the providers of appropriate cadre may be there, but they lack the necessary training to offer the service), no equipment (absence of the appropriate and/or functional equipment), no commodities or supplies (commodities and/or supplies not available or available but expired), against hospital or management policies (where there is an official policy restricting a specific service delivery at a facility), no cases requiring the procedure (no patient presented with a complication needing that particular service), against the workers’ morals/ethics (where a provider(s) cites a service as against their moral or ethical standing), and not available at this health facility level (where an official policy restricts a specific service delivery at a facility level). In this case, the respondents were read the above reasons for lack of service and asked to respond Yes or No to each reason. When a service was indicated as available, respondents were asked whether that service had been provided in the past six months. In a few cases where a facility responded that they do not provide a service but had provided it in the past 6 months, that service was coded as available. Further, the tools also captured the availability of specific post-abortion care equipment, medicines, and commodities (Supplementary file 1). Data collectors observed the stock and functionality of these essential medicines and supplies. Data collection took place between September and December 2021. The questionnaire was programmed in SurveyCTO and administered using tablets.

Statistical analysis

Statistical analysis was conducted in R version 4.3.2. Unweighted data on signal functions were summarized to describe the capacity of health facilities to deliver basic and comprehensive SAC and PAC services. Weighting was not applied, as the sample size was small and not regionally representative, limiting the extent to which the findings could be extrapolated to all public health facilities in Liberia. Drawing from previous studies by Owolabi et al. (2019) and other signal function analyses [16], we first conducted a descriptive analysis calculating the number of facilities capable of providing each signal function, overall and by facility level. Facilities were counted as being able to provide removal of retained products of conception if they responded “Yes” to the provision of either medical or surgical post-abortion care. Facilities were counted as being able to provide first-trimester legal termination of pregnancy if they reported having performed medical or surgical terminations of pregnancies less than 12 weeks in the last six months; the same rule applied for second-trimester legal termination of pregnancy (pregnancies greater than 12 weeks gestation). For the provision of parenteral antibiotics, uterotonics/oxytocics, and IV fluids, a facility satisfied the signal function if the respondent affirmed the service was provided at the facility. For contraceptive provision, the facility had to have the commodities in stock on the day of data collection. For short-term contraceptives, facilities satisfied this signal function if they had either male condoms, female condoms, oral contraceptive pills, or injectables in stock; for long-term, stock of implants or IUDs satisfied this signal function. Facilities were counted as being able to provide blood transfusion and surgical procedures to address post-abortion complications if they reported having provided the service in the last six months. Surgical procedures included laparotomy, hysterectomy, hysterotomy, hysteroplasty, and laparoscopy. Finally, a facility satisfied the referral capacity indicator if they reported having an ambulance or another vehicle for emergency transportation of clients stationed at their facility that had fuel on the day of data collection or if the facility had access to a vehicle for emergency transportation stationed at another facility.

Next, we generated composite indicators to assess the provision of basic and comprehensive PAC and SAC. The indicators that comprise each composite indicator are reported in Table 1. For comprehensive PAC and comprehensive SAC, a facility had to fulfill all the signal function indicators designated for basic PAC and basic SAC, respectively, in addition to the indicators designated for the provision of comprehensive PAC or comprehensive SAC services, including surgical procedures such as laparotomy, blood transfusion, second-trimester legal termination of pregnancy and provision of at least one long-acting reversible contraceptives (implants or IUDs). The provision of basic PAC was a prerequisite for the provision of basic SAC, meaning all facilities identified as capable of providing basic SAC also had to meet all indicators for basic PAC. The same applies for comprehensive SAC and comprehensive PAC. Clinics were excluded from the comprehensive PAC and SAC analysis, as clinics in Liberia are not expected to provide comprehensive PAC and SAC services. Compared to previous signal function analyses, a key departure is our inability to assess facility hours and staff availability. The tool did not assess whether the facility was open 24 h a day nor the number of full-time and part-time staff trained in PAC and/or SAC, two indicators defined in other signal functions analyses as essential for the provision of SAC and PAC. We did not weight the dataset as the sample size was fairly small and does not necessarily reflect the universe of facilities in Liberia or any singular part of Liberia.

