Study design and setting

The study on which this article is based was cross-sectional, using in-depth interviews to collect qualitative data. Public and private “high volume” healthcare facilities, including general and referral hospitals at Levels 4, 5 and 6, were selected, based on the number of PAC cases reported in the facilities in the most recent data [7,49,50]. Public facilities are owned and run by the government, and private facilities are managed by individual proprietors, non-governmental organizations or faith-based organizations. In Kenya, the Ministry of Health, under the Kenya Essential Package for Health (KEPH), defines six levels of preventive and curative health services, both for public and private facilities, ranging from level 1–6. Level 1 is the lowest level, and forms the foundations of service delivery at the community level; it includes village health committees and community health units. Levels 2 and 3 (dispensaries, health centres, and maternity/nursing homes) offer promotive, preventive, and curative services. Levels 4 and 5 (primary, secondary and tertiary hospitals) offer curative and rehabilitation services, with a limited number of preventive/promotive care programs, while level 6 is the highest level of care, comprising the national referral hospitals. Sampling for this study was purposive to include only facilities that managed a high number of PAC cases, so allowing us to recruit sufficient cases over a short period (one week for data collection) when trained qualitative interviewers were able to work at each facility.

Sampling

Sixteen hospitals in three regions − Nairobi (5), Central (5) and Eastern (6) − were sampled purposively by regional area, level, and reported quality of care. These facilities were classified into “high,” “medium” or “low” quality [37] to ensure representation of facilities at different levels of care and to gain a wide view of provider and patient experiences according to the facility level. This categorization was based on data from an earlier study from which these facilities were sampled. The study used the essential elements of PAC framework to categorize all participating facilities into three categories above (for details, see Mutua et al., 2017)). To achieve greater diversity and in-depth opinions on care from patients we purposively selected patients from a sub-sample of six of the 16 facilities, two from each of the three levels of quality stated above. This selection was mainly based on facility distance from Nairobi and size, with the intention to focus this selection on high-volume facilities able to provide sufficient numbers of interviews within a short period. Five patients in each of the six facilities were recruited and interviewed, together with one service provider at each of the 16 facilities. This yielded a sample of 30 patients and 16 provider interviews. However, due to a low number of patients in some facilities, only 21 patients were interviewed of the targeted 30 patients. All service providers were interviewed: of the 16 interviewed, eight were nurses and eight clinical officers.

Data collection

Two interview guides were developed and reviewed by the lead author and all co-authors. The guides were structured specifically to encourage discussions with service providers and patients on their understanding of the different components of the quality of PAC framework. The service provider interviews lasted on average 34 minutes while patient interviews lasted 21 minutes. Both provider and patient interviews were conducted in a private room in the health facility to ensure maximum privacy of information shared. All interviews were conducted either in English or Kiswahili, and transcribed and translated into English if the interview was conducted in Kiswahili. Field data collection was conducted between May and July, 2017.

Data

The main PAC service providers in selected facilities were recruited and interviewed by one of a team of three experienced and well-trained qualitative field researchers (all female). Women who were treated for PAC in selected healthcare facilities were interviewed after discharge, prior to their return home. The principal investigator (first author, PI, male) trained and worked with the field researchers to ensure that they had a good understanding of the study objectives and were mindful of its ethics. Interviews were adapted throughout the process through a constant review of data collected. All interviews were audio-recorded, and quality control and improvement were ensured through listening to recordings and discussions for providing feedback to interviewers. Once each interview was completed, it was forwarded electronically to a central office and archived on a password-protected computer, accessible only to the PI. Once all data were received, a transcriber worked with the PI; she was well trained to understand the survey objectives and the ethics of handling sensitive data. All data were transcribed into Microsoft word and as required, translated from Kiswahili into English. All patient and provider identifiers were replaced with codes to minimize the risk of identification of patients during data coding and analysis.

Survey ethical consideration

Ethics approval was obtained from the Ethical and Scientific Review Council of African Medical Research Foundation (AMREF), the Kenyatta National Hospital (KNH) and University of Nairobi (UoN) Ethical Review Committee (ERC), and the Human Research Ethics Committee (Medical) at University of the Witwatersrand. Survey approvals were also obtained from the Kenya National Commission for Science, Technology, and Innovation (NACOSTI) and the Ministry of Health. Written informed consent as approved by AMREF were obtained from each participant before any interview was conducted. All minors who were already securing PAC services were deemed to be independent and able to provide their own informed consent, without parental consent.

