Introduction: The ministry of health (MOH) of Ethiopia recommends 4 or more focused antenatal care (ANC) visits at health centre (HC) or at a higher level of health facility (HF). In Ethiopia, few studies investigated time dimension of maternal health continuum of care but lack data regarding place dimension and its effect on continuum of care. The aim of this study is to estimate effect of place of ANC-1 visit and adherence to MOH’s recommendations of MOH for ANC visits on continuum of care rural in Ethiopia. Methods: We used data collected from 1431 eligible women included in the National Health Extension Program (HEP) assessment survey that covered 6324 households from 62 woredas in nine regions. The main outcome variable is continuum of care (CoC), which is the uptake of all recommended ANC visits, institutional delivery and postnatal care services. Following descriptive analysis, Propensity Score Matching was used to estimate the effect of place of ANC-1 visit on completion of CoC. Zero inflated Poisson regression was used to model the effect of adherence to MOH recommendation of ANC visits on intensity of maternal health continuum of care. Result: Only 13.9% of eligible women completed the continuum of care, and place of first antenatal care (ANC) visit was not significantly associated with the completion of continuum of care (β = 0.04, 95% CI = -0.02, 0.09). Adherence of ANC visit to the MOH recommendation (at least 4 ANC visits at higher HFs than health posts (HPs)) increased the likelihood of higher intensity of continuum of care (aIRR = 1.29, 95% CI: 1.26, 1.33). Moreover, the intensity of continuum of care was positively associated with being in agrarian areas (aIRR = 1.17, 95% CI: 1.06, 1.29), exposed to HEP (IRR = 1.22, 95% CI: 1.16, 1.28), being informed about danger signs (aIRR = 1.14, 95% CI: 1.11, 1.18) and delivery of second youngest child at HF (IRR = 1.16, 95% CI: 1.13, 1.20). Increasing age of women was negatively associated with use of services (IRR = 0.90, 95% CI: 0.87, 0.94). Conclusion: Completion of maternal health continuum of care is very low in Ethiopia, however most of the women use at least one of the services. Completion of continuum of care was not affected by place of first ANC visit. Adherence to MOH recommendation of ANC visit increased the intensity of continuum of care. Intensity of continuum of care was positively associated with residing in agrarian areas, HEP exposure, danger sign told, delivery of second youngest child at health facility. To boost the uptake of all maternal health services, it is crucial to work on quality of health facilities, upgrading the infrastructures of HPs and promoting adherence to MOH recommendations of ANC visit.
Ethiopia is located in the Eastern part of Africa, which is administratively divided into five agrarian (agriculture as the main way of living) regions, two pastoralist (livestock raising as the main way of living), two regions with both agrarian and pastoralist areas, and two City administration at time of data collection. Each region was further administratively divided into zones, then into woredas, and finally into kebeles. Kebeles are the lowest government administrative unit and it has an average household size of 500–1000 and a population of 2500–5000. The country has an estimated 100.8 million population of which four-fifth of the population resides in rural settings, with a 4.7 average family size, and a 2.6% average annual population growth rate. Females constitute around 49.8% of the national population and half of these females are within the reproductive age [30]. Maternal health services are delivered in a three-tier health system that includes primary, secondary, and tertiary levels. At the primary level of health service delivery there are health posts staffed with Health Extension Workers (HEWs), health centers staffed with nurses and health officers, and primary hospitals. The health extension program encompasses health posts (HPs) and HEWs, and it is the main service delivery modality at the primary level for the community [31]. We used data from the National HEP assessment survey which field data collection was conducted from March to May 2019. The rural component of the survey covered 62 woreda distributed across all the 9 agrarian and pastoralist regions using multistage sampling design. Three kebeles per woreda, and 34 households (HHs) per kebele were randomly selected. A total of 6324 HHs were recruited from 185 kebeles for the survey. Thus the data is collected from health posts and HHs from the selected kebeles. The respondents for the HH survey were women and their husbands, and HEWs for health post survey [29]. The study population includes women of reproductive age (15–49 years) who delivered a child in the last two years. The study included 1431 women from 6324 HHs who were part of a fixed cohort from antenatal to postnatal care and the catchment health posts in the selected kebeles. The selected women were asked about their use of maternal health services for their last delivery in the last two years. The health post assessment includes service availability, equipment and human resource characteristics. Two outcome of interest were analysed in this paper: There are two main exposure variables. These variables include individual, household, and kebele level factors. Individual level factors include: age, marital status, whether she is a household head or not, family size, maternal and paternal education (whether the mother or husband attended grade one or more education coded as formal education), if she has exposure to HEP (Yes response if a woman took any services at a HP or if she is visited by a HEWs at her home, and No response if she is not visited anywhere), having access to media, awareness of husband and wife about MHS (ANC, ID and PNC) availability, if she was told about danger sign, and place of delivery for the previous child. Household level factors included: wealth index which is categorized into three groups (high, medium and low). The kebele level variables are the number of medical equipment at HP (a continues variable ranging from 0 to 11 which measures the availability of 11 essential medical equipment at the HP like different guidelines, statoscope, BP apparatus …), and human resource at HP which is categorized as “Presence of at most level-3 HEWs” if HPs have level 3 or below HEWs and “Presence of at least level-4 HEWs” if the HPs have at least level 4 HEWs. Access to health facility it is categorized as “Accessed HP/HEW” if the women have a nearby HP or HEW for use and “Accessed other HF” if the women have a nearby health centre or other higher level health facility. We have adapted the social ecological model since the model considers the complex interplay of multiple levels factors and interactions between individuals, household and kebele or health post level factors [32], which will affect the utilization of maternal health service in the continuum (Fig. (Fig.11). The conceptual framework showing the multilevel factors affecting maternal health continuum of care We cleaned and analysed the data using Stata version 16.1. We did weighed analysis to account for disproportionate stratification of number of different regions, use of multistage sampling to recruit study participants and to be able to generalize the finding to the national reference population. Wealth quintiles were used as a proxy measure of socio-economic status. Descriptive analysis including frequencies, crosstabulations, and graphical presentations were used to summarize characteristics of study participants across different characteristics. We used Propensity Score Matching (PSM) analysis, which is one of the treatment effect model in stata, to estimate the effect of place of ANC-1 visit on the completion of CoC which is the binary response. We used Zero inflated Poisson (ZIP) regression to model the effect of adherence to MOH recommendation of ANC visits on intensity of maternal health continuum of care that has values ranging from 0 to 6. The model is selected after checking the validity of required assumptions [33]. The exposure variable (adherence to MOH recommendation) has some degree of overlap with the outcome variable, intensity of CoC. We have conducted re-analysis of the model after removal of the overlap. We reported the findings as statistical significant whenever p-value was less than 5%.
N/A