Preprint
Article

This version is not peer-reviewed.

Racial Inequalities in Maternal Mortality: Population-Based Retrospective Study Using Data from the ReMaP Project

Submitted:

10 November 2026

Posted:

12 November 2026

You are already at the latest version

Abstract
Objective: To investigate racial disparities in maternal mortality and access to obstetric care in the state of São Paulo, comparing maternal health outcomes and access to healthcare services between Black (Black and Brown) and White women. Methods: This was a retrospective, cross-sectional, population-based observational epidemiological study conducted using secondary data extracted from the Restructuring and Redesign of the Maternal and Neonatal Care Network (ReMaP) panel, integrated into the Brazilian Obstetric Observatory. A total of 6,781 maternal deaths registered in the state of São Paulo between 2000 and 2025 were analyzed. Variables included race/color, education level, marital status, type of maternal death, period of death, medical care, place of occurrence, and investigation by Maternal Mortality Committees. Descriptive analysis was performed using absolute and relative frequencies, with calculation of 95% confidence intervals. Conclusions: Maternal mortality in São Paulo remains associated with significant social and healthcare inequalities, with a predominance of potentially preventable causes and a high occurrence in the postpartum period. The findings reinforce the need to strengthen epidemiological surveillance, improve obstetric care, and expand strategies aimed at promoting equity in maternal health.
Keywords: 
;  ;  ;  ;  ;  

1. Introduction

Maternal mortality remains one of the most striking expressions of social inequalities and health inequities. Although most deaths related to pregnancy, childbirth, and the postpartum period are considered preventable through timely interventions and qualified access to health services, significant differences persist between population groups defined by race, ethnicity, income, and territory. The social and structural determinants of health, including poverty, spatial segregation, education, institutional discrimination, and unequal access to services, directly influence the occurrence of maternal morbidity and mortality, making maternal mortality an important marker of social and health injustice [1,2].
In recent years, international organizations have highlighted that the sustainable reduction of maternal mortality requires approaches that transcend clinical care, incorporating strategies aimed at addressing the structural inequalities that affect women in vulnerable situations. Structural factors related to racism, social exclusion, and living conditions are significant determinants of the persistence of unfavorable maternal outcomes, even in contexts where formal obstetric services are available [1,2].
In Brazil, despite the progress observed after the implementation of the Unified Health System (SUS) and the expansion of prenatal and hospital care coverage, maternal mortality continues to show significant racial inequalities. Recent national analysis demonstrated that Black women had higher Maternal Mortality Ratios (MMR) in all years analyzed between 2017 and 2022, across all geographic regions, age groups, and main obstetric causes. Race/color remains a significant determinant of maternal outcomes in the country, highlighting the inadequacy of universal policies when not accompanied by explicit strategies to promote racial equity [3].
National literature also demonstrates that racial disparities in maternal mortality cannot be explained solely by biological or behavioral differences. On the contrary, such inequalities result from the interaction between structural racism, socioeconomic inequalities, reduced access to specialized services, and differences in the quality of care received. Black women frequently face greater barriers to accessing adequate prenatal care, are more likely to travel long distances in search of childbirth assistance, and are more exposed to diagnostic and therapeutic delays in the face of obstetric complications. These factors contribute to an increased risk of maternal death even in regions with ample availability of health services [2,3,4,5].
A recent retrospective study involving more than forty thousand maternal deaths in Brazil showed that approximately 60% of deaths occurred among Black women. Furthermore, the maternal mortality ratio was about 39% higher in this group compared to white women. The authors also identified an association between maternal mortality, low education levels, residence in socially vulnerable regions, difficulties accessing services, and less investigation of deaths, reinforcing the role of structural inequalities in determining these outcomes [4].
The discussion on racial equity in maternal health is closely related to the concept of structural racism. Structural racism operates through institutional mechanisms that influence the unequal distribution of resources, opportunities, and services, directly affecting access to and quality of care received by Black women throughout the pregnancy and postpartum period [2,6,7].
In obstetric care, the availability of services does not necessarily guarantee equitable access. Access involves multiple dimensions, including availability, geographic accessibility, acceptability, cultural appropriateness, quality of care, and the problem-solving capacity of services. Women belonging to racially marginalized groups often face additional barriers even when residing in areas with a wide range of obstetric care services, indicating that institutional and organizational inequalities can limit the benefits produced by the expansion of healthcare infrastructure [8,9].
In this context, the organization of Health Care Networks (HCNs) assumes strategic importance for reducing preventable maternal deaths. Regionalized and integrated obstetric networks have the potential to expand timely access to services, reduce care delays, and ensure adequate flows between primary care, routine risk maternity hospitals, specialized services, and referral centers for obstetric emergencies. However, the effectiveness of these networks depends on the territorial distribution of resources, regional governance, care coordination, and the capacity to respond to the specific needs of historically vulnerable population groups [10,11,12].
In the state of São Paulo, the organization of obstetric care occurs through Health Regions, Regional Health Departments (DRS), and Regional Health Care Networks (RRAS), structures that seek to promote greater integration of services and reduce territorial inequalities. In this context, the Restructuring and Redesign of the Maternal and Perinatal Care Network (ReMaP), developed in collaboration with the Brazilian Obstetric Observatory (OOBr) and the São Paulo State Department of Health, represents a strategic initiative to strengthen maternal and perinatal healthcare through evidence-based planning and regional health system management. The initiative involves the integration and analysis of multiple public health and socioeconomic databases, including the Live Birth Information System (SINASC), the Mortality Information System (SIM), the Hospital Information System (SIH/SUS), the National Registry of Health Facilities (CNES), and data from the Brazilian Institute of Geography and Statistics (IBGE). Using data science techniques, artificial intelligence, and machine learning, ReMaP seeks to identify regional disparities, epidemiological trends, structural deficiencies, and gaps in the availability and distribution of obstetric and neonatal services. Its operational activities include the development of interactive dashboards, continuous monitoring of maternal, fetal, and neonatal health indicators, identification of critical points in the healthcare network, and the provision of evidence to support decision-making by state and municipal health managers. Through these strategies, the initiative aims to inform the regional reorganization of healthcare services, improve the coordination of referral and counter-referral pathways, and support the development of interventions tailored to the specific needs of different territories [11].
Despite investments made in organizing the Maternal and Perinatal Network (ReMaP) in the state of São Paulo (Brazil), it remains unknown whether strengthening the obstetric network has effectively contributed to reducing racial inequities in maternal mortality and expanding equitable access to obstetric care services between Black and white women. The persistence of high racial disparities in maternal mortality highlights the need to assess not only the magnitude of the problem, but also the effectiveness of the public policies implemented to address it. Although the literature shows that Black women have a higher risk of maternal death and face additional barriers to accessing health services, there are still important gaps regarding the role of regionalized obstetric care networks in reducing these inequalities. Thus, the objective is to investigate racial disparities in maternal mortality and access to obstetric care in the state of São Paulo, comparing maternal health outcomes and access to healthcare services between Black (Black and Brown) and White women.

