Background
Amid recognition that too many women die during pregnancy, childbirth and postpartum, the health system in Kenya continues to face challenges that may be contributing to lack of improvements in maternal survival. Recent estimates indicate that the maternal mortality ratio (MMR) remains unacceptably high, at 488 maternal deaths per 100,000 live births [
1]. Kenya was also among the 11 countries contributing to 65% of all maternal deaths on a global scale in 2008, and one of the 23 countries in sub-Saharan Africa making insufficient progress towards Millennium Development Goal Five [
2]. At this level of magnitude, improvements in maternal survival by 2015 present a key challenge.
There is evidence suggesting that most maternal deaths occurring in developing countries could be reduced if all women had access to interventions for treating complications that arise during pregnancy, childbirth and postpartum. This evidence reinforces the centrality of emergency obstetric care (EmOC) [
3‐
10] in reducing maternal mortality. EmOC consist of a package of life-saving interventions or signal functions that include administration of parenteral antibiotics, uterotonic drugs, parenteral anticonvulsants, manual removal of placenta, removal of retained products of conception, assisted vaginal delivery by application of vacuum or forceps, neo-natal resuscitation, blood transfusion and caesarean section [
10]. To describe the functionality and capacity of health systems in addressing life-threatening obstetric complications, a set of process indicators exist [
10,
11]. The indicators are based on the understanding that to reduce maternal deaths, obstetric services must be available and used by pregnant women. Table
1 shows the six EmOC process indicators issued in 1997 [
11] with modifications on recommended level and the two new additional ones issued in 2009 [
10].
Table 1
Emergency obstetric care indicators, questions and acceptable levels
Do the services exist and function?
|
1. Availability of EmOC: basic and comprehensive facilities | At least five EmOC facilities (including at least one comprehensive) per 500,000 population |
Are the services geographically and equitably distributed?
|
2. Geographical distribution of EmOC facilities | Equitably distributed in an area |
Are the services being used by pregnant women?
|
3. Proportion of all births in EmOC facilities | Recommended level set locally |
Are the services being used by women with complications?
|
4. Met need for EmOC services | 100% |
Do they provide critical life saving services?
|
5. Caesarean section as a proportion of all births | 5-15% |
Do they provide good quality care?
|
6. Direct obstetric case fatality rate | <1% |
7. Intrapartum and very early neonatal death rate | To be set |
8. Proportion of maternal deaths due to indirect causes | None set |
It is recommended that countries that intend to reduce maternal mortality should attempt to include EmOC process indicators into routine health management information systems to track progress in safe motherhood both nationally and at the lower levels of the health system [
10,
11]. However, the indicators are not routinely used in many countries. This includes Kenya, where only partial data from sub-national studies exist [
12,
13]. Although EmOC process indicators have been in existence since 1997, monitoring progress in maternal health goals in Kenya has relied heavily on the MMR, albeit the complexities surrounding this indicator [
14‐
16]. Lack of sufficient information on obstetric care in Kenya may suggest that in a country with a high burden of maternal mortality, policy and decision makers are often unaware of the extent of need and, therefore, where to intervene. Another challenge is that obstetric care in Kenya is presented as a “crude” coverage, not taking into account the actual care provided in the health facilities. On the Ministry of Health website, all health centres are ‘automatically’ classified as basic EmOC [
17], meaning that on paper, coverage is considered “good”. There is need for countries to classify EmOC facilities after direct inspection. This provides a distinction between how a facility is supposed to function and the reality. The distinction provides policy and decision makers with information necessary to improve coverage of services that can prevent maternal mortality and morbidity.
This study aimed to determine the actual situation in terms of existence, functionality and provision of critical life-saving services [
10]. Malindi is one of the districts with a high MMR in Kenya, estimated at 625 maternal deaths per 100,000 live births by the district’s statistics. The evidence that a majority of these deaths could be averted if women have access to EmOC calls for the need to provide information on what interventions are needed to reduce these deaths.
Data presented draws from the “Response to accountable priority setting for trust in health systems” (REACT) study, whose intervention aimed at improving equity and access to quality health care at district level in Kenya, Tanzania and Zambia [
18]. EmOC was one of the service areas selected to assess fairness and legitimacy of priority setting in health care. The larger REACT study included a baseline assessment of conditions for fairness and participation (and thus legitimacy) of priority setting and other decision making in the district health services, including EmOC. This aimed to assess whether fairness and participation in decision making could have an influence on service output and outcome. Ideally, such changes were expected to emerge after an active promotion of the fairness conditions [
18].
Findings from this study thus provide information to the district health management team on the actual availability of EmOC as opposed to theoretical coverage in the district. The findings may therefore indicate if there is need to improve priority setting processes, which influence decision making on how to achieve optimal coverage and access to life-saving obstetric services for pregnant women in the district.