Table 1 Variables used to measure the capacity of safe abortion care and post-abortion care

Further, we ran descriptive statistics to assess why some facilities said they did not provide certain signal functions. The proportion of facilities reporting non-provision were counted, along with the reason(s) given for not providing the signal function. Reasons for non-provision were not mutually exclusive, meaning a facility could cite various reasons for not providing a service in the past 6 months.

Results

Description of health facilities in signal function survey

Complete signal functions survey data from 48 facilities were included in the analysis. The distribution across facility levels was as follows: 33% of sampled facilities were clinics, 35% were health centers, and 31% were hospitals. Most of the included facilities (52%) were located in the South Central Region, while only two facilities (4%) were in the North West region. Additional details on the distribution of facilities by level and region can be found in Table 2.

Table 2 Description of facilities in the signal function survey

The capacity of health facilities to provide basic PAC

Of the 48 facilities in the survey, nearly two-thirds (65%; 31/48) were classified as capable of providing basic PAC. Basic PAC services included the removal of retained products of conception using medical or surgical procedures, antibiotics, IV fluids, uterotonics, short-acting family planning methods, and referrals. Approximately three in four hospitals (73%; 11/15) and health centers (76%; 13/17) fulfilled the definition of basic PAC, while only 44% of clinics in our sample (7/16) were classified as able to provide basic PAC (Table 3). This variation was not statistically significant (p = 0.12).

The only signal functions fulfilled by all facilities in our sample were the provision of IV fluids and uterotonics/oxytocics. Parenteral administration of antibiotics were provided in all but one health center. 96% of facilities could facilitate the removal of retained products of conception either medically or surgically, and a similar proportion (94%) had stocked at least one short-acting contraceptive method. The basic PAC signal function fulfilled by the fewest facilities was referral capacity; 71% (34/48) of all facilities had the systems to refer patients to other facilities for care; when disaggregated by facility level, only half of the clinics in our sample fulfilled this signal function.

The capacity of health facilities to provide comprehensive PAC

Compared to basic PAC, far fewer facilities could provide comprehensive PAC services, which include all basic PAC services and the provision of long-acting reversible contraceptives, blood transfusion, and surgical procedures such as laparotomic operations (Table 3). Only about one-quarter of the 32 hospitals and health centers (28%; 9/32) fulfilled all comprehensive PAC signal functions and were classified as capable of delivering comprehensive PAC. Differences were statistically significant by facility level (p < 0.05); fewer than half of the hospitals (47%; 7/15) and only two of 17 health centers (12%) could provide these services. Most health facilities (91%; 29/32) could provide long-acting reversible family planning methods. All hospitals surveyed could provide blood transfusion, but this was not true for health centers. By standards in Liberia, all health centers and hospitals should provide blood transfusion; however, only 12% of surveyed health centers could. The ability to perform surgical procedures (such as laparoscopic procedures, also a standard for both hospitals and health centers in Liberia) to address abortion-related complications was also scored poorly. Less than half of the facilities (41%; 13/32) could provide surgery or laparotomy; this service was most available in hospitals (53%) and only available in a minority of health centers (29%).

Table 3 Provision of basic and comprehensive post-abortion care

The capacity of health facilities to provide basic and comprehensive SAC

In contrast to basic PAC, far fewer facilities were classified as capable of providing basic safe abortion care (Table 4). In total, only one-quarter (27%; 13/48) of sampled facilities were classified as able to provide basic SAC, with hospitals and health centers (27% and 35%, respectively) having a higher capacity than clinics (19%), although differences were not statistically significant. As with basic PAC, all facilities were capable of providing uterotonics/oxytocics and IV fluids, and nearly all provided parenteral antibiotics and short-acting contraceptive methods. However, only slightly more than a quarter of facilities (29%; 14/48) had provided first-trimester legal termination of pregnancy in the past six months.

Only five health centers and hospitals in our sample (16%) were classified as able to provide comprehensive SAC. Hospitals in our sample were most likely to fulfill all basic and comprehensive SAC indicators, although fewer than one-third (27%) of all hospitals could provide comprehensive SAC. Only one health center was classified as able to provide comprehensive SAC. Differences by facility level were not statistically significant. Only one out of every five facilities (22%) had reported performing a second trimester medical or surgical legal termination of pregnancy in the last six months; by facility level, this included five hospitals (33%) and two health centers (12%).

