The effect of distance to health facility on neonatal mortality in Ethiopia

The distance to the nearest health facility influences the uptake of maternal and neonatal health services [1]. Lack of access to healthcare facilities may cause delays in timely care and influence maternal service utilisation due to associated travel costs and suffering [2], thereby raising the risk of newborn mortality [3]. Only a few studies have investigated the effects of distance to health facilities on neonatal mortality in low and middle-income countries and revealed the influence of distance on neonatal survival [4, 5]. A study from Zambia and Bangladesh showed that a lack of geographic access to emergency obstetric care services is a key determinant factor for the use of facility-based delivery [5, 6].

The Zambian study indicated that every doubling of distance to the nearest facility resulted in a 29% decrease in facility delivery [5]; in Bangladesh, every five-minute increase in travel time to the closest emergency obstetric care facility decrease the likelihood of facility delivery by 30% [6].

Studies have also shown that increased distance between home and health services contributed to neonatal and childhood mortality [7‐10]. A systematic review in low and middle-income countries found that distance to health facilities is a significant predictor of under-5 mortality, particularly during the perinatal and neonatal periods [11]. However, other literature reported no association between distance to healthcare facilities and neonatal mortality. There has also been reported elsewhere that neonatal mortality decreased as the distance from home to health facilities increased [4]. However, the researchers cautioned that the result might be affected by the quality of the data. These inconsistent findings might also be attributed to inappropriate modelling of the relationship between distance, maternal service use and neonatal mortality. Some studies considered antenatal care attendance, place of delivery, and postnatal care at the same level with distance [4]. Instead, the relationship between distance to health services and neonatal mortality can be better investigated by taking health service use factors into account as an intermediary between distance and neonatal mortality (see Fig. 1).

The conceptual framework presented in Fig. 1 is based on the components of the continuum of care for maternal and neonatal care integrating health service use during pregnancy, birth and immediately after birth, with the success of each service building upon the previous stage's success. Antenatal care attendance to a skilled health care provider can avert problems during pregnancy and increase the possibility of attending a facility delivery [12]. Giving birth at a health facility again promotes postnatal health service visiting and postnatal counselling about neonatal danger signs [13]. Inadequate service use in the continuum breaks the critical link in the continuity of care and can affect women's and their children's health outcomes [12].

In low-income nations such as Ethiopia, where most of the population lives in rural areas with limited infrastructure, the distance to health care is out of reach for many households. Despite the positive efforts made by the Ethiopian government to improve health outcomes by boosting the number of health workers and expanding health facilities, health service utilisation remains low. In 2016, only 31.8% of Ethiopian pregnant women received prenatal care, with 26.2% of deliveries taking place in health facilities [14].

Ethiopia has had persistently high neonatal mortality rates with considerable disparities by geography and administrative regions [14, 15]. In our previous study, we found that people experienced difficulties accessing emergency newborn care, with only 46.5% of all births in Ethiopia having access to emergency neonatal care facilities within two hours of travel time. The geographical distribution of healthcare facilities as, well as availability of roads, transport and communication are also highly variable across administrative regions and geographic areas. Addis Ababa and Harari administrative regional states had 100% emergency newborn care services coverage, while the Afar and Somali regional states had the lowest access with only 15.3% and 16.3%, respectively [16]. This indicates that pregnant women experience long distances to travel and challenging transportation conditions to access emergency neonatal care services.

The current literature on neonatal mortality in Ethiopia showed that determinants can be grouped into sociodemographic, healthcare-related, obstetric, and medical factors. Key contributing factors to neonatal mortality include area of residence, wealth, education, birth interval, maternal age, tetanus toxoid vaccination, breastfeeding, maternal health service use, place of birth and facility type, the performance of EmONC functions, premature rupture of membrane, gestational age, twin pregnancy, baby size, birth complications, mode of delivery, skilled birth attendance and postnatal health services use [17‐21]. We have identified key contributing factors to neonatal mortality in prior work on factors contributing to neonatal mortality that included a survey-based comprehensive analysis of socioeconomic, healthcare, and geographic factors [22]. In the current study, we have incorporated all significant contributing factors from the previous analysis. This study aimed to explore the effect of distance to health facilities on maternal service uptakes and neonatal mortality in Ethiopia.

We included a total of 6,941 (51.8% male and 48.2% female) most recent live births in the five years before the 2016 EDHS survey. There were 142 neonatal deaths in the study population, with a mortality rate of 21.5 (95% CI: 18.4, 25.1) per 1,000 live births. The neonatal mortality rates within the first, second, third and fourth quartiles of distance were 16.4 (95% CI:14.1 to 23.6), 18.2 (95% CI:12.8 to 25.7), 20.7 (95% CI:15.0 to 28.6), and 26.5 (95% CI: 19.9 to 35.2) per 1000 live births, respectively. Only 31.7% of births took place in health facilities, and 1057 (14.2%) of mothers applied dung, oil, butter or ointment items on their child's umbilical cord during the days following birth (see Table 1). The Euclidean distance in kilometres was computed between the 622 clusters and the nearest 480 health facilities. The median distance from the centre of clusters to health facilities was 3.8 km (Interquartile range: 1.5 km to 6.9 km). The shortest distance from the centre of a cluster to a health facility was 50 m, and the longest distance was 49 km. All the live births that were residing in areas farther than 18 km from health facilities were from emerging regional states (Somali, Afar, Benishangul Gumuz and Gambella regional states) and no live births of the other regional states reside further than 18 km.

