The “Birth in Brazil” study is the first nationwide hospital-based study that assessed the structure of public and private maternity services in Brazil. It used a sampling process that can generate estimates for the five macro-regions of the country, type of service (public, private and mixed) and service location (capital or non-capital), which are relevant to policy makers and those involved in the planning and delivery of childbirth care in Brazil.
However, the present analysis has some limitations. The Birth in Brazil study only examined hospitals that handle more than 500 deliveries per year, which is where more than 80 % of births occur nationally. Hospitals with lower birth volumes were not included and could represent a group with inferior care infrastructure, resulting in overestimation of the overall level of care.
We obtained data related to the structure of hospitals through interviews with hospital directors, not by direct observation on the part of the researchers, leading to the introduction of potentially significant biases, if hospital directors reported a more adequate structure, which could have led to an overestimation of adequacy.
Private hospitals and women who attended mixed hospitals, but who had their care paid for by private insurance, were not included in this study. Although these women cared for in mixed hospitals had a higher prevalence of obstetric risk, this most likely did not significantly alter the estimated risk in mixed hospitals owing to the low proportion (11.7 %) of private-funding cases in these settings.
Finally, we used equal weights for all items assessed and the items may have different importance in obstetric care. However, all items are essential to the quality of maternal care and we didn’t find published studies that could inform the weights to use.
The results of this study demonstrate a low rate of adequacy of public and mixed hospitals in Brazil: we rated only 34.8 % of the hospitals as adequate. Public hospitals presented a better score for human resources except for the presence of a nursing director with specialized training in obstetrics in hospitals with ICU, where only approximately half of both public and mixed hospitals met this criterion. While other studies have concluded that good outcomes in labor and delivery care often depend upon the number of qualified professionals available [19, 20] the present study only assessed the presence of physicians and nurses. When tasks are under the direct responsibility of these professionals, such as exams for admission to maternity wards, the absence of physicians and/or nurses can make the detection of high-risk obstetric conditions much more difficult. In the same way, the lack of an available obstetrician and/or anesthesiologist on a maternity ward  can lead to errors in the management of evolving labor, and consequently, cause delays between the decision to intervene with cesarean section and the actual procedure . The disastrous effects of late care and insufficient professional attention on the needs of laboring women have been documented by various authors [5, 22].
Public hospitals also scored higher for support services, while mixed hospitals scored higher for medications. In hospitals without ICU, mixed hospitals also had a higher score for equipment. Overall, public hospitals with or without ICU scored higher than mixed hospitals. Specific characteristics of the private sector can explain some of these differences, such as the contract of support services, instead of having these services at the hospital. However, these differences can affect the quality of care. Ambulance services were least available in mixed hospitals without ICU that also had deficiencies in human resources, blood banking or transfusion units, and laboratory and clinical pathology facilities. These issues highlight inequalities in the distribution of staffing, materials and services indispensable to the proper support of labor and birth care.
Hospitals without Intensive Care Unit beds had lower global rates of adequacy: we rated only 24.4 % of public hospitals and 10.6 % of mixed hospitals without ICU as adequate. Public hospitals with an ICU scored better in all the assessed domains while mixed hospitals with ICU scored better in human resources and support services. A similar observation was made by Magluta et al  after an investigation of maternity wards in Rio de Janeiro. Those authors concluded that the infrastructure of the hospitals improved as the level of complexity increased.
Indeed, in our study, nearly 40 % of public hospitals without ICU lacked the basic equipment needed for the care of women with obstetric emergencies. This same group of hospitals also had the highest rate of missing medications. Because maternal hemorrhage is one of the leading causes of maternal mortality in Brazil, it is worrying that 43 % of maternity wards without an ICU also lacked blood banks or transfusion units. This is a cause of great concern as Brazil has high rates of caesarian sections, even in low risk women , and caesarian sections are associated with increased risk of bleeding complications  and need of blood transfusion [25, 26].
The distribution of public and mixed services, according to the availability of ICU, was unequal among the Brazilian macro regions. Public hospitals with ICU were also less available in non-capital cities. Leal and Viacava  pointed out this situation in an analysis of Brazilian maternity wards conducted in the year 1999. In that study, ICU facilities were less available in non-capital cities and in less developed regions (North and Northeast). The persistence of this unequal distribution after more than a decade reinforces the need of investments in maternity care in these regions that have the highest burdens of maternal morbidity and mortality [15, 28]. These are also the regions where national studies have uncovered failures of the public system to adequately provide prenatal care and patient transfer to center-of-excellence hospitals, which are often located in the capitals [4, 28, 29].
Studies have suggested that great strides in maternal health can be made by guaranteeing women with obstetric complications access to maternity wards that are qualified to meet their needs [7, 30]. Therefore, it is problematic that, according to our findings, hospitals without ICU or in those with inadequate or partially adequate ICU  admitted a high percentage of women at obstetric risk. Previous studies have noted a limited access to hospitals with a higher level of care in the city of Belo Horizonte , in the state of Paraná  and in the public hospital system .
We identified a high proportion of women classified as low obstetric risk admitted to labor and childbirth care in hospitals with ICU. This was more frequent in hospitals located in the Southeast/South/Central-west regions and in capitals, reaching 70.6 % among mixed hospitals located in capital cities. This type of allocation represents unnecessary spending of resources and potential exposure of low-risk pregnant women to unneeded interventions . Results from the Birth in Brazil study regarding obstetric interventions carried out in Brazilian maternity wards show that pregnant women are subjected to routine procedures that do not take into account their individual needs . A study carried out in the United Kingdom compared care setting to maternal and neonatal outcomes in low-risk women; those who gave birth at freestanding midwifery units or alongside midwifery units were not at greater risk for complications, lending support to the lower use of obstetric interventions for pregnant women attended by these types of services .