2. Materials and Methods

2.1. Study Design

This is a population-based, cross-sectional, retrospective, analytical-quantitative epidemiological study. The methodological design followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) [13] methodological guidelines for structuring cross-sectional studies.

2.2. Location, Population, and Study Period

The study setting encompassed the national epidemiological territory of Brazil, using geographic and institutional data from official health information systems maintained by the Brazilian Ministry of Health and the Brazilian Institute of Geography and Statistics (IBGE). The primary data sources included the Mortality Information System (SIM), the Live Birth Information System (SINASC), the Hospital Information System of the Unified Health System (SIH/SUS), and the National Registry of Health Establishments (CNES), complemented by IBGE demographic and territorial databases. These systems provided information on maternal deaths, live births, hospital admissions, the distribution of healthcare facilities, and population characteristics, enabling the assessment of regional patterns and inequalities in maternal mortality and access to obstetric care.
The target population consisted of all deaths among women of childbearing age (WCA), defined as women aged 10–49 years, registered in the SIM of the Brazilian Ministry of Health across all Brazilian states during the study period. Following the application of the eligibility criteria, the final study population comprised 6,781 notified cases eligible for epidemiological investigation. As this was a population-based retrospective study employing a census approach, all maternal deaths registered during the study period were included, and therefore, no a priori sample size calculation was required. This approach ensured the inclusion of all eligible cases within the defined study population, maximising the use of available epidemiological data while avoiding sampling error associated with sample selection.

2.3. Eligibility Criteria and Sample Selection

For the composition of the case series, the following selection criteria were adopted: death records of female individuals; chronological age equal to or greater than 10 years and less than 50 years at the time of death; declaration of the underlying cause or causes associated with death coded as maternal death according to the criteria of the International Statistical Classification of Diseases and Related Health Problems (ICD-10). This included direct obstetric causes (obstetric complications resulting from interventions, omissions, incorrect treatment, or a chain of events resulting from any of these causes), indirect causes (resulting from pre-existing diseases or those developed during pregnancy, not resulting from direct obstetric causes, but aggravated by the physiological effects of pregnancy), and unspecified causes. The primary outcome was maternal death, defined by the World Health Organization (WHO) as the death of a woman during pregnancy or up to 42 days after the end of pregnancy, regardless of the duration or location of the pregnancy, due to any cause related to or aggravated by the pregnancy or by measures taken in relation to it, excluding accidental or incidental causes. Additionally, late maternal deaths, occurring between 43 days and less than one year after the end of pregnancy, as classified by ICD-10, were included.
Exclusion criteria included: records with systemic duplicate notifications (same Death Certificate number); data with missing or inconsistently filled biological fields (age or sex) (irremediable biological incompatibility); and deaths of women in the described age range whose causes were explicitly linked to strictly accidental or incidental factors unrelated to the pregnancy-puerperal cycle.

2.4. Study Variables

The variables of interest were extracted and categorized in a standardized way to mitigate classification biases. The variables were structured along two analytical axes: sociodemographic profile (race/color: self-assigned classification collected nominally and stratified into six regulated categories: White, Brown, Black, Yellow, Indigenous, and Unknown; education: measured by the last formal level of education completed, ordinally segmented into: no schooling, elementary school I, elementary school II, high school, incomplete higher education, complete higher education, and unknown; marital status: legal or declared marital status at the time of death, categorized as: Single, Married, Common-law marriage, Legally separated/Divorced, Widowed, and Unknown).
For the clinical, obstetric, and care profile, the following were considered: type of maternal death: Physiopathological classification of death categorized as Direct Obstetric, Indirect Obstetric, or Unspecified; chronological period of death: Temporal moment of death in relation to the gestational cycle, subdivided into: During pregnancy, childbirth, or abortion; During the puerperium, up to 42 days; During the puerperium, from 43 days to less than 1 year; Not during pregnancy or puerperium; Inconsistent period; Not informed or ignored; death in childbearing age: Binary confirmatory variable (Yes or No) evaluating the inclusion in the reproductive window of 10 to 49 years; ICD-10 chapter: etiological grouping of death based on the structural chapters of ICD-10, highlighting: Chapter XV (Pregnancy, childbirth and puerperium); Chapter I (Certain infectious and parasitic diseases); Chapter V (Mental and behavioral disorders); Chapter II (Neoplasms/tumors); and Chapter IV (Endocrine, nutritional and metabolic diseases); medical assistance: care indicator of access to professional care prior to death, dichotomized/categorized into: With assistance, Without assistance and Unknown; place of occurrence: institutional or geographical environment where the death was ascertained: Hospital, Other health establishment, Home, Public road, Other and Unknown.

