Descriptive statistics

The search generated 19,653 items for screening (Fig 1). After duplicate removal, the 16,918 remaining items were screened for inclusion on the basis of title and abstract. Where exclusion could not be determined on the basis of title and abstract, the authors screened the full text. Decisions were made in favor of an inclusive approach if questions remained. In total, 150 studies met all the inclusion criteria and were included in the scoping review.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Screening results.

PRISMA-ScR flow diagram illustrating the selection of sources of evidence included in the review.

https://doi.org/10.1371/journal.pone.0237227.g001

Studies came primarily from Africa (n = 31) and Northern America (n = 35) (Table 3). Interestingly, fewer studies focused on countries in Asia (n = 19) even though a large share of the world’s population resides there. Noticeably absent are studies from the Middle East and North Africa. After the United States (n = 32), the United Kingdom (n = 15) and Mexico (n = 7) were most represented in the final inventory.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Included studies by region and country.

https://doi.org/10.1371/journal.pone.0237227.t003

The majority of studies were quantitative in nature, with 75 studies relying exclusively on quantitative methods and another 39 studies including both quantitative and qualitative methods (Table 4). Studies most often occurred in countries classified as high-income (44.0%), followed by lower-middle-income countries (17.3%) and upper-middle-income countries (16.0%). The level of geographic coverage varied with nearly one-third of studies examining abortion outcomes within health facilities and the remainder being conducted at the national (n = 31), sub-national (n = 42), local (n = 1), and other (n = 28) levels.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Characteristics of included studies (n = 150).

https://doi.org/10.1371/journal.pone.0237227.t004

Studies most frequently reported the costs of abortion provision and policies within health systems and communities (n = 116), followed by mesoeconomic impacts (n = 40) and mesoeconomic benefits and value (n = 27). To facilitate the analysis, studies on mesoeconomic benefits and values are tabulated jointly. The total number of data points (i.e. reported mesoeconomic outcomes) (n = 184) exceeds the number of studies in the final mesoeconomic inventory (n = 150), since some studies presented multiple mesoeconomic outcomes. Each of the economic themes that emerged from the synthesis of data is discussed in further detail below.

Costs of abortion provision and policies

Approximately three-quarters of studies present data on mesoeconomic costs to health systems (S1 Table). Across these studies, four themes emerged: (1) within health facilities and health systems, the costs of providing abortion services vary greatly, particularly given the range with which services are identified and costed; (2) financial savings can be realized while maintaining or even improving quality of abortion care services; (3) insurance coverage of abortion services is far from universal; and (4) adapting to changes in laws and policies, particularly targeted regulation of abortion provider laws, is costly for health facilities.

(1) Within health facilities and health systems, costs of providing abortion services vary greatly, particularly given the range with which services are identified and costed. Numerous factors influence the costs of providing abortion services in health facilities: the gestational age at which the abortion occurs, the method of abortion (e.g. whether or not anesthesia is required, whether the patient or hospital staff administered medical abortion), the supplies and equipment used, whether contraception is provided, diagnostic tests, management of potential complications, the number of routine follow-up visits required, the number of missed appointments (“no shows”), caseload, frequency of procedures, overhead costs of a hospital stay, the legal environment, regulatory fees, facility type, and facility location [2, 18–28].

The cost incurred by each of these components may vary greatly even within the same health system. In Mexico, the average cost of treating severe abortion complications in public hospitals ranged from US$ 601 to beyond US$ 2,100 [29]. Costs to health facilities are higher in locations with patchwork abortion laws. For example, abortion reforms in Mexico’s Federal District led women from other states to come seek legal abortions there; as a result, facilities providing abortion services in the Federal District experienced increases in workloads and costs [26].

While patients can influence the cost to health facilities of providing abortions through factors such as increases in workloads and missed appointments, the health facility costs of providing abortion services also influence patients. The increased costs of providing a second-trimester abortion compared to a first-trimester abortion require clinics to charge higher fees for the procedure, thus potentially delaying or restricting women’s access to abortion services [30–32]. It is noteworthy that in many areas in the United States, median charges for abortion remained relatively constant over time; this finding may represent providers’ commitments to maintain abortion access and keep services affordable [24].

