Background: South Africas maternal mortality rate (625 deaths/100,000 live births) is high for a middle-income country, although over 90% of pregnant women utilize maternal health services. Alongside HIV/AIDS, barriers to Comprehensive Emergency Obstetric Care currently impede the countrys Millenium Development Goals (MDGs) of reducing child mortality and improving maternal health. While health system barriers to obstetric care have been well documented, patient-oriented barriers have been neglected. This article explores affordability, availability and acceptability barriers to obstetric care in South Africa from the perspectives of women who had recently used, or attempted to use, these services. Methods. A mixed-method study design combined 1,231 quantitative exit interviews with sixteen qualitative in-depth interviews with women (over 18) in two urban and two rural health sub-districts in South Africa. Between June 2008 and September 2009, information was collected on use of, and access to, obstetric services, and socioeconomic and demographic details. Regression analysis was used to test associations between descriptors of the affordability, availability and acceptability of services, and demographic and socioeconomic predictor variables. Qualitative interviews were coded deductively and inductively using ATLAS ti.6. Quantitative and qualitative data were integrated into an analysis of access to obstetric services and related barriers. Results: Access to obstetric services was impeded by affordability, availability and acceptability barriers. These were unequally distributed, with differences between socioeconomic groups and geographic areas being most important. Rural women faced the greatest barriers, including longest travel times, highest costs associated with delivery, and lowest levels of service acceptability, relative to urban residents. Negative provider-patient interactions, including staff inattentiveness, turning away women in early-labour, shouting at patients, and insensitivity towards those who had experienced stillbirths, also inhibited access and compromised quality of care. Conclusions: To move towards achieving its MDGs, South Africa cannot just focus on increasing levels of obstetric coverage, but must systematically address the access constraints facing women during pregnancy and delivery. More needs to be done to respond to these patient-oriented barriers by improving how and where services are provided, particularly in rural areas and for poor women, as well as altering the attitudes and actions of health care providers. © 2012 Silal et al.; licensee BioMed Central Ltd.
To explore access to obstetric services from the perspectives of women needing such care, this study drew on mixed methods, sequentially combining quantitative exit interviews (QUAN) with fewer, but detailed, qualitative in-depth interviews (qual) with women who had recently used obstetric services, resulting in a QUAN-qual study design [15].The qualitative phase (Phase 2), while conceptualised at the same time as the quantitative phase (Phase 1), and therefore complementary from the outset, also had the advantage of being conducted after Phase 1 and could be used to triangulate and explore some of the themes emerging from the quantitative data alongside building the larger ‘access’ picture from patient and provider perspectives. Data from both sets of interviews were integrated in the analysis, which sought “elaboration, enhancement, illustration and clarification of the results from one method with results from the other method”; a complementarity approach ( [16] quoted in [17], p.22). Two urban and two rural health sub-districts in South Africa were purposively selected as study sites. The rural sites were chosen because both are Demographic Surveillance System sites and we could draw on existing secondary datasets and contribute to building knowledge in these areas; and the urban sites were selected in large metropolitan areas in consultation with user partners – district managers and local authorities – familiar with sub-districts facing ‘typical’ access challenges. The four sub-districts were extremely geographically dispersed making contamination of information unlikely. An exit interview questionnaire was developed to collect information on the use of, and access to, maternal health services as well as socioeconomic and demographic information. Utilisation details included type of delivery and number of nights spent at the facility, while access information related to the availability, affordability and acceptability of services e.g. travel time to facility, costs incurred and health workers’ attitudes to patients. A sample size of 300 women per site was calculated based on an anticipated analysis of socio-economic inequalities in use of obstetric health services (χ2Goodness of Fit test,80 % power, medium effect size). All obstetric health facilities in the rural sub-districts were included, and the number of interviews per facility was proportional to the number of deliveries that took place in those facilities. In the urban sub-districts, obstetric health facilities were selected using the probability proportional to size methodology [18]. The interview included questions on the experience of delivery and therefore was conducted as patients left the facility. All women above the age of 18 who had been discharged