Inequalities in the coverage of place of delivery and skilled birth attendance: analyses of cross-sectional surveys in 80 low and middle-income countries

Place of delivery and type of attendant are both important determinants of maternal and newborn health [7, 26]. The purpose of this study was to estimate the frequencies of (a) SBA coverage, (b) institutional delivery, and (c) the combination of place of delivery and type of attendant in LMICs. Similar analyses were recently published for 57 countries using DHS data, with a focus on the private and public sectors in four world regions [40]. We present analyses for 80 countries with data from both MICS and DHS, stratified by place of residence and wealth asset index. Our results are analyzed according to seven world regions using the UNICEF classification. Place of delivery and type of attendant were combined into four categories of delivery assistance, to allow for the possibilities that deliveries in health facilities are not necessarily performed by SBAs, and that deliveries outside a health facility do not necessarily represent unskilled birth attendance. We did not find any systematic studies on these four categories of delivery assistance in the literature.

We show that coverage of SBA at delivery varies widely across LMICs. In 25 out of the 80 countries studied, the national coverage was over 90.0 %; these countries have reached the 2015 target proposed by the United Nations General assembly in 1999 [22, 41]. On the other extreme, 11 low-income countries [42] have national coverage below 40 %. These results are consistent with recent findings in the literature [29, 40, 42, 43]. In addition, pro-urban and pro-rich inequalities were documented, both at the national and regional level. These results are also consistent with other recent publications [29, 44, 45].

When we analyzed the correlations between SBA coverage and institutional deliveries at country level, two groups of countries were outliers. Azerbaijan, Cambodia, Indonesia, Philippines, Iraq and Tajikistan the prevalence of home SBA deliveries was over 10.0 %, being higher in rural areas and in the poorest quintiles. In Senegal and Togo, the proportion of births in a health facility by unskilled birth attendants was above 10.0 %, being higher in the poorest quintiles and – in Togo – in rural areas. We sought to understand why these categories are so prevalent in these two groups of countries.

Since the launch of the Safe Motherhood Initiative in 1987 and of the Millennium Development Goals in 2000, countries have adopted different strategies to improve maternal and newborn health outcomes [10, 46, 47]. In several countries, delivery by SBAs outside health facilities have been promoted. In the Philippines, “birthing homes” supervised by public or private health facilities provide birthing services including antenatal, normal spontaneous delivery, and postnatal care, particularly in rural and poor population [44, 45]. This service is provided by accredited health personnel, generally midwives, with a minimum of 2 years training [48‐50]. In Indonesia, the situation is very similar except that the duration of the training program is one year [51, 52]. In Cambodia, no formal program about home SBA is reported in the literature. According to Por et al., home SBA deliveries are driven by economic interest from government health personnel, who can charge women for the delivery at home, but not in a public facility [53, 54]. In Tajikistan, health care has deteriorated dramatically since 1991 [55, 56], and most rural maternities are operating without the basic conditions such as heat or running water. Thus, women consider giving birth at home safer, more comfortable and affordable than in a facility [55‐57]. In contrast, in Azerbaijan, a system known as “feldsher-accoucher points” staffed by mid-level health care providers who focus on primary health care in rural areas is in charge of assisting deliveries at home [57, 58]. In Iraq, due to decades of war and economic sanctions, many health facilities have faced serious difficulties in keeping functioning and providing adequate health care [59]. Besides, the fear of terrorist attacks has reportedly led patients to avoid public spaces such as health facilities, and some women feel more secure by requesting a health care provider to assist their births at home [59]. These are likely the reasons why home SBA deliveries are so prevalent in these countries, especially in rural areas and in the poorest quintiles.

A different pattern was observed in West-African countries such as Senegal and Togo, where more than 10 % of institutional deliveries were carried out by unskilled birth attendants. According to Kodio et al., in some villages in Senegal, no qualified midwife or nurse are presented in the health facility and deliveries are generally assisted by a traditional birth attendant (known as “matrone” in French or a community health worker (CHW)) [32]. In Togo, the situation is not different and health post at village level provide maternity services that are often run by matrones or other voluntary community health workers [60]. Institutional deliveries by unskilled attendants were also common – although not as much as in Togo and Senegal – in other African countries such as Benin (4 %), Central African Republic (5 %), and The Gambia (6 %).

