Understanding abortion-related complications in health facilities: Results from WHO multicountry survey on abortion (MCS-A) across 11 sub-Saharan African countries

Introduction Complications due to unsafe abortions are an important cause of morbidity and mortality in many sub-Saharan African countries. We aimed to characterise abortion-related complication severity, describe their management, and to report women’s experience of abortion care in Africa. Methods A cross-sectional study was implemented in 210 health facilities across 11 sub-Saharan African countries. Data were collected on women’s characteristics, clinical information and women’s experience of abortion care (using the audio computer-assisted self-interviewing (ACASI) system). Severity of abortion complications were organised in five hierarchical mutually exclusive categories based on indicators present at assessment. Descriptive bivariate analysis was performed for women’s characteristics, management of complications and reported experiences of abortion care by severity. Generalised linear estimation models were used to assess the association between women’s characteristics and severity of complications. Results There were 13 657 women who had an abortion-related complication: 323 (2.4%) women were classified with severe maternal outcomes, 957 (7.0%) had potentially life-threatening complications, 7953 (58.2%) had moderate complications and 4424 (32.4%) women had mild complications. Women who were single, multiparous, presenting ≥13 weeks of gestational age and where expulsion of products of conception occurred prior to arrival to facility were more likely to experience severe complications. For management, the commonly used mechanical methods of uterine evacuation were manual vacuum aspiration (76.9%), followed by dilation and curettage (D&C) (20.1%). Most frequently used uterotonics were oxytocin (50â 9%) and misoprostol (22.7%). Via ACASI, 602 (19.5%) women reported having an induced abortion. Of those, misoprostol was the most commonly reported method (54.3%). Conclusion There is a critical need to increase access to and quality of evidence-based safe abortion, postabortion care and to improve understanding around women’s experiences of abortion care. The study protocol of the WHO MCS-A study has been published previously, describing the methodology of the cross-sectional study with prospective data collection across health facilities.21 This analysis focuses on the primary findings from the sub-Saharan Africa region. Briefly, after a multistage sampling, the 11 participating countries were identified (Benin, Burkina Faso, Chad, Democratic Republic of the Congo (DRC), Ghana, Kenya, Malawi, Mozambique, Niger, Nigeria, Uganda), followed by provinces and facilities in each country.22 Health facilities were only eligible if they fulfilled the following characteristics: >1000 deliveries per year, a gynaecology ward and surgical capability (defined as providing the signal functions for comprehensive emergency obstetric care, which includes removal of retained products and surgical capability23 and, if available, abortion provision and/or postabortion care. To ensure each facility could contribute sufficient data to the study during the 3-month data collection period, facilities reporting <10 postabortion care patients on average over a month in the facility assessment form were excluded. In each country, data collection took place over a 3-month period between February 2017 and April 2018. All women presenting to the participating facilities with signs and symptoms or death at discharge from abortion-related complications or early pregnancy loss (including ectopic and molar pregnancies) were included. Pregnant women with a diagnosis of threatened abortion, defined as vaginal bleeding with a closed cervix were excluded.24 The criterion of all women presenting to the facilities rather than admission to the facility was used to avoid exclusion of those women who seek care in the facilities for mild complications. Women with abortion-related complications who were admitted or had a prolonged hospital stay (>24 hours), able and were willing to consent were eligible to participate in the exit survey and convenience sampling was used to invite eligible women to participate in the exit survey based on the workload of data collectors and time of day in the study facility. A hospital administrator or a healthcare provider responsible for the gynaecology and obstetrics wards at each identified facility completed the facility assessment form that collected information on availability of services and resources. For the main survey, 1-week training sessions were conducted with research assistants at the facility and country level on the objectives of the study, data collection procedures, practice sessions with the tools as well as highlighting ethical, safety and confidentiality considerations. Techniques of information gathering on this sensitive and highly stigmatised topic using the medical records were also conducted during the training session and data collectors had access to facility coordinators and principal investigators for continuous support. Based on the eligibility criteria, research assistants at each facility reviewed and abstracted information from women’s medical records that included sociodemographic data, clinical information, obstetrics characteristics, signs and symptoms due to abortion-related complication, medical procedures, clinical outcomes and vital status at discharge to identify eligible women. Abstracted medical records data were transcribed into paper-based case report forms and entered into a web-based electronic data capture system developed by the