Table 4 Provision of basic and comprehensive SAC

Reasons for the lack of basic and comprehensive PAC and SAC capacity

For each signal function reported, facilities were asked whether they provided this service in the past six-months and if not, why not. These reasons are reported in Table 5. Among the eight facilities that reported not providing any PAC service in the past six months, the main barriers to provision included not having the necessary supplies or equipment (4 facilities; 50%), against hospital policy (4 facilities; 50%) as well as not having any cases (6 facilities; 75%). For safe abortion care, approximately half of the facilities that did not provide SAC in the past six months reported not having any cases, regardless of trimester (first or second). Many facilities also reported that SAC services were unavailable at their facility level or the service was against hospital policy. Around a quarter of facilities that have not provided SAC in the past six months also reported that it was against the ethics or morals of their health workers. The main barrier to the provision of blood transfusion was the lack of supplies or equipment (10 facilities; 67%), while the main reason for the non-provision of surgical procedures to manage abortion complications was the lack of cases (11 facilities; 58%).

Table 5 Reasons for non-provision of signal functions in the past 6 months

Discussion

This paper utilizes the signal function methodology to assess the capacity of health facilities to provide post-abortion and safe-abortion care in Liberia. The analysis reveals key strengths and weaknesses in the health system’s capacity to provide critical basic and comprehensive abortion-related services. In our sample of 48 facilities, nearly two-thirds (65%) were classified as able to provide the full spectrum of basic PAC services, even though the capacity for basic PAC varied by facility level, with 73% of hospitals and 76% of health centers able to provide basic PAC, compared to only 44% of clinics. Our findings show greater capacity to provide basic PAC among our sample of Liberian health facilities compared to what has been reported in previous studies from other countries, such as Burkina Faso (12.1%), Ethiopia (15.0%), Nigeria (8.6%), Kenya (6.3%), Côte d’Ivoire (37.5%), and Uganda (17.8%) [23, 27, 28]. Similarly, Owolabi et al., in a 2019 multi-country study across 10 low and middle-income countries, reported that less than 10% of primary-level facilities could provide all elements of basic post-abortion care [16]. Compared to the caseload of PAC patients that visit each of the facility levels, a recent national abortion survey in Liberia reported that only 14% of PAC patients seek care in Hospitals, 12% in health centers, and the vast majority of patients (74%) visit the clinics for post-abortion care services [2]. In contrast, only 44% of clinics could offer the full complement of basic PAC services. In addition, more than 67% of patients seeking PAC go to public health facilities [2].

Our findings align with those from a study in Senegal, where 53% of facilities were defined as capable of providing basic PAC when excluding staff availability and referral indicators [16]. However, there exist slight variations in how the above studies were designed and how the datasets were analyzed that may have also contributed to the findings reported. For instance, there is no consensus on a fixed standard for assessing post-abortion care (either basic or comprehensive), and various studies have included or excluded certain variables (such as availability of staff, facility hours, nature of staff training, and referral services).

The analysis also revealed limited capacity to facilitate referral of patients needing further medical care, with less than three-quarters of health facilities in our sample (71%) fulfilling this signal function. When referral capacity is removed from the analysis, the capacity to provide basic PAC increases by 17%, from 71 to 88%. Referral capacity was lowest among clinics, which signals a significant gap since primary-level facilities often have sub-optimal capacity to provide comprehensive care, and PAC patients visiting these facilities are likely to require further medical care in secondary and tertiary-level facilities. The lack of referral capacity likely results in delays to the referral process, worsening of complications, and ultimately poorer care outcomes, including longer hospital admission periods, higher costs of care, and death [29]. This weak referral capacity is not unique to Liberia but has also been reported in Burkina Faso, Kenya, and Nigeria, among other countries [23].

