Barriers to providing quality emergency obstetric care in Addis Ababa, Ethiopia: Healthcare providers’ perspectives on training, referrals and supervision, a mixed methods study

Background: Increasing women’s access to and use of facilities for childbirth is a critical national strategy to improve maternal health outcomes in Ethiopia; however coverage alone is not enough as the quality of emergency obstetric services affects maternal mortality and morbidity. Addis Ababa has a much higher proportion of facility-based births (82%) than the national average (11%), but timely provision of quality emergency obstetric care remains a significant challenge for reducing maternal mortality and improving maternal health. The purpose of this study was to assess barriers to the provision of emergency obstetric care in Addis Ababa from the perspective of healthcare providers by analyzing three factors: implementation of national referral guidelines, staff training, and staff supervision. Methods: A mixed methods approach was used to assess barriers to quality emergency obstetric care. Qualitative analyses included twenty-nine, semi-structured, key informant interviews with providers from an urban referral network consisting of a hospital and seven health centers. Quantitative survey data were collected from 111 providers, 80% (111/138) of those providing maternal health services in the same referral network. Results: Respondents identified a lack of transportation and communication infrastructure, overcrowding at the referral hospital, insufficient pre-service and in-service training, and absence of supportive supervision as key barriers to provision of quality emergency obstetric care. Conclusions: Dedicated transportation and communication infrastructure, improvements in pre-service and in-service training, and supportive supervision are needed to maximize the effective use of existing human resources and infrastructure, thus increasing access to and the provision of timely, high quality emergency obstetric care in Addis Ababa, Ethiopia. A census, in March 2013, of the maternal health providers in the St. Paul’s Hospital Millennium Medical College (SPMMC) hospital- health center network, identified 138 maternal health providers. The sample for the quantitative survey consisted of 111 of 138 (80%) of the healthcare workers providing maternal health services (antenatal (ANC), post natal, and delivery care) in the referral network. Staff providing these three services often rotate between the ANC, post natal and delivery wards, so providers from all three levels of care were included in the survey. Three providers declined to complete the questionnaire, a refusal rate of 2% (3/111). Twenty-four providers (17%, 24/111) were not available to complete the survey because of annual leave, off site training or being off duty. The data collection tools were piloted and pretested in a comparable hospital-health center network in Addis Ababa: the Ghandi Memorial Hospital and the Kirkos health center. The quantitative data collection instrument can be found in Additional file 1. Data were collected by six trained data collectors at the health facilities. The key informant interview participants were purposively selected from the population of maternal healthcare providers in the referral network based on their professional responsibilities. The participants for the in-depth interviews included the medical directors from each of the networked health centers, the liaison officers who are responsible for maternal referrals, maternal and child health coordinators, and senior nurses and midwives from the labor rooms. The structured interview guide that was used can be found in Additional file 2. Of the 31 respondents identified, two were unavailable to participate, and none refused to participate. This resulted in a sample size of 29 with three participants from the hospital and 26 from health centers. Each respondent was interviewed individually at his/her place of work by a team of trained data collectors from the Addis Continental Institute of Public Health (ACIPH): one interviewer and a note taker. Both data collectors were trained in the use of the interview guide and qualitative methodologies before data collection began. Interviews were conducted in private offices, and ranged in duration from 22to 74 minutes. Interviews were audio recorded, transcribed in Amharic, and then translated to English [11]. During the coding process, translations from Amharic were back-translated and validated within the context of the interview by a native Amharic speaker (HG); no discrepancies from the original translation to English were identified. Qualitative and quantitative data was collected in March of 2013. Data collectors and study participants were aware that the goal of the study was to assess factors (including provider knowledge and confidence in their skills and facility resources) that influence the provision of emergency obstetric services in order to develop interventions to improve the quality of care. In-country ethical clearance was obtained from the ACIPH Institutional Ethical Review Board. Informed consent was obtained from all study participants. The study is registered with the National Institutes of Health in the United States of America: {“type”:”clinical-trial”,”attrs”:{“text”:”NCT01802957″,”term_id”:”NCT01802957″}}NCT01802957. A priori themes were coded based on the study objectives and emergent themes were identified based on the narratives of research participants. Interviews were analyzed and coded by one researcher (AA) and the coding was verified by a second researcher (MB). This analysis draws on participating providers’ responses to queries about three main issues: the functioning of the newly developed referral system; current staff capacity and the need for technical training to appropriately handle obstetric emergencies; and status of and satisfaction with the supervisory visits. The quantitative survey provides complementary information about 1) healthcare provider characteristics 2) receipt of and satisfaction with in-service training 3) perception of supervision and job satisfaction. Data were double entered using Epi Info software version 3.5.1 and descriptive statistics were generated using SAS V9.2. When comparing and contrasting the qualitative and quantitative findings during the interpretation, we adapted the classic “three delays” model that has been developed by Thaddeus and Maine and fruitfully applied by other researchers, to conceptualize the barriers to quality obstetric care identified by providers in Addis Ababa, Ethiopia [8,12-16].