The distance between residence and health-care facilities was found to have an impact on neonatal mortality. However, while use of health services and application of substances to umbilical cords were determinants of neonatal death in the prior study, they are not statistically significant in our current analysis. Closer proximity to health facilities was associated with lower neonatal mortality. These finding is in agreement with a previous study in Ethiopia [30] that found closer proximity to health facilities was associated with lower early neonatal mortality, as well as studies from Vietnam [8] and India [31] that showed closer proximity to health facilities had an association with lower neonatal mortality. However, our finding is inconsistent with a study conducted in Malawi and Zambia that showed no association between distance and neonatal mortality in Malawi, while in Zambia, ten km further distance was associated with lower neonatal mortality [4]. A finding from India also showed that neonatal mortality decreased as the distance from home to primary health care services increased [31]. This disparity could be related to a difference in statistical approach, as we examined the mediated effect of distance on neonatal mortality, whereas the former studies considered distance variables at the same level with maternal health service use variables. The other explanation could be that the difference in the quality of services among different countries.

There was no significant association between neonatal mortality and maternal health services use (ANC, FD and PNC). However, previous studies showed that antenatal care use and the place of delivery influenced neonatal mortality [8].

We also investigated distance to health facilities and its influence on the utilisation of antenatal care, facility delivery and postnatal care services. Our findings showed that the farther the distance to health services, the lower the utilisation of antenatal, perinatal and postnatal counselling services. In agreement with our findings, studies from Kenya [32, 33], Ghana, Malawi and Zambia showed that further distance to care was negatively associated with facility use for delivery services [4, 34‐37]. Distance to health facilities was also associated with lower use of antenatal care or a delay in attending ANC visits. In support of our finding, a systematic review in sub-Saharan Africa showed that further distance to health facilities reduces the uptake of antenatal care service use [38]. Another study from northwest Ethiopia showed that distance to health facilities caused a delay in ANC attendance [39], and a study in Burkinafaso found that proximity increased the odds of getting four or more ANC visits[36]. This is because distance can determine the accessibility to health facilities, cause delays to care, and influence the ability to use health services during pregnancy, at birth, and during the postnatal period [2].

Attending four or more antenatal care was found to influence facility delivery, getting counselling services and postnatal care service use. A systematic review in Ethiopia found that mothers who received one or more antenatal care visits were more likely to deliver at health facilities and use postnatal care services, which was consistent with our findings [13]. Studies from the Ethiopian DHS on factors associated with facility delivery and postnatal care service use also showed that attending four or more ANC was found to increase the probability of facility delivery and postnatal care use [40, 41]. This is because ANC is an important entry point for subsequent use of delivery and PNC services; hence, it promotes women to have a birth preparedness and complication readiness plan [42] and then give birth at a health facility and return for a postnatal visit [13, 43].

We found that women residing in rural areas were less likely to use antenatal care, facility delivery, and counselling services on neonatal danger signs and postnatal care services than women residing in urban areas. In agreement with our finding, a study from the 2011 Ethiopian DHS further analysis showed that women from urban areas were more likely to use ANC, facility delivery and postnatal care services than women from rural areas [41]. A systematic review and meta-analysis in low- and middle-income countries also showed that postnatal care service use was lower among women from rural areas [44]. This can be related to the health services accessibility and better infrastructure coverage in urban areas. The rural areas in Ethiopia suffer from shortages of basic health services, road, electricity and water infrastructures [45]. We found that giving delivery at a health facility increased the probability of postnatal care service use. A study from Uganda on determinants of early postnatal care use showed that facility delivery was associated with increased uptake of postnatal care service use [46]. It can be because health professionals can educate women and their families about healthy parenthood and the necessity of postnatal care during birth. Even though it was not statistically significant, the association between preterm pregnancy and maternal health service use was strange. Preterm births were less likely to seek postnatal care and counselling services than term births. This could be related to mothers' fear of social isolation due to having small babies and unfavourable attitudes about their baby in the early days of birth [47].

Distance to health services was found to impact the use of maternal services (such as antenatal care visits, perinatal care service use) as well as neonatal survival. A longer distance to the nearest health facility can be an obstacle to seeking care. In Ethiopia, majority of the community are disadvantaged as they often lack road infrastructure and transportation. Antenatal care use was associated with facility delivery service use, whereas both antenatal care and facility delivery services positively influenced postnatal care. The GSEM model simultaneously depicts the relationship between distance, maternal health service utilisation and neonatal mortality. The association between distance and infant mortality was shown to be mediated by factors such as antenatal care use, facility delivery, and postnatal counselling services. This supports our hypothesis that maternal health service use mediates the relationship between distance and neonatal mortality. Policymakers can leverage the findings from this study to improve health services coverage. Increased access to maternity and neonatal health care should be prioritised in the rural community. Policymakers and programs should contemplate mapping the updated distribution of health care facilities, road and emergency transport access, and population groups beyond the timely reach of emergency newborn care. Further consideration should be given to geographic disparities when planning facility expansions and service distribution. Finally, it would be beneficial to investigate the impact of quality on maternal service utilisation in addition to geographic access.