A cross‐sectional analysis of Kenyan postabortion care services using a nationally representative sample

To assess quality of postabortion care (PAC) offered by Kenyan healthcare facilities.

service quality and sustainability from the perspective of both patients and healthcare providers. 4 The five-point PAC Consortium framework takes a public-health approach to respond to women's broad sexual and reproductive health needs. These five elements include community and service provider partnerships to prevent unwanted pregnancies and unsafe induced abortion, counseling, treatment, contraceptive and family planning (FP) services, and links to reproductive and other health services. 4 A high proportion of women who undergo unsafe induced abortion will require medical care 5 ; however, access to quality PAC remains a challenge, especially among low-income countries. 6 Furthermore, induced abortion is illegal in many low-income countries, and women often fail to seek intervention owing to a fear of legal repercussions. 7 Conversely, many patients who actively seek PAC do not receive immediate and adequate treatment at healthcare facilities, mainly due to too many patients in healthcare facilities and other access restrictions such as cost. 8 Such delays to provision of PAC can lead to severe complications such as sepsis, perforated uterus, and obstetric hemorrhage, 9 which can all cause death if advanced medical or surgical intervention is not provided in time. Together with stigma, cultural, socioeconomic, and religious factors minimize any conversations on induced abortion. 9 Although complications from unsafe induced abortion can be life-threatening, many women fail to seek PAC because they fear reprimand from medical personnel, whereas others do not seek care due to poverty and cultural practices, 9 even among countries where PAC is highly subsidized. 10 The leading known cause of unsafe induced abortion is unintended pregnancy, 11 which is associated with unmet need for FP services. Contraceptive interventions are therefore an important pathway to prevent unsafe induced abortions through reducing the risk of unwanted or mistimed pregnancies before and after induced abortion. The PAC Consortium model advocates the provision of FP counseling and services to increase contraceptive uptake and so reduce repeat unintended pregnancies. 12 In Kenya, maternal mortality remains high. 13 The majority of such deaths are associated with the type of care sought once women are faced with pregnancy complications. 14 The PAC Consortium framework provides fundamental elements of quality of care and is generally accepted as the standard approach to estimating facility-based quality of care. 15 Nonetheless, little research has been conducted to date on the quality of PAC in Kenya.
The aim of the present study was to evaluate PAC services among Kenyan healthcare facilities.

| MATERIALS AND METHODS
An analysis was conducted using cross-sectional data derived from the Incidence and Magnitude of Unsafe Abortions (IMUA) study, which was conducted among PAC-providing Kenyan healthcare facilities from March 13 to June 30, 2012. 16 Approval was obtained from the ethical review committees of the Kenya Medical Research Institute, the University of Nairobi, and Kenyatta National Hospital (all in Nairobi, Kenya). Service providers and patients signed informed consent forms before participating in the interview.
The present analysis used data from two components of the IMUA study: the retrospective health facility survey and the prospective morbidity survey. Data for the health facility survey were collected from healthcare managers and senior staff by trained field workers.
The prospective morbidity survey collected data from 2631 (83.2%) of the 3161 patients who presented for PAC services at a participating healthcare facility in a 30-day period. The remaining 530 patients did not provide consent and so were excluded, as were those who sought a pregnancy termination. Trained providers who offered PAC services at the sampled healthcare facilities collected prospective morbidity survey data, including sociodemographic characteristics, reproductive and clinical histories, diagnosis, treatment, contraception, and clinical outcomes.
The Kenya Essential Package for Health classifies all healthcare facilities into six levels of preventive and curative services, based mainly on functionality. 17 Level 1 refers to preventive approaches in the community; therefore, level 2 was considered to be the lowest level of care (e.g. dispensaries and clinics), whereas level 6 was the highest level of care (e.g. national referral hospitals). 17 The IMUA study sampled 350 (12.3%) of 2838 nationally representative level-2-6 healthcare facilities, and 326 (93.1%) participated in the study. All level-1, and some level-2, healthcare facilities were excluded from the IMUA study as they did not offer PAC services. A stratified random sample was drawn, with the geographic region and level of care used as the strata. The present analysis assessed the quality of PAC services offered at the level-2-6 Kenyan healthcare facilities by using the PAC Consortium model. 4 The indicators for each component of PAC assessment are presented in Table 1. These indicators focused on the use of appropriate technologies, provider technical performance, information and counseling, and equipment, supplies, and medications.
All data were captured on paper forms and then entered into computer systems using Census and Survey Processing System (CSPro)