Fewer facilities in the sample (28%) fulfilled all comprehensive PAC signal functions, among which 47% were hospitals and 12% were health centers. Our findings are slightly higher than those reported among referral-level facilities in Ethiopia, in which 11.2% were defined as having comprehensive PAC capacity [28]. However, our findings were lower than those from similar studies in the region. For instance, a study by Juma et al. reported that about 30% of secondary and tertiary facilities in Burkina Faso and 42.9% in Kenya could deliver the entire package of comprehensive PAC services [23]. Similarly, the multi-country analysis by Owolabi and colleagues reported that among referral-level facilities, 33% in Rwanda and 32% in Senegal could provide comprehensive PAC services [16]. Nevertheless, the same study reported much higher proportions in Tanzania (53%) and Malawi (58%) [16]. Differences in the ability to provide basic and comprehensive PAC by facility level are expected, as lower-level facilities are typically less-resourced and may lack critical human and material resources for providing abortion care. Indeed, several studies have reported differences in the capacity to provide PAC by facility level and even across regions in a country [22, 30]. Variations in basic and comprehensive PAC capability across countries may reflect differences in the structure of health systems and the expectations for each level of health facilities. More broadly, these disparities may also be linked to country-level prioritization and investments to strengthen health services linked to maternal health. In this study, we attempted to broadly align the classification of facilities into primary and referral levels similar to classifications in Service Provision Assessment (SPA) surveys.

Nonetheless, this study highlights important gaps in service delivery across facility levels, particularly with regard to the lower-performing items, including uterine evacuation procedures, provisions of short-acting and long-acting contraceptives, referral capacities, surgical procedures, and blood transfusion. Considering the critical role of these services in the treatment of post-abortion complications, these gaps certainly affect the quality of services women receive when they seek PAC. Notably, while our assessment qualified facilities that could provide at least one short-acting or long-acting contraceptive method as capable of this service, it is essential to highlight that this categorization does not reflect an optimal situation. Indeed, when a facility only has one contraceptive method available, this limits women’s ability to have the choice of a desired method, which could be associated with low uptake, dissatisfaction and discontinuation [31, 32]. Other previous studies have reported the consequences of poor quality of PAC services, including delays in seeking care, repeat unintended pregnancies, repeat abortions, severe complications, and higher abortion-related death rates [15, 29]. Our review of why facilities had not performed certain procedures over the last six months highlights key leverage points to improve basic and comprehensive PAC capacity in Liberia. While the lack of cases was an important reason for the non-provision of certain PAC/SAC services in the past six months, it implied that no single patient presented with a complication that needed that particular service. The lack of PAC cases needs further contextualization since the signal function survey was embedded within a larger study that implemented the HFS tool (which captured, among other things, the monthly or annual PAC caseloads in each facility) and also observed a sample of nationally representative health facilities for all cases presenting with post-abortion complications over a one month period (30 days) [2]. Within the HFS, there were several facilities, especially those in urban environments and also in very remote settings and of lower facility levels, that reported less than six cases in a year. We, however, also appreciate that PAC cases are sometimes recorded differently or deliberately concealed for various reasons.

Nevertheless, other important barriers to PAC/SAC provision included lack of equipment/supplies, moral or ethical objections, and belief that the service was against hospital policy. No supplies/equipment was also the foremost reason for facilities not offering blood transfusion; stock-out of blood products has also been widely reported in other sub-Saharan countries [11].

Regarding safe abortion care services, only one in four facilities in our sample (27%) could provide basic SAC, with hospitals (27%) and health centers (35%) showing greater capacity compared to clinics (19%), although the differences were not statistically significant. The signal function met by the fewest number of health facilities was the capacity to provide first-trimester legal termination of pregnancy; only 29% of facilities said they had provided this service in the past six months. Only 16% of hospitals and health centers could deliver the full complement of comprehensive SAC services in Liberia. This is mostly attributed to the fact that only 22% of facilities had reported providing second-trimester medical or surgical SAC in the last six months, coupled with the low capacity for blood transfusion and surgical/laparotomy capacity reported above. A study by Stillman et al. in Ethiopia reported significantly higher figures for both basic and comprehensive SAC [28]. For instance, the indicated study reported that up to 74% (primary) and 96% (referral-level facilities) could deliver basic SAC. Similarly, about 32.6% of referral-level facilities in Ethiopia could deliver comprehensive SAC [28]. Of course, it deserves mention that the legal framework for abortion in Ethiopia is much more liberal than in Liberia, where legal termination of pregnancy is not widely available. Indeed, the lack of cases was the foremost reason for the non-provision of second-trimester SAC mentioned by respondents. The lack of cases here may imply that very few SAC patients are captured in hospital medical records as they remain largely invisible. Sometimes, even when they present at facilities for legal SAC, they may be turned away for various reasons linked to stigma and discrimination associated with abortions.