2.5. Data Collection

The data were obtained through the Restructuring and Redesign of the Maternal and Neonatal Care Network (ReMaP) panel, an analytical platform integrated with the Brazilian Obstetric Observatory (OOBr) [14] ReMaP is an epidemiological monitoring tool that gathers information from national health information systems, allowing the visualization, extraction, and analysis of indicators related to maternal and neonatal health. For this study, maternal death records available on the panel for the period 2000 to 2025 were used. Data extraction was performed systematically, encompassing sociodemographic, obstetric, clinical, and care variables available for public consultation. Because it is a consolidated database previously processed for epidemiological surveillance purposes, the data were only subjected to additional consistency, standardization, and organization procedures for analytical purposes.

2.6. Data Analysis

The statistical approach was divided into complementary descriptive and inferential steps: The descriptive analysis consisted of calculating absolute and relative frequencies. In terms of probabilistic modeling, for each individual category of a variable, the relative frequency was formally treated as the point estimate of the probability of success of a Bernoulli distribution (indicating mean of the category), where the value 1 represented belonging to the category and 0 non-belonging. To establish the mathematical precision and statistical uncertainty associated with each proportion estimate, the 95% Confidence Interval (95% CI) was calculated using the Wilson Score epidemiological method. This method was selected for its superior performance and robust coverage in large samples, overcoming the limitations of the Wald normal approximation for categories with very high or extremely low frequencies (close to 0% or 100%). The inferential analysis evaluated the intragroup distribution pattern. Initially, the overall chi-square goodness-of-fit test was applied to determine if the observed distribution in the macrovariables differed from an expected uniform pattern. Subsequently, to obtain the significance level of each individual row, a two-tailed exact binomial test was performed for each category separately. The null hypothesis tested assumed that the observed proportion of each category would be statistically equal to the theoretical probability of a balanced uniform distribution. A significance level of 5 was adopted for all tests, with results of p<0.05 considered statistically significant.

2.7. Ethical Aspects

The research was conducted in accordance with national ethical guidelines governing research involving human subjects. Since it exclusively used publicly available secondary data, without the inclusion of identifying variables, the research met the criteria for exemption from submission and analysis by the Research Ethics Committee System / National Research Ethics Commission (CEP/CONEP), as per the parameters established by Resolution No. 466 of December 12, 2012, both from the National Health Council [15].