Centro Rosario de Estudios Perinatales (Rosario, Argentina) for the study. Data entry was performed at the health facility or at a central level, dependent on logistics and available infrastructure. For the exit survey, data were collected on tablets using the audio computer-assisted self-interviewing (ACASI) system developed by Tufts University. The system allowed participants to respond to the exit survey at a private location with a focus on maintaining participant confidentiality. Data collected in the exit survey consisted of abortion safety characteristics (method used, provider, setting) prior to coming to the facility, and women’s experience of abortion care related to effective communication, respect and dignity and emotional support during their time in the facility. The women who participated in the exit survey were compensated for their time by approximately US$2 worth of mobile phone airtime. Data managers in Argentina continuously monitored the study data flow and data quality by use of validation procedures and progress reports for all countries. Data inconsistencies were identified and corrected by contacting the study principal investigators as they emerged. These procedures have been used in previous multicentre studies.25 Based on indicators present at time of hospital admission including clinical, laboratory and management-based markers, abortion-related complications were classified into five hierarchical and mutually exclusive categories based on severity: (1) deaths, (2) near miss, (3) potentially life-threatening complications, (4) moderate complications and (5) mild complications (figure 1). Study flow diagram for severity of abortion-related complications. Severe maternal outcomes (SMOs) (n=323, 2.3%). *Status at discharge. †WHO maternal near-miss criteria (organ dysfunction of either one or more of the following: cardiovascular, respiratory, renal, coagulation, hepatic, neurological or uterine dysfunction). ‡WHO potentially life-threatening complications (severe haemorrhage, severe systemic infection or suspected uterine perforation). §Moderate complications (heavy bleeding, suspected intra-abdominal injury or infection). ¶Mild complications based on abnormal physical examination findings on initial assessment (vital signs, appearance, mental status, abdominal examination, gynaecological examination). Based on WHO criteria for near miss, women who died or identified as a near-miss case were classified as a severe maternal outcome.26 Women presenting with severe haemorrhage, severe systemic infection or suspected uterine perforation were classified based on WHO’s criteria for potentially life-threatening conditions.26 Moderate complications included bleeding, suspected intra-abdominal injury and infection. Mild complications included any abnormal signs from initial physical examination (vital signs, appearance, mental status, abdominal examination, gynaecological examination). Death was based on woman’s status at discharge. Online supplemental file 4 includes the identification criteria used for each severity category in detail and, online supplemental file 1 includes information on timing of abortion-related complications based on facility admission. bmjgh-2020-003702supp004.pdf bmjgh-2020-003702supp001.pdf Gestational age at presentation was grouped as <13 weeks, ≥13 weeks or undetermined weeks (online supplemental file 1). Clinical management of abortion-related complications was categorised as medically managed by uterotonics only, by uterine evacuation only or both methods. Uterotonics use was further divided into: misoprostol alone, oxytocin alone, ergometrine only and their combinations. Uterine evacuation was further examined by type of procedure: manual vacuum aspiration (MVA), dilation and curettage (D&C) and both. Development of the protocol used evidence from qualitative research exploring women’s experiences with abortion care. The been published, the results will be disseminated for professional and non-professional audiences in participating countries. Descriptive bivariate analysis was performed for national level and facility level characteristics (online supplemental file 1), as well as sociodemographic, obstetrics and clinical management characteristics by severity of abortion-related complications. The χ2 test was used to compare proportions of descriptive characteristics across severity categories. Severity of abortion-related complications is presented across countries as ratios calculated based on the prevalence of each category of complication per 1000 women with complications. Descriptive analysis was also performed to evaluate the methods used, information received and help sought to end pregnancy for self-reported data collected in the exit interview via the ACASI platform. Experience of abortion care during facility stay was assessed by comparing responses across severity of abortion-related complications using χ2 test. Regression methods were used to evaluate women’s characteristics potentially associated with the outcome of abortion-related complication severity. Generalised linear models, adjusting for facility clustering effect and differences across countries, were fitted to estimate the odds of severe maternal outcomes, potentially life-threatening complications and, moderate complications for women’s characteristics. The independent variables were categorical variables including sociodemographic characteristics (age, marital status, education, gainful occupation) and, obstetric characteristics (prior pregnancies, gestational age and expulsion of products of conception before arrival to the health facility). Data analysis was conducted using SAS (V.9.4, Cary, North Carolina, USA).