Among studies that costed or reported costs of abortion services, significant variation exists in how studies report the costs of abortion to health facilities (Table 5) and health systems (Table 6). Studies report on the costs of different methods of abortion (e.g. manual vacuum aspiration, dilation and curettage, medical abortion, surgical abortion), the costs of abortion in specific locations (e.g. in a nurse-led clinic, in a tertiary hospital), and the type of abortion service (e.g. abortion at 20-weeks gestation, post-abortion care, treatment for septic abortion, treatment of incomplete abortion). Some researchers specify all of the direct and indirect components that are involved in the delivery of abortion services and cost each of these components, making it easy to understand how the total cost was derived [33]. Other researchers do not. Some researchers report the costs for an abortion without specifying the type of abortion (e.g. medical, surgical, early/late gestational age). Other researchers are quite specific about the type of service being costed (e.g. manual vacuum aspiration for treating incomplete abortion patients), especially researchers who reported costs for different pilot programs in the United Kingdom [34–36].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Costs of abortion services in health facilities.

https://doi.org/10.1371/journal.pone.0237227.t005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Costs of abortion services in health systems.

https://doi.org/10.1371/journal.pone.0237227.t006

In settings where individuals seek unsafe abortions, immense health system resources are required to treat complications resulting from unsafe abortions [2, 3]. Public health services often pay for treatment of complications at tertiary hospitals, where costs are greatest [4, 5]. For example, in Ouagadougou, Burkina Faso, the average cost of treating abortion complications in a tertiary hospital (US$ 51.09) was 40% higher than the cost in a secondary-level hospital (US$ 36.50) [5]. In Nigeria, while clinic-based manual vacuum aspiration and hospital-based manual vacuum aspiration are equally effective, hospital-based manual vacuum aspiration is the more costly option [62]. When possible, shifting to the provision of simpler procedures to lower-level hospitals or primary clinics would thus minimize costs and lead to financial savings among health systems.

(2) Financial savings can be realized while maintaining or even improving quality of abortion care services. Reducing costs to health facilities and health systems does not mean reducing the quality of care. For example, without sacrificing quality, it is possible to reduce costs by choosing a local generic abortifacient over an imported one [63]. Health systems can also reduce costs by shifting the location in which abortion care services are provided. In Mexico, shifting from hospital-based manual vacuum aspiration to clinic-based manual vacuum aspiration would save the health system nearly US$ 100,000 per 1,000 women [64]. In the United Kingdom, shifting outpatient consultations to a nurse telephone clinic service could reduce costs per patient by 18% while simultaneously providing patients with greater convenience [35]. Using a local anesthetic outpatient clinic reduced use of the operating room to one streamlined abortion patient list per week, thus generating approximately GB£ 60,000 in annual savings [65]. Health facilities can even improve service delivery whilst reducing their costs of service delivery. In two Latin American study sites, reorganizing services to an outpatient basis resulted in reductions in the average length of patient stay and significant financial savings that facility administrators were then able to pass on to patients via fee reductions [53].

With treatment of post-abortion complications consuming a significant proportion of health system resources in many settings [66], decentralizing services and legalizing abortion can produce substantial financial savings. Financial models from Uganda show that moving from a conventional, legally restrictive setting to a planned, decentralized, legally permissive setting decreased the mean cost per unsafe abortion complication case by 86% from US$ 45 to US$ 6 [11]. Offering menstrual regulation, a method of fertility control, care can also reduce the cost of treatment for severe abortion complications, since menstrual regulation care costs 8–13% of the cost of treating severe abortion complications [67]. When feasible and appropriate, service delivery for post-abortion care can be moved from an inpatient basis to an outpatient basis. Outpatient manual vacuum aspiration for post-abortion care is less expensive than inpatient dilation and curettage, according to findings from multiple studies in Africa and Latin America [41, 68].