These findings suggest the need for improvements, and they call for the full and comprehensive implementation of the National Guidelines for Comprehensive Abortion Care [8], which would address some of the specific service gaps pointed out, especially within lower-level facilities that attend to a majority of PAC and SAC clients. To address training gaps and ensure the availability of PAC providers, there is a need to invest in in-service PAC training for mid-level health providers (such as nurses and clinicians) to ensure the provision of quality and safe services. This training would also address the confusion around hospital and ministry policies, as all signal functions for basic PAC should be available at all public health facilities. The need for Values Clarification and Attitudes transformation (VCAT) training sessions and workshops for both PAC and SAC cannot be overstated. Such training would mitigate the socio-cultural norms and religious leanings that may hinder the provision of PAC/SAC, positively shift provider attitudes, enhance respectful care, and improve the experiences of PAC/SAC patients as they seek services. Such training should also be accompanied by establishing robust referral mechanisms that address situations where provided conscientiously object to the provision of SAC [33]. This should also be accompanied by the strengthening of supply-chain systems for PAC commodities and supplies [34]. Further, putting more resources and investment into strengthening primary-level facilities has yielded greater dividends since about 75% of PAC cases are initially attended to in these lower-level facilities [35]. There is also a need to enhance referral systems to ensure women needing higher-level care can get the care they need.

This study also makes clear that legal termination of pregnancy is not a commonly provided service among health facilities in Liberia; this is expected, of course, as abortion is primarily restricted in Liberia. However, the proposed Public Health Law in Liberia may liberalize abortion and expand access to legal abortion in the public health facilities that form our sample. While the potential liberalization would pave the way for more SAC cases in health facilities and change current restrictions around policy, the barriers posed by the lack of equipment/supplies and the moral and ethical objections of healthcare providers would likely still impede access. As such, our study makes clear that any changes to the abortion law in Liberia must also be accompanied by; comprehensive guidelines that outline the responsibilities of health workers to provide or facilitate care, enhancements in supply chain systems and comprehensive training on VCAT.

Study limitations

While the study is the first to report data on the capacity of health facilities to provide PAC and SAC in Liberia, it is nonetheless not without limitations. First, certain key indicators (such as the availability and number of trained providers on PAC/SAC by cadre) were missing in the signal function survey, and this invariably hampered the comprehensive assessment of the health system’s capacity for PAC and SAC. The tool also lacked certain general measures of facility capacity, forcing us in some cases (e.g., first and second trimester SAC, blood transfusion) to use the questions asked about facility service provision in the last six months. In addition, the relatively small sample of health facilities that had complete datasets for the signal function survey also limits the extent to which we can extrapolate these findings to the whole of Liberia. The fact that 75% of facilities in our sample were located in the more developed and urbanized South-Central region means that our sample may capture better-resourced facilities than is representative of the country. As such, our findings may be biased towards better PAC and SAC capacity. The exclusion of specific critical indicators, such as staff availability and facility operation hours, is a crucial missed opportunity that deprives us of the ability to understand the functioning of facilities in Liberia and also the extent to which the health personnel workforce is available for PAC/SAC. This was inadvertent, and we recommend that future studies include these variables. Notably, the study sample size was fairly small and does not adequately represent the universe of public health facilities in Liberia. As such, the study findings cannot be extrapolated beyond the facilities targeted in the study. Nonetheless, our facilities sample is diverse with a range of facility levels, offering an idea of the current characteristics of health facilities by level for PAC and SAC. Future studies should attempt to collect data from a nationally representative sample of health facilities and include data points beyond just the structural indicators of PAC and SAC, but also include the patient care processes and clinical outcomes that can comprehensively inform the quality of PAC and SAC. These are particularly important as the Liberian government debates liberalization of the abortion law.

Conclusion

The study highlights the limitations to the provision of basic PAC and SAC among our sample of public health facilities in Liberia and the poor capacity of these health facilities to provide comprehensive PAC and SAC services in particular. This is an extremely important study in anticipation of possible liberalization of the abortion provisions within the public health law. Improving the availability and accessibility of the package of PAC and SAC services is critical to ensuring better maternal health outcomes. This requires strengthening the health systems through training of providers, better supply chains for PAC commodities, supplies, and equipment, as well as removing legal and policy barriers to accessing PAC and SAC services.