Analysis

Data were analyzed deductively, using a predefined set of themes from the essential elements of PAC framework. Additionally, inductive analysis was used by identifying any additional themes on quality of care at the facility as described by the providers and patients. A double coding approach was used to compare independent codes between a trained and experienced coder and the PI, and an agreement was built on the codes before final coding in NVIVO. Final analysis involved in-depth interrogation of the data for both visible and underlying meanings assigned to quality of care from the perspectives of the service providers and patients. Specific verbatim quotes were extracted from the data, which express respondent views concerning certain propositions of quality care. Below, we identify people only by broad class (provider, patient) when we quote them directly.

The constitution and the law

For fear of legal repercussions, patients often fail to disclose their abortion history, withholding information from providers on the method used to induce an abortion, for instance, and estimated gestational age. Under such circumstances, especially when physical examination is also misleading, inadequate disclosure can result in inaccurate management and impedes the use of appropriate technologies, with potential risks to the woman:

In the process of taking the history, there is another challenge because a lady can decide not to tell at which gestation stage the child was. You can say my last menstrual period (LMP) was the previous week yet you find at the expulsion of the fetus you see the fetus was already developed. (Provider, level 4, private, low quality facility)

How policy is derived from law, and how it is translated

In 2016 changes in law led to the registration and inclusion of Misoprostol and Mifepristone as essential drugs for the management of obstetric and gynecological indications, resulting in increased access of these as abortifacients. This significantly reduced the use of crude means of abortion termination (such as the insertion of a sharp object, ingesting concoctions or heavy massage). However, some service providers found challenges with the unregulated access of these drugs, often without proper controls on drug prescription and usage:

You find that abortifacients could not be sold in chemists without a doctor's prescription. Nowadays they are very available, even a standard 8 girl can go to town, purchase the abortifacients, come here bleeding in shock … the drug is now an over-the-counter drug. (Provider, level 5, public, high quality facility)

The organization of healthcare services from national to county to facility level poses additional restrictions to quality PAC access. Processes such as the acquisition of appropriate equipment and medication, integral to quality PAC, are managed by healthcare administrators who often lack requisite medical skills:

You are dealing with someone who is not a medical person in the procurement department. You are dealing with a matron that is not interested maybe” (Provider, level 4, public, low quality facility).

There was a lack of guidelines on the provision of PAC and legal abortion care in all facilities and at all levels. In the absence of such clarity, some service providers opted to provide safe abortion services at night, but during the day, these cases would be recorded and managed as PAC cases. Service providers emphasized the need for interventions by county and national health management officials to provide effective guidelines for abortion care and PAC; guidelines, they felt, would improve their confidence to provide PAC, because they offer clarity in relation to the law:

I think in the county we have the RH (Reproductive Health) department. They are doing a good job but mostly they are concentrating on the family planning. In the county government, the RH department should really concentrate on how to reduce cases of aborting and on how to increase awareness on PAC. (Provider, level 4, private, medium quality facility)

Service providers acknowledge the importance of collaboration between different RH actors. Despite the cost implications of such ventures, service providers identified the need for service collaboration between county governments and other RH stakeholders to support identifying and addressing existing quality gaps in PAC services.

Delayed care due to limited facility capacity

Lack of requisite capacity to offer quality PAC in facilities resulted in women receiving “partial referrals” for certain services such as radiological examinations and surgical obstetric or gynecological care. These referrals extended the duration of treatment and increased the risks of excessive bleeding and untreated infections, causing concern among patients and service providers alike. Lack of ultrasound equipment in most level 4 facilities, for example, meant that patients were regularly referred to other facilities for obstetric or abdominal scans, after which they were then to present back to the referring facility for further management. The need to go to a different facility for a scan led to delayed intervention and increased the risk of severe complications. As one service provider in Central Kenya emphasized, less time was lost when patients were referred to a single facility for both the scan and completion of treatment. However, even when this was possible, in some cases patients were asked to return to the referring facility because of heavy workloads and lack of space in the larger facilities:

Our nearest referral hospital is [Facility A] level 5, where they usually can go and first do the scanning, if they wish to complete their procedures there they can but some of them due to the workload in [Facility A], after scanning they usually come back here, so we do still the MVAs here. If it is septic we have the wards and thus we admit them. (Provider, level 4, public, high quality facility)

Patient referral from smaller facilities was common. In the larger facilities, patients could wait overnight to be seen by a doctor. For example, one patient was referred to a facility, having already expelled one of a twin fetus; the second fetus was only partly expelled. Overnight, this patient remained under ineffective pain killers, awaiting a doctor to perform an evacuation.