3. Results

Between 2000 and 2025, 6,781 maternal deaths were recorded in the state of São Paulo. A predominance of self-declared white women (56.5%) was observed, followed by mixed-race (30.7%) and black (9.4%) women, while records among Asian (0.5%) and Indigenous (<0.1%) women were less frequent. The high participation of mixed-race and black women, who together accounted for 40.1% of deaths, highlights the persistence of racial inequalities in maternal mortality and reinforces the need for analyses focused on health equity. Regarding education, a high proportion of records were missing (60.7%), indicating a significant limitation in the quality of information. Among the cases with available information, secondary education predominated (20.3%), followed by middle school (9.4%) and completed higher education (5.3%). The low frequency of women with incomplete higher education (1.5%) and no schooling (0.2%) suggests that the educational profile of women who died as mothers was predominantly of intermediate school level, although the high incompleteness of the data requires caution in interpreting the results. Regarding marital status, single women predominated (47.3%), followed by married women (32.1%) and those in stable unions (7.1%). This pattern suggests greater vulnerability among women without formal marital ties, possibly related to differences in social support networks, socioeconomic conditions, and access to health services during the pregnancy-puerperal cycle (Table 1).
Analysis of clinical characteristics showed that the majority of maternal deaths were classified as having a direct obstetric cause (61.2%), followed by indirect causes (36.0%), while only 2.9% remained unspecified. This finding indicates that a large proportion of deaths occurred as a result of complications directly related to pregnancy, childbirth, or the postpartum period, many of which were potentially preventable through timely and qualified obstetric care.
Regarding the timing of death, almost half of the cases occurred during the postpartum period, up to 42 days after delivery (49.9%), followed by the period of pregnancy, childbirth, or abortion (26.9%). It was also observed that 5.4% of deaths occurred in the late postpartum period, between 43 days and less than one year after the end of pregnancy. These results reinforce the importance of continuous monitoring of women after childbirth, a period frequently neglected by health policies and services. All records corresponded to women of childbearing age. Regarding the underlying cause, there was a large predominance of deaths classified under Chapter XV of the ICD-10 (96.5%), relating to pregnancy, childbirth, and the postpartum period, while the other chapters showed residual participation. This result confirms the consistency of the database used and the predominance of causes directly related to maternal health.
Regarding medical assistance, 68.8% of women received assistance before death, while 2.5% did not receive assistance and 28.6% had missing information. Although most had some contact with health services, the high proportion of missing records limits a more accurate assessment of the quality of care provided. As for the place of death, almost all deaths occurred in hospitals (93.2%), followed by other health facilities (3.2%) and homes (2.6%). The predominance of deaths in hospital settings suggests that the most serious maternal complications reach specialized services, but also points to the need to evaluate the quality and timeliness of care offered in these facilities.
Table 2. Characteristics of maternal deaths according to type of death, period of death, ICD-10 chapter, medical assistance and place of occurrence, São Paulo, 2000–2025 (n=6,781).
Table 2. Characteristics of maternal deaths according to type of death, period of death, ICD-10 chapter, medical assistance and place of occurrence, São Paulo, 2000–2025 (n=6,781).
Variable/Category n (%)
Type of maternal death
Direct 4.147 (61.2)
Indirect 2,438 (36.0)
Not specified 196 (2.9)
Period of death
During the postpartum period, up to 42 days 3.383 (49.9)
During pregnancy, childbirth, or abortion. 1,824 (26.9)
Not informed or ignored 709 (10.5)
Inconsistent period 380 (5,6)
During the postpartum period, from 43 days to less than 1 year 365 (5,4)
Not during pregnancy or the postpartum period 120 (1.8)
Death during childbearing age
Yes 6.781 (100,0)
ICD-10 Chapter
XV – Pregnancy, childbirth and the postpartum period 6,547 (96.5)
I – Some infectious and parasitic diseases 227 (3,3)
V – Mental and behavioral disorders 4 (<0.1)*
II – Neoplasms (tumors) 2 (<0.1)*
IV – Endocrine, nutritional and metabolic diseases 1 (<0.1)*
Medical assistance
With assistance 4,668 (68.8)
Missing data 1,941 (28.6)
Without assistance 172 (2.5)
Extracted from the Restructuring and Redesign of the Maternal and Neonatal Care Network (ReMaP) panel, integrated into the Brazilian Obstetric Observatory (OOBr), 2000–2025. *Percentages calculated in relation to the total number of maternal deaths that occurred during the period (n=6,781). Values ​​less than 0.1% were presented as “<0.1”.
Analysis of maternal death investigations revealed that 32.4% of cases were investigated by Maternal Mortality Committees (MMCs), while 24.0% were not investigated and 43.6% had no information regarding the investigation. The high proportion of records without information highlights weaknesses in maternal mortality surveillance and monitoring systems, potentially compromising the identification of determinants and opportunities for preventing these events. Regarding the place of occurrence, hospitals again predominated, accounting for 93.2% of deaths, regardless of the investigation status. Deaths occurring at home (2.6%) and in other health facilities (3.2%) represented substantially smaller proportions. This pattern demonstrates that maternal mortality remains heavily concentrated in the hospital environment, reinforcing the relevance of improving obstetric care and monitoring deaths as fundamental strategies for reducing preventable maternal mortality.
Table 3. Distribution of maternal deaths according to place of occurrence and investigation by the Maternal Mortality Committee (CMM), São Paulo, 2000–2025 (n=6,781).
Table 3. Distribution of maternal deaths according to place of occurrence and investigation by the Maternal Mortality Committee (CMM), São Paulo, 2000–2025 (n=6,781).
Variable/Category n (%)
Investigation by CMM
No information 2,956 (43.6)
Yes 2,200 (32.4)
No 1.625 (24.0)
Location of occurrence
Hospital 6.318 (93.2)
Another health facility 220 (3.2)
Home 177 (2,6)
Others 53 (0.8)
Public road 12 (0.2)
Missing data 1 (<0.1)
A predominance of self-declared white women was observed, corresponding to 56.5% of the analyzed population (95% CI: 55.31–57.67), followed by mixed-race (30.7%; 95% CI: 29.62–31.81) and black (9.4%; 95% CI: 8.77–10.16). The categories Asian (0.5%) and Indigenous (<0.1%) presented the lowest proportions. Regarding education, a high proportion of missing records was found (60.7%; 95% CI: 59.53–61.86), highlighting significant incompleteness of this variable. Among the completed records, secondary education predominated (20.3%; 95% CI: 19.34–21.25), followed by middle school (9.4%; 95% CI: 8.74–10.13) and completed higher education (5.3%; 95% CI: 4.76–5.82). The lowest frequencies were observed among women with no schooling (0.2%) and those with incomplete higher education (1.5%). Regarding marital status, the highest proportion was observed among single women (47.3%; 95% CI: 46.08–48.45), followed by married women (32.1%; 95% CI: 31.00–33.23) and those in stable unions (7.1%; 95% CI: 6.55–7.77). The remaining categories showed frequencies below 4%, including ignored records (3.2%), separated/divorced (3.0%), and widows (0.9%). The simultaneous presence of male and female marital status categories (single, married, widowed) is noteworthy, suggesting possible distinct patterning of records throughout the analyzed period (Table 4).
Regarding the type of maternal death, direct maternal deaths predominated, accounting for 61.2% of cases (95% CI: 59.99–62.31), followed by indirect deaths (36.0%; 95% CI: 34.82–37.10). Deaths classified as unspecified represented only 2.9% of the sample (95% CI: 2.52–3.32). As for the period of death, approximately half of the cases occurred during the postpartum period up to 42 days after the end of pregnancy (49.9%; 95% CI: 48.70–51.08). Next, deaths occurring during pregnancy, childbirth, or abortion were the most frequent (26.9%; 95% CI: 25.86–27.97). A considerable proportion of records were also found to be unreported or ignored (10.5%; 95% CI: 9.75–11.21), in addition to records classified as inconsistent period (5.6%; 95% CI: 5.08–6.18) and deaths occurring between 43 days and less than one year after delivery (5.4%; 95% CI: 4.87–5.95). Deaths outside the pregnancy-puerperal period accounted for 1.8% of the records (95% CI: 1.48–2.11).
All deaths analyzed occurred in women of childbearing age, corresponding to 100.0% of the records (95% CI: 99.94–100.00). Regarding the classification of underlying causes according to ICD-10, there was a wide predominance of Chapter XV – Pregnancy, childbirth and the puerperium, responsible for 96.5% of deaths (95% CI: 96.09–96.96). Infectious and parasitic diseases (Chapter I) represented 3.3% of cases (95% CI: 2.95–3.80), while the other chapters presented frequencies below 0.1%. Regarding medical care, the majority of women received assistance before death (68.8%; 95% CI: 67.73–69.93). However, a high proportion of missing records was observed (28.6%; 95% CI: 27.56–29.71), which may limit more precise interpretations regarding access to care. Only 2.5% of cases occurred without registered medical assistance (95% CI: 2.19–2.94). Regarding the place of occurrence, hospitals concentrated the vast majority of maternal deaths (93.2%; 95% CI: 92.55–93.75), highlighting the predominantly institutional nature of these deaths. The remaining deaths occurred in other health establishments (3.2%; 95% CI: 2.85–3.69), at home (2.6%; 95% CI: 2.26–3.02), in other locations (0.8%; 95% CI: 0.60–1.02), and in public places (0.2%; 95% CI: 0.10–0.31). The number of missed records was residual (<0.1%) (Table 5).