| RESULTS
Characteristics of the healthcare facilities and patients included in the present analysis are outlined in Table 2. Most of the healthcare facilities sampled were level 2-3, and most were publically owned. The median age of the patients was 26 years; 1700 (64.4%) were married or living with their partner; 952 (35.5%) had received secondary education; and 908 (34.6%) were nulliparous.
Only 408 (41.8%) of all first-trimester PAC cases were treated using appropriate technology (as defined in Table 1). In all, appropriate technology for second-trimester PAC was used in 826 (82.6%) of all healthcare facilities used. Use of appropriate technology for first-trimester PAC was higher among public healthcare facilities than among private facilities (Table 3).
Approximately two staff members per healthcare facility had received in-service or preservice training in PAC (Table 3). A substantial proportion of the healthcare facilities reported that at least one staff member had received PAC training (Table 3). For both public and private healthcare facilities, there was a marked difference between the mean number of staff trained at level 2-3 versus level 4-6 ( This indicator is based on two main types of measurements: patients (as observed and recorded by the healthcare providers) or healthcare facilities (as reported by a key informant within each center). Values for the following measures are given as the percentage: • Whether patients reported that they were given pain medications for their procedure (yes or no) • Whether patients were given antibiotics for their procedure (yes or no) • Whether patients were given modern FP methods on discharge from the facility (yes or no) Sourced from the prospective morbidity survey component of the Incidence and Magnitude of Unsafe Abortions study. 16 All prospective morbidity survey data were collected by trained service providers (predominantly the main providers of PAC in the sampled healthcare facilities). c Sourced from the health facility survey component of the Incidence and Magnitude of Unsafe Abortions study. 16 All health facility survey data were collected by trained field workers using paper questionnaires, which were administered to health managers and senior hospital management staff.

| DISCUSSION
The present study found a low level of adherence to the predefined PAC service standards among all healthcare facilities sampled.
Notably, the use of appropriate technology was low, especially among private healthcare facilities. In providing the technology, a broad approach should ensure support in contraception and pregnancy testing, improvement in assessment of gestational age, supply of manual vacuum aspiration kits, and improved availability and ease of access to medical abortion. 18 The use of invasive methods such as dilation and curettage (especially for first-trimester evacuation procedures) raises concern over the adherence to guidelines, as well as the process through which service providers decide on treatment and the factors that inform the decision. The present finding differs from past studies, which showed that dilation and curettage equipment were available in healthcare facilities yet remained unused. 19 Interestingly, almost all dilation and curettage cases recorded in the IMUA study occurred at level-4-5 private facilities. 16 Use of appropriate technology for PAC in the first and second trimesters can be seen against the backdrop of high cost of some of the highly invasive and inappropriate procedures  c F value. Probability (F t >F); significance was cutoff at 0.05, meaning any F value below 0.05 was considered to represent a significant difference between the means. d Data given for healthcare facilities that always offer FP counseling. such as dilation and curettage. Some providers are thought to have shown a preference for these inappropriate methods for financial gain rather than as patient-centered treatment choices. 6 Unwanted pregnancies are a leading cause of unsafe induced abortion, the majority of which can be attributed to an unmet need for contraception. 20  women's health. Improving the quality of PAC services should focus not only on availing contraceptive services, but also on ensuring that these services have the potential to reach the intended users.
Enhanced community involvement in PAC increases awareness of the dangers of unsafe induced abortions and improves utilization. 24 Additional evidence 25 demonstrates that healthcare facilities with specialized obstetrics and gynecology services have increased odds of improved PAC. This evidence preempts the need for increased service provision and improvement in low-level healthcare facilities. Having a separate evacuation room for all PAC cases is also vital for quality improvement and enhanced patient satisfaction with services. 25 Potential limitations of the present study pertinent to interpretation of the results included the patient selection, provider recall bias, and the inability to assess all five elements of the PAC Consortium model. 4 In conclusion, the present study revealed an urgent need to improve the capacity of level-2-3 healthcare facilities to offer quality PAC services, by equipping service providers with the requisite skills to offer effective FP and counseling. Although these efforts could easily focus on public healthcare facilities (which deliver the bulk of PAC in Kenya), an inclusive approach that promotes private-sector participation will reduce the risk of complications attributable to delays in seeking care, often owing to lengthy referral times. 14

AUTHOR CONTRIBUTIONS
MMM conceptualized the present study; led the data analysis; and wrote the first draft of the manuscript. TNOA and BWM reviewed the manuscript drafts. COI conceptualized the present study; was the Principal Investigator of the Incidence and Magnitude of Unsafe Abortion study (which provided data for the present analysis); and reviewed the manuscript drafts. All authors read and approved the final manuscript before submission.

CONFLICTS OF INTEREST
The authors have no conflicts of interest.