Evaluation of emergency obstetric and neonatal care services in Kumba Health District, Southwest region, Cameroon (2011–2014): a before-after study

Worldwide, about 830 women die from complications of pregnancy and childbirth daily. Ninety-nine percent of these maternal deaths occur in low and middle-income countries and most are preventable [1]. In 2015, maternal mortality ratio was reported to be 239 per 100,000 life-births in low and middle-income countries [2], with the leading causes being direct obstetric complications notably, haemorrhage, hypertension, infections and unsafe abortion [3]. Sub-Saharan Africa alone accounts for 179,000 pregnancy and childbirth-related deaths each year [2]. In Cameroon, about 7000 women die from complications of pregnancy and childbirth yearly. The country had been ranked 15th from bottom on a global league table of maternal mortality ratios [4]. In 2000, 2005, 2010 and 2015, maternal mortality ratios in Cameroon were 750, 729, 676 and 596 respectively [5]. Thus, despite the enormous efforts made by the Cameroon Ministry of Health and its partners, the country had failed to attain its target of reducing maternal mortality ratio to 350 by the year 2015.

The main causes of maternal death (direct obstetric complications) can be prevented by a recommended package of nine evidence-based interventions (signal functions). These interventions which are collectively called emergency obstetric and newborn care (EmONC) services comprise: use of parenteral antibiotics, use of parenteral anticonvulsants, administration of parenteral uterotonics, manual removal of the placenta, removal of retained products of conception, assisted vaginal delivery, resuscitation of the newborn, emergency surgery (notably, emergency caesarean section) and blood transfusion. Comprehensive EmONC includes all signal functions while basic EmONC does not include emergency surgery and blood transfusion [6]. Basic EmONC only can help avert as much as 40% of intra-partum related newborn deaths and a considerable proportion of maternal deaths [7].

While the international scientific community acknowledges that quality EmONC services can significantly reduce maternal and neonatal deaths [8, 9], there has been much uncertainty regarding the status of these vital services in Cameroon. Maternal and neonatal mortality rates have been persistently high despite the introduction of EmONC, with national trends revealing considerable increase in mortality rates during the 2004 to 2011 period. Under these circumstances, it is imperative to assess and report the status of EmONC services in the country. This could reveal strategies of sustainably improving these essential services. This study aimed to assess coverage, functionality and quality of EmONC services in Kumba health district. This is one of the largest rural health districts in Cameroon and the largest in the Southwest region of the country. Because of the large size of the district and inadequate resources, it has been a challenge to the district health service to organize large-scale evaluations and generate comprehensive reports on the status of EMONC services. Consequently, since the introduction of EmONC services in the district, data on their availability, utilization and quality have remained obscure.

All health facilities in KDH were eligible for the study given that EmONC had been instituted in the district since 2011. However, only 12 (39%) of these had staff trained on EmONC. Among the 12 EmONC facilities, 3 (25%) were not evaluated because they were geographically inaccessible during the data collection phase: two facilities could not be accessed because of difficult road conditions during the rainy season while one facility could only be accessed using canoes which were hardly available. Among the 9 facilities that were evaluated, 58% were public facilities (Fig. 2). Among the assessed centres, 4(44%) had been performed designated signal functions, with 2 qualifying as CEmONC and 2 as BEmONC.

This study evaluated the availability, utilization and quality of EmONC in KHD, the largest health district in the Southwest region of Cameroon. The data assessed was for the period spanning from 2011 when EmONC was introduced in the district to the end of 2014, i.e., just before 2015 when the level of attainment of Millennium Development Goals had to be assessed. Although previous studies have evaluated EmONC services in sub-Saharan African settings, reports based on results from before-after studies or other quasi-experimental evaluation methods are scarce. According to this study, the number of facilities expected to provide EmONC services was slightly above the required UN minimum coverage. A considerable drop in deliveries in the included health facilities was noticed in the second segment of the period under review. Caesarean section rates remained very low. Met needs increased after EmONC were fully instituted. Direct obstetric case fatality rates remained high.