4. Discussion

The results of this study show that maternal mortality in São Paulo remains strongly associated with social and racial inequalities, despite the state’s extensive healthcare infrastructure, which includes a large network of hospitals, specialized healthcare facilities, and high-complexity services. This paradox highlights that the availability and geographical distribution of healthcare resources do not necessarily translate into equitable access, timely obstetric interventions, or consistent quality of maternal care. Although White women represent the largest absolute proportion of deaths, the combined participation of black and mixed-race women reached approximately 40% of registered maternal deaths. This finding should be interpreted in light of the demographic composition and structural inequalities that condition access to, use of, and quality of obstetric care. Black women systematically present higher maternal mortality ratios, regardless of geographic region, age group, or cause of death, indicating that race/color remains one of the most important social determinants of maternal health in Brazil [3,4].
The high frequency of records with unknown schooling represents a significant limitation for epidemiological surveillance. The quality of information is a central component for monitoring the social determinants of maternal mortality and for planning public policies guided by equity. International evidence indicates that information systems with high levels of incompleteness tend to underestimate inequalities and hinder the identification of priority groups for intervention, compromising the formulation of effective strategies for preventing and reducing maternal mortality [16].
A complex epidemiological profile characterized by social and racial inequalities, a predominance of direct obstetric deaths, a substantial concentration of deaths during the postpartum period, and important weaknesses in the completeness of information and investigation of maternal deaths. Taken together, these findings indicate that maternal mortality should not be understood exclusively as a clinical outcome, but also as an indicator of the performance of healthcare networks, the distribution of social vulnerabilities, and the capacity of health systems to identify and prevent avoidable deaths. Although São Paulo has a comparatively extensive healthcare infrastructure, the persistence of maternal deaths demonstrates that the availability of services does not necessarily ensure equitable access, timely interventions, or adequate continuity of care [1,2,5,17].
The sociodemographic findings require careful interpretation. White women represented 56.5% of recorded deaths, while brown and Black women accounted for 30.7% and 9.4%, respectively. However, the absolute distribution of deaths cannot be interpreted as evidence that White women experienced a higher risk of maternal mortality, because the present analysis did not calculate race-specific maternal mortality ratios using the corresponding live-birth denominators. This distinction is essential when interpreting population-based mortality data. Silva et al. [3], analyzing Brazilian national data from 2017 to 2022, found that the maternal mortality ratio among Black women was almost twice that observed among White women, with disparities persisting across geographical regions, age groups, and causes of death. A recent retrospective cohort study of maternal mortality in Brazil also reinforces the importance of examining mortality through a racial equity perspective [4]. Thus, the predominance of White women in the absolute counts in the present study should be interpreted alongside the population composition of São Paulo and the need to calculate race-specific risks, rather than as evidence that racial inequalities are absent.
These findings also need to be situated within the broader evidence on structural racism and unequal obstetric care. A systematic review by Hailu et al. [2] identified structural racism as an important determinant of adverse maternal health outcomes. In Brazil, Leal et al. [17], using data from the Birth in Brazil national survey, documented racial inequalities in prenatal and childbirth care. Compared with White women, Black women had higher odds of inadequate prenatal care, lack of linkage to a maternity facility, absence of a birth companion, and having to seek more than one facility for childbirth. Brown women also experienced disadvantages in selected indicators. These differences suggest that racial inequalities may operate through multiple stages of the maternal care pathway, from access to prenatal services to the treatment received during childbirth. The present study did not directly measure discrimination or quality of care by race; therefore, it cannot establish that these mechanisms caused the observed deaths. Nevertheless, the existing literature provides a plausible framework for interpreting maternal mortality as an outcome shaped by unequal social conditions and healthcare experiences [2,3,6,7,8,9,17].
The high proportion of missing educational information is another important finding. Education is frequently used as a marker of socioeconomic position and may help identify differences in access to information, material resources, and healthcare. However, 60.7% of records in this study lacked information on schooling, severely limiting the assessment of social gradients in maternal mortality. Peterson et al. [16] emphasize the methodological challenges involved in estimating maternal mortality from civil registration and vital statistics systems, including the consequences of incomplete and inaccurate data. Missingness should not be regarded as a merely administrative issue: when information is systematically absent, surveillance systems become less capable of identifying vulnerable groups and informing targeted policies. Nevertheless, the present analysis cannot determine whether missing educational information was more frequent among specific racial groups or types of death. Future analyses should investigate the pattern of missingness and consider sensitivity analyses or appropriate methods for incomplete data, where the available variables and assumptions permit.
The predominance of direct obstetric deaths, which accounted for 61.2% of cases, is particularly relevant from a prevention perspective. Direct obstetric deaths are associated with complications of pregnancy, childbirth, or the puerperium, including hypertensive disorders, hemorrhage, and pregnancy-related infections. In a global systematic analysis, Say et al. [18] found that direct obstetric causes accounted for approximately 73% of maternal deaths during 2003–2009, with hemorrhage, hypertensive disorders, and sepsis representing major contributors. Although these global estimates cover a different period and population, they demonstrate the continued importance of recognizing and treating acute obstetric complications. The proportion of direct deaths in São Paulo should therefore be interpreted as evidence of the continuing burden of obstetric complications, rather than as proof that every death was individually preventable. Establishing preventability requires clinical records, information on the circumstances of care, and systematic review of the events preceding death.
From a health-system perspective, preventing deaths from direct obstetric causes requires more than the availability of specialized facilities. It depends on the timely recognition of warning signs, appropriate risk assessment, access to effective treatment, reliable referral pathways, and coordinated responses to obstetric emergencies. The World Health Organization’s most recent global estimates continue to emphasize the substantial burden of preventable maternal deaths and the need for quality care before, during, and after childbirth [19]. In the context of São Paulo, these priorities are particularly relevant to the organization of regional maternal and neonatal care networks, including communication between primary care, maternity services, emergency departments, and referral hospitals. However, because the present study did not examine individual clinical pathways, delays, or specific treatment protocols, these mechanisms should be considered priorities for further investigation rather than explanations directly demonstrated by the data [10,11,12,18,19].