Given that clinical staff costs tend to be highest for physicians, facilities in which abortion services are provided by a range of providers (e.g. nurse practitioners, physicians’ assistants) were able to keep staff costs low without sacrificing time with patients [56]. In the United Kingdom, a nurse-led service for termination of pregnancy could maintain the clinic while saving nearly 40% in costs annually compared to a medically-led clinic with a staff-grade doctor [49]. However, task shifting has logistical and financial costs that can raise health faculties’ expenses in the short-term [69]. Removing administrative barriers that limit trained midwives’ ability to provide routine post-abortion care services and implementing a policy on task shifting could free up physicians to focus on more complicated cases [41]. Yet, the setting (e.g. rural facility) and context (e.g. supply chain, sustainability) must be considered when assessing the feasibility of task shifting. For example, low-volume, low-income nurse midwives in Ghana reported less sustainable access to manual vacuum aspiration than physicians [39].

Costs of providing medical abortion and surgical abortion can be quite similar in some settings, like China [70]. As a result, abortion providers do not need to factor in remuneration when recommending a method to patients. In other settings, however, the costs of providing medical and surgical abortion differ. Surgical abortion in the British National Health Service cost more than medical abortion due to inpatient standard costs and used 8% more resources than medical abortion [51, 71]. Costs and savings from surgical abortion techniques also vary by setting. Numerous hospitals in Asia and Latin America found manual vacuum aspiration to produce lower costs to health facilities than sharp curettage [25, 45]. In El Salvador, a manual vacuum aspiration procedure and hospital stay (US$ 54) cost 13% less than sharp curettage and reduced patients’ time in hospital by 28% [25]. In some African countries, the reverse was found: dilation and evacuation was less expensive than medical abortion in South Africa [72].

(3) Insurance coverage of abortion services is far from universal. Health insurance and public health systems effectively disincentivize abortion outright or disincentivize certain methods of abortion, particularly medical abortion. Because Australian insurance companies that provide medical indemnity initially assessed the risks of medical and surgical abortions to be the same, physicians considering incorporating medical abortion into their practices experienced a prohibitive increase in their medical indemnity insurance [73]. Once this issue was resolved in 2014, medical abortion access in rural areas improved. In some countries, including Germany and the United Kingdom, the public health systems have not covered the costs to physicians and health facilities of providing medical abortion using Mifegyne (RU 486) [74]. When combined with a requirement for special licenses to offer medical abortion services and the high costs of purchasing Mifegyne, the economic disincentives resulted in surgical abortion becoming the more economic method [74]. These decisions limit patients’ access to the full range of abortion care services.

In the United States, private insurance coverage of abortion is restricted in some states, and not all clinics accept third-party payers [75]. When subsidized or public insurance programs like Medicaid exist, staff members at abortion facilities may play an important yet underutilized role in helping patients enroll [76]. Even when women can obtain Medicaid coverage of abortion, providers are not guaranteed reimbursement for qualifying abortions. In six states where Medicaid coverage of abortion is limited to cases of rape, incest and life endangerment, providers were reimbursed for only 41.6% of abortions that should have qualified for Medicaid reimbursement [58].

Policies to implement safe abortion services in the public sector can backfire if provider reimbursement is insufficient. Following passage of the Safe Abortion Service Guidelines of 2016 in Nepal, provider reimbursement for free abortion services under the national government scheme was less profitable than abortion services in which clients paid out of pocket [77]. Financial incentives could motivate providers to perform abortions in the private sector rather than the public sector, thus reducing access to abortions in public facilities.

(4) Adapting to changes in laws and policies, particularly targeted regulation of abortion provider laws, is costly for health facilities. Following implementation of a new targeted regulation of abortion provider law in Texas, the cost of providing an abortion increased: from 2001 to 2006, the cost of providing an abortion at 20 gestational weeks increased by 37% (US$ 454) compared to other states [78]. Laws can also target the cost of providing a particular type of abortion. In Texas, House Bill 2 increased the cost of medical abortion in most facilities by requiring providers to choose between a drug regimen that was considerably more expensive than the evidence-based regimen or a drug regimen supported by limited evidence [79].

Impacts of abortion provision and policies

Among the 40 studies that reported economic impacts of abortion at the mesoeconomic level (S2 Table), five themes emerged: (1) limited resources negatively affect health facilities’ ability to meet client demand and to provide quality abortion care services; (2) American policies on abortion economically impact health systems and facilities both domestically and abroad; (3) a disproportionate amount of health facility resources are required to provide post-abortion care; (4) switching abortion techniques in health facilities can have positive or negative economic impacts, depending on the reason behind the change; and (5) complex health systems impact the abortion-related trajectories of pregnant people.