from the post-natal ward were eligible for selection. Patients were selected systematically for interview until target sample sizes were reached for all facilities. Completed questionnaires and records were entered into EpiData v1.3 and analysed in Stata ®10. To develop an indicator of socioeconomic status (SES) multiple correspondence analysis was conducted on several household level variables including type of house, material of walls, type of toilet, primary source of energy for cooking and ownership of assets such as a vehicle, fridge and livestock etc. The first dimension captured 65% of total inertia and was adopted as the index for socioeconomic status. Descriptive statistics were calculated. Differences between categorical variables were tested using the Chi-square test of association and differences between ordinal variables were tested using a Chi-square test for Trend. The non-parametric Kruskal Wallis test was used to compare continuous distributions between two groups. Multiple Linear Regression analysis was also used to test associations between descriptors of the affordability, availability and acceptability of services and demographic and socioeconomic predictor variables. Where the dependent variable was categorical, logistic regression was used. Proxies were used to estimate the availability, affordability and acceptability dimensions of access. The amount of time taken in minutes for the patient to travel to the facility was used as the proxy for availability of the service. If patients travelled to a primary health care facility and needed to be moved to a district hospital by ambulance, then only the travel time from the primary health care facility to the district hospital was recorded. The total amount of money spent on the day of delivery measured as a percentage of annual household expenditure, was the estimator of affordability. This amount comprised money spent on transport, supplies such as sanitary towels and nappies, food, phoning and money used to pay someone for taking over tasks that the patient would be completing such as childcare. South Africa’s legislation prohibit user fees being charged for maternal services at public facilities and none of the sample subjects reported expenditures on medicines. In the logistic regression, the proxy for acceptability was the response to the statement “The health worker is too busy to listen to my problems”. Additionally, other acceptability variables, including levels of respect and whether health workers understood the difficulty of labour were also considered in a separate analysis. A delayed start to the quantitative phase of work in one rural sub-district meant that the qualitative phase could only be completed in three of the four sub-districts. In this phase, in-depth interviews were carried out with women chosen through a purposeful selection methodology to reflect a range of different delivery experiences, and with a particular focus on women who had obviously faced problems accessing services. As this was a facility-located, rather than community-based, sample of women, those who had given birth before they got to the facility, known as Born Before Arrivals (BBAs), were used as a proxy for women who were not able to access services. A total of 16 women who had recently delivered, in most cases at the facility were interviewed shortly after the birth. Follow-up interviews were conducted with six of these women a few weeks later in their own homes. Eight women were BBAs, five had had successful deliveries (three normal vaginal deliveries and two caesarean sections), and three had experienced stillbirths. Interview guides were loosely structured to explore patients’ life narratives, starting with their latest pregnancy and birth experience, and extending to previous pregnancies and engagements with the health system, as well as their life circumstances and backgrounds more generally. The women were interviewed by trained fieldworkers with previous qualitative research experience and interviews were conducted in the participants’ own language. These were audio-taped, transcribed and translated into English, and pseudonyms were assigned to protect confidentiality. For those women who were interviewed twice, the research team reviewed the initial interview and identified questions (for clarification and/or further exploration) for the follow-up interview. Interviews were independently coded using ATLAS ti.6 by at least two members of the research team. Coding was both deductive, using a codebook constructed around key access issues, and inductive, allowing for themes to emerge from the data. The data and preliminary analyses from both the quantitative and qualitative interviews were then integrated into a broader analysis of access to obstetric services from perspectives of women across and within the study sites. Thematic areas explored in this integrated approach included the availability, affordability and acceptability of obstetric services and barriers to access along each of these dimensions. The data were collected over a period of 15 months between June 2008 and September 2009. The Universities of Cape Town, Witwatersrand and Kwa-Zulu Natal and the South African Provincial Health Research Committees granted ethical clearance and informed, written consent was obtained from the women for the quantitative exit and in-depth interviews.
N/A