The timing of death represents another major contribution of this study. Almost half of all recorded deaths (49.9%) occurred during the postpartum period up to 42 days after the end of pregnancy. This finding reinforces the need to consider the puerperium as a period requiring active clinical monitoring, rather than a stage in which care can be reduced after childbirth. Postpartum complications may emerge after discharge, and their outcomes depend on whether women and their families receive clear guidance about warning signs, can access urgent assessment, and obtain timely treatment. Care pathways should therefore connect maternity discharge planning with primary healthcare follow-up, specialist assessment when indicated, and emergency services. These measures are especially important when pregnancy-related complications require continued management after the immediate obstetric episode.
The 5.4% of deaths classified as occurring between 43 days and less than one year after the end of pregnancy also warrants attention. This proportion should be interpreted cautiously because the study included substantial numbers of records with an unreported or inconsistent period of death, and because the classification of late maternal deaths depends on accurate identification of the relationship between pregnancy and the subsequent death. Nevertheless, evidence from outside Brazil supports the importance of examining mortality beyond the conventional 42-day period. In a population-based cohort study in Norway, Arshad et al. [20] investigated pregnancy-associated mortality during the first year after childbirth, highlighting the value of longitudinal follow-up and the analysis of health outcomes beyond the early postpartum period. Differences in health systems, study periods, and definitions limit direct comparisons with the present study, but the evidence supports the need to improve linkage between birth and death records and to investigate deaths occurring throughout the first postpartum year. Such approaches may help identify deaths associated with conditions aggravated by pregnancy that would otherwise be overlooked by routine maternal mortality monitoring.
The finding that 68.8% of women had received medical assistance before death should not be interpreted as evidence that access to care was adequate or that the care provided was effective. Contact with a healthcare service is a measure of utilization, not necessarily of clinical quality, timeliness, continuity, or patient safety. A woman may reach a facility but experience delays in diagnosis, treatment, referral, or escalation of care. Moreover, 28.6% of records had no information about medical assistance, limiting the extent to which this variable can be interpreted. The predominance of hospital deaths (93.2%) similarly indicates where deaths were recorded, but does not establish whether the hospital caused the death or whether the care provided was inadequate. Hospitals often receive women with the most severe complications; determining whether deaths were preventable requires information on severity at admission, clinical management, available resources, referral delays, and the progression of complications. Consequently, the results support the need to evaluate the quality of obstetric care, but do not independently demonstrate deficiencies in hospital performance.
The investigation of maternal deaths represents a further area of concern. Only 32.4% of cases were recorded as investigated by Maternal Mortality Committees, while 24.0% were reported as not investigated and 43.6% had no information about investigation status. The high proportion of missing information prevents a clear distinction between deaths that were not investigated and those for which investigation information was not entered into the database. Nonetheless, these findings identify an important limitation in the surveillance process. According to the World Health Organization [21], maternal and perinatal death surveillance and response should not end with the notification or classification of a death; it should also identify modifiable factors, formulate recommendations, and support actions to prevent similar events. Smith et al. [22], reviewing implementation experiences across countries, similarly emphasized the importance of reliable information systems, institutional support, a constructive and non-punitive review culture, and monitoring whether recommendations are implemented. In São Paulo, improving documentation of investigation status and strengthening the connection between mortality reviews and concrete changes in clinical practice could improve the usefulness of surveillance data.
The long observation period is a strength of this study because it provides a broad overview of the sociodemographic and care-related characteristics of maternal deaths in São Paulo. However, the 2000–2025 period also raises the possibility that changes in death certification, coding practices, health information systems, and the completeness of reporting affected the observed distributions. Without a detailed temporal analysis of these changes, it is not possible to determine whether differences across the entire period reflect changes in maternal health, changes in surveillance practices, or both. In addition, the descriptive design does not allow causal inferences, and the absence of individual-level data on income, pre-existing conditions, clinical severity, timing of interventions, and quality of care limits the analysis of mechanisms underlying maternal mortality. The results should therefore be interpreted as a characterization of recorded deaths rather than as estimates of the independent effects of race, education, or healthcare access on mortality risk.
Despite these limitations, the findings have important implications for public health policy and research. First, surveillance systems should improve the completeness and consistency of information on race/color, education, period of death, medical assistance, and maternal death investigation. Second, maternal mortality analyses should incorporate appropriate population denominators to calculate race-specific mortality ratios and examine inequalities across time and place. Third, maternal care networks should strengthen the continuity of care from prenatal follow-up through childbirth and the postpartum period, including mechanisms for identifying complications after hospital discharge. Finally, maternal death reviews should be linked to measurable corrective actions, professional training, service-level quality improvement, and monitoring of implementation. These priorities are consistent with evidence on structural determinants of maternal health, racial inequalities in obstetric care, and international recommendations for maternal death surveillance and response [1,2,3,17,19,21,22].
This study presents some limitations that should be considered when interpreting the results. Because it uses secondary data from the ReMaP/Brazilian Obstetric Observatory panel, it is subject to the inherent limitations of health information systems, including underreporting, data entry errors, and inconsistencies in the records. Of particular note is the high proportion of missing information for relevant variables, such as education level and investigation of deaths by Maternal Mortality Committees, which may have limited more in-depth analyses of the social and healthcare determinants of maternal mortality. Furthermore, the observational and retrospective design does not allow for establishing causal relationships between the analyzed characteristics and the occurrence of deaths. Finally, the absence of individual information on income, pre-existing clinical conditions, quality of care received, and time of access to services restricts a comprehensive understanding of the mechanisms that contribute to the observed inequalities.
Despite these limitations, the study makes relevant contributions to public health and the field of maternal health. It is a comprehensive analysis of 25 years of maternal mortality in the state of São Paulo, using a robust population-based database with broad coverage. The results allow for the identification of sociodemographic, clinical, and care patterns associated with maternal deaths, highlighting the persistence of racial inequalities and social vulnerabilities. Furthermore, the study reinforces the importance of strengthening regionalized obstetric care networks, improving the quality of maternal death surveillance, and enhancing the quality of records in information systems. The findings can support managers, health professionals, and policymakers in implementing strategies aimed at reducing preventable maternal mortality and promoting health equity, contributing to the achievement of the Sustainable Development Goals.