(1) Limited resources negatively affect health facilities’ ability to meet client demand and to provide quality abortion care services. Studies examining the economic impact of limited resources for abortion focused on the impact of staff shortages, insufficient physical resources, low Medicaid reimbursement rates, and costly regulations. Shortages of trained staff affected patient care in countries of all income levels. In France, a hospital lacked trained staff to meet patients’ needs for psychological help [80]. In Kenya, gaps in capacity resulted in longer hospital stays and increased costs of care to the health system [81]. In Malawi, staff shortages resulted in one or two nurses covering up to 70 patients, the discontinuation of manual vacuum aspiration, and limited on-the-job training [82]. A large and increasing proportion of unsafe abortion admissions can stretch human resources even further, leading health care staff to likely prioritize the easier option over the patient’s best interests [82].

Shortages of physical resources, such as manual vacuum aspiration instruments, led health care providers in Malawi to automatically resort to dilation and curettage even though dilation and curettage is more expensive, is slower and more difficult to perform, requires more staff, and results in more in-patients [82]. Medication stock-outs due to poor supply chain management resulted in women in Nepal being denied legal abortions [77].

Financially, rates of reimbursement vary. However, some providers in the United States experienced Medicaid reimbursement rates so low that the rates would put providers out of business: one provider was reimbursed only US$ 212 for an abortion that cost US$ 420 [58]. To minimize the financial impact of Medicaid’s insufficient reimbursement rates, providers depended on abortion funds, minimized their work with Medicaid, or avoided Medicaid outright by absorbing up to US$ 60,000 annually in free or reduced-cost services [58].

Due to costly regulations, existing health facility resources may be stretched to the point where facilities must accommodate a smaller clientele or close their doors. In the United States, requiring abortions to be performed in ambulatory surgical centers would cost a facility over US$ 750,000 to comply with the regulations’ physical renovations; as such, these costly requirements may result in providers stopping abortions or significantly raising the prices so that quality abortion care is beyond some women’s reach [75].

(2) American policies on abortion economically impact health systems and facilities both in the United States and abroad. Domestically, abortion legislation hinders or prevents health providers from performing abortions by influencing the demand of abortion care and increasing the costs of care. The implementation of House Bill 2 in Texas led clinics to close and increased the average clinic service population from 150,000 clients to 290,000 clients [83]. This increase in service population was felt especially by urban clinics. In Dallas-Fort Worth, the average service population increased by 250,000 clients [83]. As a result of clinic closures, the remaining clinics were unable to meet demand for abortion services and deliver the same quality of care as before the law [84]. In North Carolina, abortion providers attempted to minimize the burden of new regulations on patients by implementing telephone counseling in place of two in-person clinic visits [85]. Telephone counseling required significant adaptations, including hiring additional nurses, developing call-center infrastructure, and changing scheduling and work tasks. By explicitly deciding to absorb the financial burden of these changes rather than pass them on to patients, providers worked more hours uncompensated or fundraised to support the increased costs [85]. Clinics are occasionally unable to absorb the financial burden of abortion regulations, particularly when regulations require costly physical renovations. In these cases, providers may stop performing abortion or may increase their prices beyond some clients’ reach [75].

Abroad, the economic impact of American abortion policies extends beyond abortion care. Funding cuts due to financial restrictions commonly referred to as the Global Gag Rule resulted in disruptions to the provision of sexual and reproductive health services in numerous countries, and clinic closures affected hundreds of thousands of underserved clients whose needs were not met by public health services [86, 87]. In addition to staff terminations and clinic closures that may particularly affect rural clinics, the Global Gag Rule resulted in the termination of contraceptive supply shipments in 29 countries and programmatic cuts (e.g. to maternal health and well baby care, youth outreach, HIV/AIDS prevention) [77, 86–88].

