Health systems factors influencing the delivery of emergency obstetric care in a South African district

Abstract

Background Ekurhuleni District has an unacceptably high maternal mortality ratio (MMR) that is persistently greater than the provincial average. The Gauteng Province identified Ekurhuleni as a high priority district, requiring heightened focus to reduce the MMR and end preventable maternal deaths. Most of these deaths are due to direct pregnancy related causes that can be effectively reduced by the emergency obstetric care (EmOC) intervention. The national committee for confidential enquiry into maternal death (NCCEMD) identified health system strengthening (HSS) as one of South Africa’s key focus areas to decrease maternal mortality. Strategies for health system strengthening are not clear however. This PhD work sought to investigate health systems weaknesses, which if addressed could strengthen the health system of Ekurhuleni district, in order to reduce the MMR. Purpose To identify health system factors that influence the delivery of EmOC, and understand how these factors translate to severe maternal outcomes (SMOs) in the delivery of EmOC. Methods A mixed-methods study design was used. EmOC was used to trace the health system’s support of maternal health using the WHO health systems building blocks. The extent to which EmOC services were available in the district was assessed. A rapid health system audit was performed in 15 public health facilities that offered EmOC services to the public in the district. To measure the impact of health system factors to maternal outcomes, a comparative near-miss audit was conducted over 12 months in two regional hospitals with contrasting performances in maternal outcomes. Lastly, a qualitative study was performed to explore in-depth, heath system governance issues that impacted on EmOC delivery and subsequent outcomes of the district. Analysis was methods specific. Descriptive statistics reported in proportions and frequencies of key indicatorsin the assessment of the availability of EmOC and the rapid health systems audit. Firth adjusted penalised conditional logistic regressions were used to establish associations of socio-demographic, obstetric and health system factors with severe maternal outcomes (SMOs). Relationships were confirmed by multiple regression analyses. Thematic analysis was performed for qualitative data to determine governance and leadership issues explaining the reasons behind the health system performance measured by maternal outcomes. Results All 7 hospitals were classified EmOC facilities and none of the 8 community health centres (CHCs) qualified to be classified. There was a shortfall of at least 18 EmOC facilities to meet the current population need in the district. Midwives in hospitals had a significantly higher workload than in CHCs (p<0∙0001). Skilled health providers were not always available for women requiring EmOC in CHCs. None of the health facilities (0.0%, N=15) scored 100% on any checklists measuring drugs, equipment, supplies, or EmOC guidelines. Health system effect on quality of EmOC and corresponding outcomes The overall hospital MMR was high at 139.8 per 100 000 live births. Haemorrhage (35.4%) and hypertension (32.5%) were the most common underlying causes of SMOs. Anaemia was an important contributing factor as it was present in 21% of women. Antenatal care (ANC) was protective against SMOs (p<0∙001). Poor overall quality of EmOC, administrative problems, and health provider related problems had the highest odds for severe maternal outcomes (SMOs). EmOC facilities were often inadequate to meet need and could not therefore deliver care adequately. Impact of health system factors on SMOs The likelihood for SMOs from lack of access into the operating theatre (OT) was 17 fold (OR=17.4; 95%CI=4∙28-70.56). Women were nearly 71 times more likely to have SMOs from lack of access into the intensive care unit (ICU) (OR=70.6; 95%CI=4.39-1135.35) when required. Ambulances (OR=4.0; 95%CI=0.73 - 22.11) were often unreliable and this predicted SMOs. Health providers sometimes gave substandard care (OR=10.0; 95%CI=1.24-80.48) or delayed to give care (OR=15.2; 95%CI=4∙87-47∙72) and this significantly raised the risk for SMOs. EmOC health providers were sometimes fewer than the need for EmOC by women, raising the risk for SMOs significantly (OR=3.1; 95%CI=0.51-6.33). Important patient factors were women delaying to present for care in hospitals (p<0.01) and lack of xi information on pregnancy danger signs (p<0.05).Patient factors also had a high odds for SMOs in bivariate analysis, but were less significant (p=0.12) in the multivariate analysis where health provider (p<0.001) and administrative factors (p<0.001) were included. Management experiences of the health system In the interviews, district managers identified important health system weakness in relation to all six building blocks. Senior management posts were often vacant or manned by acting managers with a high turnover. Other important health systems issues included segmented EmOC services with suboptimal coordination between them; inadequate health providers to administer EmOC; insufficient numbers of ambulances available for EmOC; overwhelmed health facility infrastructure; inadequate routine health system indicators to rapidly reflect performance on EmOC; and insufficient overall funding for EmOC delivery. Conclusions The overall state of the health system is indeed weak in Ekurhuleni district, and results in poor quality of EmOC given, which corresponds to the occurrence of SMOs. EmOC medicines and supplies were not always available in health facilities and this limited the availability of EmOC in the district. Health system factors have a significant impact on SMOs, and therefore need to be addressed in order to prevent avoidable SMOs. To strengthen the health system, more EmOC facilities are needed to have EmOC readily available for women and decentralise EmOC. The organisation of EmOC services also needs revisiting to improve coordination by managers and subsequent service delivery. Furthermore, ANC is effective in protecting women from SMOs, and is a strength of the health system. More needs to be done to improve ANC attendance to 100% and to assure more than 4 visits by all women in the district. This PhD also contributes methodological innovations in measuring the size of the impact of health system factors on SMOs