In 1997, the World Health Organization document stated that “birth can take place in a range of appropriate settings, from home to tertiary referral centres, depending on availability and need [61]. It also recognized that home delivery could be appropriate for normal deliveries, provided the person attending the delivery is well trained and equipped [61]. Over time, the focus on SBA moved from coverage to quality of care in facilities, as many countries adopted strategies for promoting institutional deliveries. The World Health Report 2005 promoted care close to home – e.g. with midwives deployed in health centers and referral backup hospitals staffed by doctors, nurses and midwives [62]. Over time, the focus is also changing from measuring coverage to assessing and improving quality of care for facility births [63]. The shift from coverage to quality is at least in part motivated by studies showing that increased coverage of SBA and/or institutional deliveries does not necessarily improve maternal or newborn outcomes [64‐68]. The growing focus on institutional birth should take into account the fact that in a few countries, many such deliveries are carried out by unskilled birth attendants, and therefore do not contribute to achieving the recommendations for increasing SBA coverage [18].

There is little evidence about the effectiveness of home SBA deliveries in the literature. Most research regarding home birth strategies has focused on the training of TBAs [15, 21], while home SBA deliveries have received little attention. In our study, pro-poor and pro-rural inequalities on home SBA deliveries at the national level were observed, consistently with previous publications [55, 69]. In spite of the existing inequalities among home SBA utilization, information on the quality of care provided at home remain scant. Constraints encountered by SBAs during home delivery are numerous and include inappropriate environment for delivery, insufficient supplies and equipment, lack of security, inadequate training for home delivery, lack of transportation for referrals, and the social pressure in life-and-death situations, all of which affect the quality of care [25, 70]. Some authors have stressed the limitations of home SBA deliveries for reducing maternal and newborn outcomes [51, 64]. On the other hand, several advantages have been reported for home SBAs deliveries, such as lower rate of medical interventions (episiotomy, forceps, vacuum extraction among others), social support, privacy and higher proportions of birth receiving skin-to-skin practices in immediate breastfeeding within one hour after birth [24, 71, 72]. Assessing the quality of home deliveries by SBAs is beyond the scope of the present analyses, but our contribution lies in highlighting that such births are common in several countries, and deserve further evaluation.

Our analyses have some limitations. The first is related to the definition of SBA. As initially proposed by WHO, SBAs included doctors, nurses and midwives [18]. In many countries, others cadres such as auxiliary nurses, auxiliary midwifes, community health workers, and even TBAs or matrones who received some degree of formal training may be considered as skilled, making it difficult to compare among countries [22]. We believe that analyses such as ours may help identify countries with unusual patterns, which may be associated with non-standard definitions of SBA. Second, SBA coverage in national surveys was based on maternal reports, and some women may be unable to provide accurate information. For example, Hussein et al. described the difficulties of women in discriminating accurately among different types of birth attendants in Ghana [73]. We believe that this situation could be similar in others countries, especially where SBA coverage is low. On the other hand, studies on the validity of self-report SBA question during delivery conducted in Kenya and Mexico have shown that, while this indicator is not recommended for use at the individual level, it could be used to generate acceptable estimate of SBA coverage at population level [74, 75]. Third, our analysis uses available data for countries for the period 2005–2014, but only three surveys were carried out prior to 2007, and with a few exceptions SBA coverage is increasing slowly in most countries [76]. Because not every country was included, there may be additional examples of the outlier patterns we described above. Fourth, wealth asset indices were used to analyze economic status; such indices may vary according to the choice of assets and are affected by issues of comparability between urban and rural household [77, 78]. In addition, wealth quintiles are relative measures, that is, the poorest quintile in a middle-income country might be the wealthier than the third or fourth quintile in an extremely poor country [27].

Nevertheless, use of asset indices allows systematic comparisons of inequalities in health that would not be possible with other more complex measures of socioeconomic position [27]. Finally, we defined inequality based on wealth asset indices and place of residence, a common tool for evaluating inequalities within populations [78]. Other determinants such as education, distance to health facility, ethnicity, occupation and religion, among others may be equally or even more important in affecting access to delivery care.

The above limitations, however, are unlikely to affect our main conclusions. To our knowledge, this is the first study that analyzes inequalities in these four categories of delivery assistance in a large number of LMICs countries using both DHS and MICS survey databases.