Methods
This was a facility-based cross-sectional survey, conducted between October and December 2010 in Malindi District, Kenya. The district is located in the southern coastal region, covering an area of 7, 792 square kilometers. Four divisions: Malindi, Langobaya, Marafa and Magarini constitute the district. The total population in the district was 400,514 people in 2009, with urban–rural distribution of 140,739 and 259,775 persons, respectively [
19]. Malindi division has a higher population density than the other three divisions as it has favourable topographic features and economic factors affecting human settlement. Malindi town, which is located in Malindi Division, has been labeled “Little Italy”, with an estimated 3,000 Italian residents. The district has a total of 105 public and private health facilities [
17]. Of these, 42 (40%) offer delivery services. The total fertility rate in the district was 4.8 children per woman of reproductive age and crude birth rate of 38.1/1000 [
20].
All the 42 facilities (private and public) that offer delivery services in Malindi District were listed for inclusion in the study. Since it was feasible to study all the facilities listed, no sampling was done. Two facilities were, however, not reached due to bad road conditions.
Although there are a total of eight process indicators, the study focused on the first, second and fifth indicators, since the aim of the study was to describe the actual situation in terms of existence and functionality of EmOC and provision critical life saving services. The first indicator examined the availability of EmOC. This was measured by obtaining data on the number of facilities that perform the complete set of signal functions. A standard tool was used to interview the in-charge of maternity unit, whether the nine signal functions had been performed at least once during the previous three months (Yes/No) [
10]. If any of the signal functions had not been performed, reasons were recorded. A review of facility registers to ascertain that the signal functions were performed was done. In addition, observations to indicate the availability of equipment and drugs were conducted.
A strict WHO definition of a basic EmOC facility is one that has performed all the first seven signal functions in the last three months. A comprehensive EmOC facility is one that has performed caesarean section and blood transfusion in addition to basic functions in the past three months. In some instances, a signal function such as assisted delivery, is not performed in some countries as a matter of policy. According to the WHO handbook of assessing EmOC, “If a signal function is systematically absent in a region, it is possible to use the designation comprehensive “minus one” or basic “minus one” as a temporary measure while policies are reviewed and programmatic interventions planned to remedy the lack” [
10].
The second indicator examined equity in distribution of facilities. This was achieved through mapping of facilities to identify gaps in geographical distribution of services and acknowledge added barriers such as distance to facilities. Geographical coordinates of different facilities were collected using a handheld Geographical Positioning System (GPS) device (Garmin eTrex). The device automatically logged in longitude and latitude values. Facility name, administrative location and type of facility were keyed in the device. The GPS data were downloaded into a spreadsheet and mapped onto an administrative map within ArcGIS 9.3 software environment. The map contained data from the survey department, with the most up to date official administrative boundaries. Road infrastructure and key features like settlements and water bodies were overlaid with the administration boundaries data to produce base maps. The GPS data were analysed in relation to administrative locality of facilities. This facilitated identification of underserved areas and approximate distance as an independent indicator of limitation to access. The conditions of roads and various terrain barriers were not considered since the buffer tool assumes a straight line distance function that would mean in real-time land travel. The buffer proximity analysis provided the shortest distance it would take to reach the comprehensive care facility.
The fifth indicator assessed the provision of critical life saving services for pregnant women as measured by caesarean section rates in the district. To obtain this data, a form was completed for every woman who underwent a caesarean section to obtain information on the indications for the intervention, geographical origin of the women and outcome for mother and newborn. The data were collected retrospectively for the periods 1
st January 2008 to 31
st December 2009. The data, together with district population figures [
19] were used to calculate caesarean section rates by division and rural–urban residence of the women. The differences in rates between urban and rural women were compared using Pearson’s Chi-square test of association. The strength of the association was estimated using odds ratios, with corresponding 95% confidence interval.
Approval to conduct this study was obtained from the Kenya Medical Research Institute’s Ethical Review Committee (Scientific Steering Committee Number 1808). Written permission was obtained from the Medical Officer of Health in the district prior to visiting the health facilities. All data have been maintained as confidential and no individuals will be identified in dissemination of findings.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
EE was involved in the study design, data collection, analysis and initiated the manuscript. YK coordinated data collection and critically revised the manuscript. DD participated in the study conception, analysis and critically revised the manuscript. AM contributed substantially to writing and critically revising the manuscript. BEO co-coordinated the work package on EmOC and critically revised the manuscript. JB and OEO initiated and coordinated the overall “REACT” study and contributed substantially to editing of this manuscript. MM participated in data analysis and critically revised the manuscript. RM contributed to critically revising the manuscript. All authors read and approved the final manuscript.