The availability of EmONC meeting the required minimum coverage in KHD is a pertinent deviation from the low availability which is characteristically reported in low-income settings especially during the period prior to the target Millennium Development Goal for maternal and newborn health [10‐13]. Notwithstanding the apparently optimal number of facilities designated to provide EmONC in KHD, there seemed to be major gaps in the geographical distribution, performance of signal functions and quality of EmONC. Thus, the considerably high neonatal mortality rate observed in EmONC facilities was possibly linked with lack of the neonatal resuscitation signal function, but it is plausible that the full institution of EmONC also improved the recording of deaths. To be more certain about these links, there is need to conduct analytic studies comparing mortality rates between facilities without and facilities with EmONC (in general) and neonatal resuscitation (more specifically), and using appropriate statistical tests to assess if there is a statistically significant difference in mortality rates between the two types of facilities. However, this was not within the scope of this study which sought to investigate changes in EmONC indicators (including mortality data) before and after full implementation of EmONC. The discordance in availability of EmONC and quality of services observed in our study was also reported by Ameh et al in a large cross-sectional study evaluating the status of EmONC in six African countries [13]. A common lack of trained and qualified staff and lack of equipment to carry out signal functions could explain the discordance noted in our study and the report by Ameh et al. Large cross-sectional studies by Hanson et al in Tanzania [14], Wilunda et al in Uganda [15] and Abegunde et al in Nigeria [16] also revealed persistently high maternal and neonatal death rates even after the institution of EmONC. Contrary to these findings, Kim et al in Afghanistan reported optimal EmONC quality with case fatality rates as low as 0.7% in their low-income study setting [17]. In line with these, there is evidence of a reduction in maternal and neonatal mortalities after ameliorations in training of personnel and performance of EmONC signal functions in a low-income context [18]. These contrary reports are possibly due to the higher level of technical and infrastructural facilities providing EmONC (Kim et al) and the scale-up of quality EmONC services (Ameh et al). In KHD, even though the lack of adequate resources may explain the increase in mortality after EmONC services were introduced, it is very likely that systems for registering and reporting mortality improved with the introduction of these vital services and this may partly explain the observed increased mortality after the introduction of EmONC. These suggest that scaling up quality EmONC in rural areas (like KHD) where most maternal and neonatal deaths occur ought to be be prioritized as a means of averting maternal and neonatal mortality and improving death reporting systems.

Some reports indicate serious short falls in EmONC indicators as the study by Admasu et al in Ethiopia [11]. These translated into negative outcomes such as low delivery rates, low caesarean section rates and high obstetric case fatality. Differences in population knowledge, attitudes and practices as well as differences in implementation gaps in different settings could explain these contrasting findings. Continuous sensitization of populations on the availability and effectiveness on EmONC and consistent monitoring of EmONC services by health authorities could alter behaviors and improve the status of EmONC services respectively in low-income settings. In our study, apart from the provision of EmONC, the increase in caesarean sections and overall deliveries as a proportion of expected deliveries could be also due to other factors such as obstetric kits whose pilot phase was launched by the ministry of health and the UNFPA by 2012 [19]. Obstetric kits lead to faster and cheaper access to obstetric care [19]. Moreover, healthcare providers inducing demand for obstetric services such as caesarean sections (as a result of reduced revenue associated with the policy of obstetric kits) is plausible especially in this low-income context.

This study is not void of limitations. There were reporting inconsistencies among the included facilities especially because of the retrospective nature of data collection. We tried to reduce the effect of this on the quality and internal validity of our results by introducing an observational grid which enabled direct observation of signal function performance although this mainly improved the quality of data for the second segment of the time series. The observational grid also helped to address this source of measurement bias by including data from the most standard sources that were available i.e., demographic health survey and district health information system. Furthermore, the records available did not allow us to consistently report other indicators of quality notably, indirect causes of maternal deaths. Nonetheless, we were able to capture data on direct obstetric case fatality rates and intrapartum and very early neonatal deaths which are very important indicators of EmONC. The findings of this study should be generalized with caution since three of the 12 health facilities that were eligible for inclusion in this study were not assessed.

As theory of change, the burden of maternal and neonatal mortality is expected to significantly reduce if EmONC is available, of good quality and effectively utilized. For the period 2011–2014, KHD fully implemented EmONC during the period 2013–2014. This study found that albeit UN minimum coverage was met following the full institution of EmONC, there were major gaps in the performance of signal functions as well as the quality and utilization of EmONC. These possibly explain the persistently high maternal and neonatal mortality rates even though it can be argued that full implementation of EmONC also led to improvements in reporting of mortality data. The gaps in EmONC identified in this study suggest the urgent need to sustainably scale up quality and accessible EmONC services in KHD. This could re-enforce the quality of staff and equipment in EmONC facilities and contribute to sustainable improvements in EmONC. Notwithstanding the findings of this study, their interpretations should be made with consideration of dynamics in Cameroon’s maternal health policies and how these may influence indicators used in grading EmONC services. Further research is warranted to explore contextual factors that could influence the performance of signal functions and ultimately the functionality, quality and coverage of EmONC services in KHD.