5. Conclusions

Maternal mortality in the state of São Paulo remains a significant public health challenge, highlighting the persistence of sociodemographic and healthcare inequalities throughout the analyzed period. The results demonstrated a predominance of deaths among white, mixed-race, and black women, a high frequency of direct obstetric causes, and a concentration of deaths in the postpartum period, especially within the first 42 days after delivery. Most deaths occurred in hospital settings and after some prior contact with health services, suggesting that access to healthcare alone is insufficient to guarantee favorable maternal outcomes. Therefore, further efforts are needed to improve the quality, timeliness, and effectiveness of obstetric care.
The findings reinforce the need to strengthen Maternal and Neonatal Care Networks by expanding integration across different levels of care, improving monitoring throughout prenatal care, childbirth, and the postpartum period, and enhancing mechanisms for the surveillance and investigation of maternal deaths. Furthermore, improving the completeness, accuracy, and consistency of health information systems is essential to address substantial missing data in available epidemiological records. Strengthening data collection, documentation, and reporting practices is particularly important to ensure the reliable identification of maternal mortality patterns, support the analysis of sociodemographic and racial inequalities, and inform evidence-based decision-making. Strategies aimed at addressing social and racial inequalities should also be incorporated into public maternal health policies, recognizing maternal mortality as a phenomenon strongly influenced by the social determinants of health. The results provide relevant evidence for monitoring maternal health in the state of São Paulo and can support managers, healthcare professionals, and policymakers in implementing actions to prevent avoidable maternal deaths, improve obstetric care, strengthen health information systems, and promote health equity, thereby contributing to the achievement of national and international maternal mortality reduction targets.

6. Patents

Author Contributions

Conceptualization, G.G.d.S.; methodology, G.G.d.S; software, G.G.d.S; validation, G.G.d.S.; formal analysis, G.G.d.S, G.d.S.C, K.C.P.M.V, C.M.P.d.S and C.d.O; investigation, G.G.d.S, G.d.S.C, K.C.P.M.V, P.W.P, C.M.P.d.S, B.C.R.C.M, R.J.O.M, A.d.S.O, A.S.G.T, M.C.d.G.M, C.d.O, M.J.J.G, R.P.P and E.M; writing—original draft preparation, G.G.d.S, G.d.S.C, K.C.P.M.V, P.W.P, C.M.P.d.S, B.C.R.C.M, R.J.O.M, A.d.S.O, A.S.G.T, M.C.d.G.M, C.d.O, M.J.J.G, R.P.P and E.M.; writing—review and editing, G.G.d.S, G.d.S.C, K.C.P.M.V, P.W.P, C.M.P.d.S, B.C.R.C.M, R.J.O.M, A.d.S.O, A.S.G.T, M.C.d.G.M, C.d.O, M.J.J.G, R.P.P and E.M.; visualization, G.G.d.S.; supervision, G.G.d.S; project administration, G.G.d.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

This study did not directly involve human beings or animals and, therefore, did not require approval from a research ethics committee.

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Public Involvement Statement

No public involvement in any aspect of this research.

Guidelines and Standards Statement

This manuscript was drafted against the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) [13].

Use of Artificial Intelligence

AI or AI-assisted tools were not used in drafting any aspect of this manuscript.

Acknowledgments

Research Group on Nursing Care in Women’s Health/Observatory of Technological Innovations in Women’s Health—Intersectionality, Race and Gender (GPCESM/OTIRG-Mulheres), Guarulhos University (UnG).

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

CNES National Registry of Health Establishments
DRS Regional Health Departments
HCNs Health Care Networks
IBGE Brazilian Institute of Geography and Statistics
ICD-10 International Statistical Classification of Diseases and Related Health Problems, 10th Revision
LD Linear Dichroism
MMCs Maternal Mortality Committees
MMR Maternal Mortality Ratio
OOBr Brazilian Obstetric Observatory
ReMaP Restructuring and Redesign of the Maternal and Perinatal Care Network
RRAS Regional Health Care Networks
SDGs Sustainable Development Goals
SIH/SUS Hospital Information System of the Brazilian Unified Health System
SIM Mortality Information System
SINASC Live Birth Information System
STROBE Strengthening the Reporting of Observational Studies in Epidemiology
SUS Brazilian Unified Health System
TLA Three-Letter Acronym
UNICEF United Nations Children’s Fund
WHO World Health Organization
WCA Women of Childbearing Age

