Using a linked population and facility based survey and appropriate modeling framework, this study assessed the effect of individual, household, communal and health facility characteristics in utilization of antenatal, delivery, and postnatal care by a skilled provider. The results suggest that utilization of skilled maternal care by individual women is very low in the study area. The finding is consistent with the recent EDHS report and previous studies conducted in the surrounding area [5, 19, 20].
Our analysis indicates that utilization of skilled maternal care by individual women depends on the joint effect of individual, household, communal and facility characteristics. According to the intra-class correlation results, the contribution of unobserved communal level characteristics, were 31% for antenatal care, 12% for delivery care and 25% for postnatal care. In all the three intercept-only models, the contributions were significant and indicated that determining association without the control of variables at different levels would give a misleading result. This was also observed during analysis where many of the significant associations disappeared when the effect of clustering by kebele was controlled. Previous studies based on a similar analysis showed consistent findings [12, 13].
At the individual level, variables related to awareness and perceptions were found to be much more relevant for skilled maternal service utilization. Women may get knowledge on the importance of skilled maternal care in different ways such as previous exposure to skilled maternal services, community based health educations, through community media or due to their better educational status. In this study, women with at least one antenatal care in previous pregnancies were significantly more likely, to use skilled maternal care for all the three indicators of their recent pregnancy. The positive effect of maternal service in the previous birth on current maternal care was observed in different studies done elsewhere [12, 21, 22].
Having awareness about health facilities to get skilled professionals was a significant predictor of both skilled antenatal and delivery care. The effect of awareness was also observed in community-based interventions focusing on awareness creation. For example, in Ethiopia, higher rate of household visits and awareness by health extension and community health workers were associated with improved antenatal care use and postnatal care visits . Better educational status is believed to be an important factor for better awareness and positive attitude related to maternal health service utilization. As expected, skilled maternal care during pregnancy, delivery, and postnatal period increased steadily with education. Women with secondary and above education were also significantly associated with antenatal and delivery care by a skilled provider. Previous studies support our findings [13, 19, 24, 25].
One of the expected effects of knowledge about an issue is change on individual attitudes. Women who have confidence on skilled providers and their care tend to use maternal health services by a skilled provider. In our analysis, significantly higher utilization rate of all skilled maternal health service indicators was observed among women who perceive that health professionals are better and safer (prefer skilled providers) for their maternity care. The implications of this finding is that community based health education about the benefit of skilled maternal care by targeting women who prefer non-skilled providers as well as improving the quality of care by providers will bring a positive contribution for utilizing skilled maternal care.
The predictor birth order indicates that women tend to use skilled maternal care if their birth is the first, and its significant effect is observed in antenatal and delivery care. The variations observed in the odds ratios can be related with the risk perception of women that varies overtime. Many women (48.3%) believe the first pregnancy is risky compared with the next consecutive pregnancies. However, as the number of pregnancies exceeds a certain limit, they start to think about another risk. Because of this, many women (41.3%) believe that having too many births is risky for a woman.
Women’s health seeking behavior is also influenced by the cost of maternity services and their capacity to cover the expected expenses. For instance, a substantial proportion of antenatal care users did not deliver or use postnatal care by a skilled provider. On the one hand, maternity services, especially delivery care, are expensive. Studies indicate that delivery care use among antenatal care users is highly correlated with wealth . On the other hand, delivery occurs suddenly. As a result, women have home birth by non-skilled providers due to transportation problems and lack of preparation. The positive contribution of better wealth status for all maternity service indicators and its significant contribution to postnatal care are also observed in our analysis. The study further evaluated the contribution of wealth indicators at kebele level. Women living in kebeles with mixed (farming and trading) source of income were utilizing significantly more skilled delivery and postnatal care than those living on farming only. The finding indicates that the presence of different sources of income for covering payments of transportation and other services contributed to the existence of higher rate of expensive maternal services in the communities. In this study, payment requirement at the time of delivery was an important barrier to delivery service by a skilled provider. The negative effect of the cost of maternity services and the need for reducing it was also observed in a study conducted in Tanzania. In the Tanzania study, standard maternity services were costly. For most households, maternal health care could take more than half of their annual consumption. The study further indicated that by subsidizing maternity care, women, especially those in the lowest socio-economic category would experience the greatest increase in service utilization . The findings of our study and previous evidences imply that for better achievements in the utilization of skilled maternal health services, there should be mechanisms to reduce or avoid the costs of maternal services.
Another strong facility level predictor for skilled maternal care utilization was the performance of health facilities. The presence of all the six signal functions in the nearby basic essential obstetric care facility (health center) positively contributed to the utilization of all indicators of skilled maternal services, and its effect was significant on skilled attendance rate. Functioning obstetric facility means performing the essential services for normal situations and complications and these services should be available 24 hours a day and 7 days a week. The presence of all signal functions reflects better performance (quality) of a health facility. Furthermore, the findings of this study imply that improving the quality of care and monitoring facilities to perform the expected function at their level will be one of the strategies to increase the utilization rates of maternity care by a skilled provider.
One of the limitations of this study was that women faced difficulties to differentiate the type of skilled providers during interview. To reduce the problem, data collectors gave further clarifications by collecting information on types of providers (doctor, health officer, nurse, midwife, health extension worker or others) from health institutions that gave services. Errors may occur in knowing the categories of health professionals. However, there were no serious difficulties in categorizing service providers as skilled or non-skilled, so it did not affect the interpretation of results. The potential limitations expected during analysis of determinant factors were minimized.