In smaller level 4 facilities, patients often needed specialized surgical interventions, which only a specialist gynecologist could perform. Yet a referral was often not feasible, due to the long distance between facilities and transfer costs. Some of these facilities lacked ambulances with advanced life support (ALS) to ensure patients of a seamless transfer to a level 5 hospital. Even for facilities with ALS ambulances, certain complications such as postpartum hemorrhage and ruptured uterus could still result in death of the woman while in transit.

Patient referral and transfer gaps were evident, with providers experiencing difficulties to identify the right facility to which they could refer patients. This caused delays, leading to severe complications and occasionally the death of a patient:

You might find that when you are referring a patient there could be some difficulties you encounter before the patient can be accepted in the facility where you are referring her. So they should try and harmonize in terms of linking to the system. ~ (Provider, level 4, public, medium quality facility)

Infection control prevents sequelae from unsafe abortions and must be controlled to prevent future pregnancy loss, and pain medication is important to ensure patient-centered care. However, lack of infection-control medications including antibiotics was common, as was the lack of analgesics. Patients were often referred to facilities such as local chemists to purchase these medications; others endured painful procedures without effective analgesics. Other health facilities referred patients because of the lack of blood or blood products, or lack of capacity or equipment to cross-match blood or transfuse. Commonly, there were insufficient instruments and lengthy instrument sterilization procedures, which with high numbers of patients hindered service provision and increased the number of referrals.

Privacy is essential for a conducive environment for patients to openly discuss their condition, for informed management of their complications, and to discuss their reproductive history and aims. However, patient admission facilities were often insufficient, resulting in patients having to share beds, premature patient discharge, and referrals to other facilities. In addition to infringements of patient privacy and right to confidential treatment, there was evidence of lack of capacity by lower level facilities to manage the number of cases that presented for PAC at these facilities. One patient had been admitted to three other facilities before a fourth admission at a level 5 hospital, and by the time she was admitted, she was unconscious. Many facilities also lacked sufficient evacuation procedure rooms, especially when a service provider was required to manage more than one patient at a time. As a result, PAC procedures were often conducted in maternity wards, despite concerns from providers of the potential psychological distress to patients by combining patients who had lost pregnancies with those who had successful births.

Technical performance

Lack of clarity on abortion law, and lack of care guidelines, impacted negatively on the capacity of service providers to provide quality PAC. Some patients were critical of the quality of support they received from staff at health facilities in relation to surgical and medical care, counseling and family planning advice. Repeat procedures, often the result of incomplete evacuation due to inadequate provider skills, were common. According to patients, these “revacuations” were mainly due to provider’s inability to make a specific diagnosis, and this increased the costs of care to both the patient and the health system.

Staff turnover, due to transfer of trained personnel between facilities, also impacted on quality PAC. While transfers can improve the capacity of other facilities to offer PAC, within facilities, necessary transfers demotivated other providers and caused confusion among patients. According to service providers, skill improvement lacked a focus on the capacity to prevent patients’ risk of repeat abortion:

To many people, the knowledge is limited to just provide what it takes to take this fetus out and then the other things is not important. The only important thing is that the pregnancy is out. For me, there is that gap. I think it is a real discrepancy because tomorrow they will go get pregnant again and like that. (Provider, level 4, public, low quality facility)

Interactions between women and service providers or staff

Uncertainty around the legality of abortion often led to patient discrimination at healthcare facilities. Service providers often condemned abortion and discriminated against women who secured abortion, and did not always support patients’ healthcare needs. In some cases, they designated duties such as within hospital transfers to patients or their caregivers, causing delays in care provision. Providers were often unsupportive and unresponsive to the patients’ needs, and failed to treat them with courtesy. According to patients, providers unreasonably withheld guiding them through the facilities. One patient described how two or more patients used one stretcher, without measures to avoid infection, equally exposing patients and caregivers to infections:

Yes. Then we came with the stretcher, they returned us and told us that it wasn’t the stretcher for casualty, but by good luck there was a patient that was being discharged and that is when we took the stretcher; it was just by chance. (Patient, level 6, public, high quality facility)

Some providers also spoke of disrespect to patients (from providers other than themselves) and argued the need for positive attitudes by health providers to encourage timely seeking of PAC services. Young women were particularly affected by negative attitudes, from providers who believed that they should not have been pregnant in the first place. These attitudes have also been associated with the high incidence of unsafe abortion in the country, and severe complications when women opt to self-medicate complications from unsafe terminations. In addition to age, discrimination occurred on the basis of marital status, with service providers likely to conclude that an abortion was induced if a patient was unmarried. Unmarried patients were therefore more likely to lie about their marital status, in order to gain access to the same level of care as provided to married patients.

Women are discouraged from accessing care and delay doing so

Quality care requires the continued availability of PAC services, and any service disruption can lead to life-threatening complications or result in lifetime consequences on patients’ reproductive and psychological health. During the study period, however, there were recurrent service disruptions due to industrial disputes: a national doctors’ strike lasted over three months; a national nurses’ strike for six months; and a clinicians’ strike for one month. Holidays and weekends, when the few available providers were off-duty, further increased the number of days when services were unavailable. In addition to delays in seeking care occasioned by legal challenges, as discussed earlier, other forms of service unavailability and stigmatization influenced general patterns of healthcare seeking. Women were aware of service disruptions, and they often presented to public facilities only when their conditions were too grave for further self-management and when they could not afford private facilities. Furthermore, most PAC cases in Kenya are treated in public facilities [7]. Even during the strike period, most public facilities always had nurses on duty, but care was delayed when the nurses were unable to perform a particular procedure or the procedure was not the nurse’s responsibility as set out in the service guidelines and Ministry of Health regulations:

Because if you come here at night, there is no doctor, there were just nurses, there are some things that the nurses can’t do like that one of mine they were telling me that I had to wait for the doctor, what if it was something serious that could kill me? (Patient, level 6, public, high quality facility)

Can women afford to access PAC?

The essential elements of the PAC framework do not explicitly clarify cost of care, whether direct or indirect. In Kenya, healthcare services are not free; they are subsidized in public facilities but not in private facilities. In cases of emergencies at night, or during weekends and holidays, most primary care public facilities are closed, and patients travel for long distances to reach an available secondary and tertiary facility. In such cases, these patients would seek care in private facilities, but these services are not generally affordable. Patients who could not pay for the services in private facilities would be referred to public facilities, while those in public facilities would miss certain services or procedures until they can raise funds to pay for these services. Others were discharged without receiving critical treatment procedures, owing to non-payment. This exposed patients to further complications, such as septic abortion when a patient could not afford to pay for antibiotics, or a perforated uterus for a patient who was delayed before and evacuation procedure.

Regardless of ability to pay, the Kenyan constitution guarantees every Kenyan of the right to emergency care. However, patient referrals increase the out-of-pocket cost of treatment, particularly when referred from public to private facilities. For other patients, the requisite mode of payment for services posed additional challenges. Lack of payment options for patients necessitated the emergence of brokers, who would assist patients to make payments at a fee. Patients were often required to make cash payments for procedures, including large sums of money for surgical procedures. This requirement would restrict services even to patients who could make these payments using an alternative payment mode such as a credit or debit card. In large government hospitals, use of cash was restricted, as facilities favored mobile money transfers (M-Pesa), yet patients were not always able to make such transactions. Consequently, PAC patients were at times required to walk for long distances in search of mobile money services:

Maybe you have cash and not in M-Pesa [mobile money], when you are sent to buy drugs or pay for something, you are told that you should pay through M-Pesa, now with that it delays because you have to move from there and go to M-Pesa, deposit the money so that you can then come back, and pay. If it was someone in a critical condition, he/she can just die.”~ (Patient, level 6, public, high quality facility)

The requirement that patients paid prior to each procedure restricted access to timely care within facilities. Similarly, inability to pay for services led to multiple referrals, with patients typically moving from a private facility near their residence to a public facility, far from their residence. In private facilities, providers reported that patients often could not afford services at all, even after the services had been rendered. In addition to direct treatment costs, even in public facilities where all services are supposedly free, other indirect costs such as toiletries and bathing accessories such as bath basins further increased their out-of-pocket expenses.