(3) A disproportionate amount of health facility resources are required to provide post-abortion care. Studies agreed that the costs of post-abortion care were a burden to health systems, requiring a disproportionate amount of health facility resources and stretching overstretched health system resources [4, 38, 53, 89]. Some studies reported the use of physical resources for post-abortion care, such as up to 50% of hospital gynecological beds in Bangladesh or 25% of the main hospital’s blood supply before Guyana liberalized its abortion law [4]. Other studies reported financial costs of post-abortion care: more than seven times the Tanzanian Ministry of Health’s annual per capita budget [4] or more than half of facilities’ budgets for obstetrics and gynecology [53]. Estimates of the annual health system costs of post-abortion care in Africa and Latin America are considerable, ranging from US$ 159 million to US$ 476 million [38].

To reduce the number of patients seeking post-abortion care, it is cost effective to expand contraceptive services and supplies and to provide menstrual regulation and related care. Data from Nigeria show that if health systems provided contraceptive services and supplies to prevent unintended pregnancies that result in unsafe abortions, the cost-benefit ratio would be US$ 4 to US$ 1 when compared with the cost of post-abortion care [68]. In Bangladesh, menstrual regulation prevents unnecessary and expensive complications from unsafe abortions, with each case costing the health system a fraction (8–13%) of the cost of treatment for severe abortion complications [67].

(4) Switching abortion techniques can have positive or negative economic impacts on health facilities, depending on the reason behind the changes. Switching abortion techniques (e.g. performing abortion in ambulatory surgery centers, mandating multiple clinic visits) may result in negative economic impacts on health facilities when the changes result from laws designed to make abortion more expensive and difficult to provide [75, 85]. Other times, directives aimed at improving one aspect of abortion services, such as minimizing complications, may negatively impact a different aspect of abortion services. For example, by requiring abortion in Moldova to be performed by obstetrician-gynecologists in hospitals, hospitals had to centralize their services that resulted in higher costs of care and decreased client access to abortion [90].

Changes in abortion techniques can also facilitate financial savings and the expansion of abortion services. In Nepal, medical abortion has expanded access to first-trimester abortion in rural areas, since it does not require facilities to have capacity for surgical abortion [77]. In Sweden, a randomized-controlled equivalence trial found provision of medical abortion by nurse-midwives superior to provision by physicians: provision by nurse-midwives was more efficacious and less expensive [47]. In Malawi, manual vacuum aspiration delivered cheaper (i.e. financial and human resources), easier, and faster abortions (i.e. time to perform the abortion and time to discharge) than dilation and curettage [82]. In Latin America and the Caribbean, however, dilation and curettage produced similar outcomes to manual vacuum aspiration; a reduction in opportunity costs and an increase in service efficiency happened only after the reorganization of services [53].

(5) Complex health systems impact the abortion-related trajectories of pregnant people. Health system requirements and responses to new abortion methods impact the pathways of pregnant people. Where misoprostol is sold by prescription only in Latin America, pregnant people seeking the medication must pay for prescriptions, lie to providers about intended use, ask a third-party to obtain a prescription, or seek medication through unregulated channels [91]. In Brazil, inflated commercial values of misoprostol and price fluctuations dependent on facility type encourage non-facility-based procurement with reported prices as high as US$ 144 despite the mean cost of misoprostol being US$ 6 [92]. A study of Caribbean states found that hospital-based gynecologists sought to control the provision of Cytotec (misoprostol), whilst abortion providers admitted that the cost of abortions led them to recommend pregnant people buy medications from pharmacies and return to them in case of complications [93]. Permitting conscientious objection in Mexico has led to suggestions that provider objections were based on financial incentives, with providers objecting in public facilities but providing abortion care in their private facilities [94]. Where health systems regulations interfere with access to abortion, pregnant people face more complex pathways to obtaining care, including pathways that might increase potential associated risks like seeking medical abortion from unlicensed providers.

