In Belo Horizonte, Brazil, perinatal mortality was higher in hospitals that provided care under government contracts than in private hospitals that did not receive such funding, according to a cohort study of nearly 41,000 births in 24 facilities.1 Mortality was also elevated in hospitals that scored lower on a quality measure based on structural capacity to provide maternal and infant care. When maternal education and birth weight were controlled for, infants born in government-funded hospitals had an elevated relative risk of perinatal mortality (2.1–3.1), and infants delivered at hospitals having a low quality of care were at greater risk than those delivered at hospitals of adequate quality (1.9).

Although almost all births in Brazil occur in hospitals, with a large majority assisted by doctors, the country has persistently high infant mortality (23 deaths per 1,000 live births in 2003). Most infant deaths in the country are considered preventable through access to better hospital care; however, few studies have assessed the relationship between socioeconomic status, quality of hospital care and perinatal mortality. This study assessed how mothers' socioeconomic status and hospital use influenced perinatal mortality. Almost 80% of Brazilians receive health care provided by the country's Universal Public Health System (Sistema Único de Saúde, SUS) through contracts with private hospitals and hospitals run by philanthropies and the government. The remainder, who can afford private health insurance or direct payment, use private, non-SUS hospitals.

The data for this study were drawn from a 1999 cohort study of 40,953 live births and 775 perinatal deaths. Perinatal deaths were defined as fetal deaths and infant deaths occurring within seven days of birth (with a birth weight of at least 500 or a gestational age of at least 22 weeks). Hospitals were classified by whether they provided care under a contract with the SUS system and by their quality of care; maternal education was used as an indicator of socioeconomic status.

Twenty-two percent of all live births took place in private, non-SUS hospitals; the rest occurred in private (40%), philanthropic (22%) and public (15%) SUS hospitals. Forty-five percent of births took place in hospitals with a low quality of care, 26% in those with intermediate quality and 24% in those with adequate quality. Eleven percent of all live births had a low or very low birth weight (less than 2,500 g).

Seventy-four percent of perinatal deaths occurred among low or very low birth weight infants, and about a third of deaths took place at hospitals of each level of care (32–35%). The large majority of perinatal deaths occurred at SUS hospitals (87%), with 36% at public SUS institutions; only 13% of deaths took place at private, non-SUS hospitals.

Overall, the perinatal mortality rate was 19 deaths per 1,000 live births. Compared with the rate at private, non-SUS hospitals, crude rate ratios for perinatal mortality ranged from 1.3 in private SUS hospitals to 4.2 in public SUS hospitals.

Slightly fewer than one-third (31%) of mothers who gave birth in SUS hospitals had eight or more years of education, compared with 85% of those who gave birth at private, non-SUS hospitals. For both more educated mothers and less educated mothers, rates of perinatal mortality were higher at public SUS hospitals (21.7 per 1,000 and 23.3 per 1,000, respectively) than at private, non-SUS hospitals (6.6 per 1,000 and 6.2 per 1,000). Compared with the rates in private, non-SUS hospitals, rate ratios for perinatal mortality ranged from 1.2 among more educated mothers at private SUS hospitals to 3.5 among less educated mothers at public SUS hospitals.

Rates of perinatal mortality from specific causes varied by hospital type and by birth weight: Among low-birth-weight infants, for example, the rate of perinatal death from asphyxia (36.7 per 1,000 live births overall) ranged from 18.5 per 1,000 in private, non-SUS hospitals to 50.2 per 1,000 in public SUS hospitals, while the rate of death from immaturity (34.3 per 1,000 live births overall) ranged from 25.1 per 1,000 in philanthropic SUS hospitals to 48.0 per 1,000 in public SUS hospitals. Among infants of normal birth weight, the rate of death from asphyxia (2.7 per 1,000 live births overall) varied from 0.9 per 1,000 in private, non-SUS hospitals to 3.6 per 1,000 in private SUS hospitals, while the rate of death from immaturity (0.3 per 1,000 overall) varied from 0.0 in private, non-SUS hospitals to 0.6 in private SUS hospitals.

After controlling for maternal education and birth weight, multiple logistic regression analysis found that, compared with infants delivered in private, non-SUS hospitals, those born in private or philanthropic SUS hospitals had an elevated risk of perinatal mortality (3.1 and 2.1, respectively). In addition, mortality was higher at hospitals with low quality-of-care scores than at those scored as having adequate quality (1.9).

According to the researchers, even though most Brazilian women give birth in a hospital, nontimely access to care and low-quality care during delivery and the neonatal period present significant risks of perinatal mortality. They recommend that public services be increased at SUS-funded hospitals and that government resources be directed toward expanding the availability of intermediate-quality facilities in an effort to reduce the national perinatal mortality rate. They conclude that to address the unacceptably high rates of perinatal and neonatal mortality in Brazil, it is critical to tackle the "socioeconomic factors that contribute to inequities in perinatal mortality [and] to improve the quality of health care delivered to women and their babies at the health system level. "—J. Thomas "


1. Lansky S, França E and Kawachi I, Social inequalities in perinatal mortality in Belo Horizonte, Brazil: the role of hospital care, American Journal of Public Health, 2007, 97(5):867–873.