In reviewing data on studies that reported economic impacts of abortion at the mesoeconomic level, economic impacts of abortion on communities are rarely studied. The paucity of evidence on this topic does not align with the frequency that abortions occur in communities or the potential economic impact of unsafe abortions on communities. Two studies (2/150) reported the economic impact of abortion on communities, broadly speaking. In New Brunswick, Canada, a regulation from 2015 insufficiently mitigated challenges resulting from the province’s refusal to fund abortion care in health clinics [95]. Researchers found that even if the province completely eliminated the two-physician requirement, the resulting impact on access to affordable, timely abortion care would be marginal [95]. In India, evidence from a study on prenatal sex selection found that regions with an increasing male-female sex ratio experienced an increase in families’ parental education, an increase in maternal age at first birth, and a decrease in the likelihood of rural residence [96].

Benefits and value of abortion provision and policies

Twenty-seven studies reported economic benefits (n = 17) or value (n = 11) of abortion at the mesoeconomic level (S3 Table); one study reported both benefits and value, so the number of data points exceeds the number of studies. Across the studies, two themes emerged: (1) the benefits of abortion differ by method; and (2) health systems often, but not always, respond to women’s needs around abortion.

(1) The benefits of abortion differ by method. Multiple studies compare the benefits of surgical abortion methods to medical abortion methods. The health system benefits, often reported in terms of cost and efficacy, vary depending on where the study occurred [45]. In Nigeria, manual vacuum aspiration in clinics was the least expensive and most effective option relative to unsafe abortion, whereas in Ghana, medical abortion was the most cost-effective option relative to unsafe abortion [62]. In the United Kingdom, replacing surgical abortion procedures in the operating room with manual vacuum aspiration resulted in an annual savings of GB£ 60,000 [65]. In China, 33.1% of rural abortion providers and 38.1% of urban abortion providers reported that an advantage of medical abortion is that it is less expensive than surgical abortion, and medical abortion is also a low-risk alternative without surgery and anesthesia [97].

Within-method benefits can be further enhanced by task shifting, service reorganization, and streamlining. In Sweden, task shifting of medical abortion from physicians to nurse-midwives produced direct economic benefits: shorter time providing care, lower salaries among nurse-midwives than physicians, shorter waiting times among patients, and a potentially lower cost to treat complications [47]. Reorganizing services by treating clinically eligible patients in the manual vacuum aspiration unit rather than the main operating room could free up needed hospital beds by shortening patients’ stays and freeing up physicians and anesthesiologists for other cases [98]. Streamlining the protocol for medical abortion to two or three clinic visits rather than four visits, as required in Mexico City in 2005, would reduce the costs of medical abortion without affecting patient safety [64].

Of course, providing safe abortion instead of unsafe abortion is the most important factor to improve health outcomes and economic outcomes [29, 62]. Reducing unsafe abortion by increasing access to safe abortion can reduce the costs of care and benefit health systems. In two public referral hospitals in Ouagadougou, Burkina Faso, the hospitals would have saved over US$ 19,778 in 2010, had safe abortion care services been available [5]. In Mexico City in 2005, a modest increase in access to manual vacuum aspiration would improve access to safe abortion and result in savings of over US$ 50,000 among the three public hospitals and one private clinic being studied [29].

(2) Health systems often, but not always, respond to women’s needs around abortion. In settings where policies and laws do not restrict discussions about abortion to clients and providers in facilities providing abortion, a range of health providers can respond to women’s needs for pregnancy options counseling and referrals to abortion clinics [76]. By offering medical abortion patients alternative forms of follow-up care (e.g. phone calls, internet communication), health systems may support patients’ preferences and needs while simultaneously experiencing potential cost-savings resulting from patients being less likely to seek follow up care from emergency departments [99]. In Mexico, liberalizing abortion laws to permit abortion due to rape resulted in a statistically significant increase in women accessing induced abortion in the public sector [100]. Responding to women’s needs around abortion includes expanding access to health insurance that includes abortion coverage. In the United States, state Medicaid coverage of medically necessary abortion reduced the risk of severe maternal morbidity by an average of 16% [101].

At times, the health system fails to meet women’s needs around abortion. Access barriers may prevent women from obtaining legal abortions [102]. Existing abortion services may be inappropriate or even harmful, lack appropriate referrals, and deliver limited knowledge on contraceptive methods to post-abortion patients [103]. Providers’ values may result in a double discourse around abortion. Providers who consciously object to providing abortion in the public sector may willingly overlook these public values to perform abortions in the more profitable private sector [26].