References

  1. Neerland C, Slaughter-Acey J, Behrens K, Claussen AM, Usset T, Bilal-Roby S, Bashir H, Westby A, Wagner B, McAlpine D, Dixon M, Xiao M, Avila D, Butler M. An Evidence Map for Social and Structural Determinants for Maternal Morbidity and Mortality: A Systematic Review. Obstet Gynecol. 2024 Mar 1;143(3):383-392. [CrossRef]
  2. Hailu EM, Maddali SR, Snowden JM, Carmichael SL, Mujahid MS. Structural racism and adverse maternal health outcomes: A systematic review. Health Place. 2022 Nov;78:102923. [CrossRef]
  3. Silva AD, Guida JPS, Santos DS, Santiago SM, Surita FG. Racial disparities and maternal mortality in Brazil: findings from a national database. Rev Saude Publica. 2024 Jul 5;58:25. [CrossRef]
  4. Santos GGD, Njoku A, Mafetoni RR, Sanfelice CFO, Nicolau AIO, Parenti PW, et al. Maternal Mortality Among Black Women in Brazil: A Retrospective Cohort Study. Int J Environ Res Public Health. 2026 Jan 9;23(1):94. [CrossRef]
  5. Tenorio DS, de Matos Brasil AG, Nogueira BG, Rolim Lima NN, Araújo JEB, Rolim Neto ML. High maternal mortality rates in Brazil: Inequalities and the struggle for justice. Lancet Reg Health Am. 2022 Oct;14:100343. [CrossRef]
  6. Santana AT de, Couto TM, Lima KTR dos S, Oliveira PS de, Bomfim ANA, Almeida LCG, et al. Racismo obstétrico, um debate em construção no Brasil: percepções de mulheres negras sobre a violência obstétrica. Ciênc saúde coletiva. 2024;29(9):e09952023. [CrossRef]
  7. Belfort IKP, Kalckmann S, Batista LE. Assistência ao parto de mulheres negras em um hospital do interior do Maranhão, Brasil. Saude soc. 2016Jul;25(3):631–40. [CrossRef]
  8. Lessa MS de A, Nascimento ER, Coelho E de AC, Soares I de J, Rodrigues QP, Santos CA de ST, et al. Pré-natal da mulher brasileira: desigualdades raciais e suas implicações para o cuidado. Ciênc saúde coletiva. 2022Oct;27(10):3881–90. [CrossRef]
  9. Alves MTSS de B e, Chagas DC das, Santos AM dos, Simões VMF, Ayres BV da S, Santos GL dos, et al. Desigualdade racial nas boas práticas e intervenções obstétricas no parto e nascimento em maternidades da Rede Cegonha. Ciênc saúde coletiva. 2021Mar;26(3):837–46. [CrossRef]
  10. Felipe LL, Albuquerque PC, Lopes JF, Zicker F, Fonseca B de P. Desigualdades regionais no acesso ao parto hospitalar no Estado do Rio de Janeiro, Brasil: redes de deslocamento, distância e tempo (2010-2019). Cad Saúde Pública. 2024;40(5):e00064423. [CrossRef]
  11. Gryschek AL de FPL, Nichiata LYI, Fracolli LA, Oliveira MAF de, Pinho PH. Tecendo a rede de atenção à saúde da mulher em direção à construção da linha de cuidado da gestante e puérpera, no Colegiado de Gestão Regional do Alto Capivari – São Paulo. Saude soc. 2014Apr;23(2):689–700. [CrossRef]
  12. Viana AL d’Ávila, Bousquat A, Melo GA, Negri Filho AD, Medina MG. Regionalização e Redes de Saúde. Ciênc saúde coletiva. 2018Jun;23(6):1791–8. [CrossRef]
  13. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007 Oct 20;335(7624):806-8. [CrossRef]
  14. Rodrigues A, Lacerda L, Francisco RPV. Brazilian Obstetric Observatory. arXiv preprint arXiv:2105.06534. [CrossRef]
  15. Brasil. Ministério da Saúde. Conselho Nacional de Saúde. Resolução nº 466, de 12 de dezembro de 2012. Aprova diretrizes e normas regulamentadoras de pesquisas envolvendo seres humanos. Diário Oficial da União. 2013 jun 13; Seção 1:59.
  16. Peterson E, Chou D, Moller AB, Gemmill A, Say L, Alkema L. Estimating maternal mortality using data from national civil registration vital statistics systems. UN Maternal Mortality Estimation Inter-Agency Group. https://arxiv.org/abs/1909.08578.
  17. Leal MC, Gama SGN, Pereira APE, Pacheco VE, Carmo CN, Santos RV. The color of pain: racial iniquities in prenatal care and childbirth in Brazil. Cad Saude Publica. 2017;33 Suppl 1. [CrossRef]
  18. Say L, Chou D, Gemmill A, Tunçalp Ö, Moller AB, Daniels J, et al. Global causes of maternal death: a WHO systematic analysis. Lancet Glob Health. 2014;2(6). [CrossRef]
  19. World Health Organization, United Nations Children’s Fund, United Nations Population Fund, World Bank Group, United Nations Department of Economic and Social Affairs, Population Division. Trends in maternal mortality 2000 to 2023: estimates by WHO, UNICEF, UNFPA, World Bank Group and UNDESA/Population Division. Geneva: World Health Organization; 2025.
  20. Arshad N, Skjærven R, Klungsøyr K, Sørbye LM, Kvalvik LG, Morken NH. Pregnancy-associated maternal mortality within one year after childbirth: population-based cohort study. BJOG. 2025;132(3):365-74. [CrossRef]
  21. World Health Organization. Maternal and perinatal death surveillance and response: materials to support implementation. Geneva: World Health Organization; 2021.
  22. Smith H, Ameh C, Roos N, Mathai M, van den Broek N. Implementing maternal death surveillance and response: a review of lessons from country case studies. BMC Pregnancy Childbirth. 2017;17(1):233. [CrossRef]
Table 1. Distribution of maternal deaths according to race/color, education level, and marital status, São Paulo, 2000–2025 (n=6,781).
Table 1. Distribution of maternal deaths according to race/color, education level, and marital status, São Paulo, 2000–2025 (n=6,781).
Variable/Category n (%)
Race/Color
White 3,831 (56.5)
Brown 2.082 (30.7)
Black 640 (9.4)
Missing data 193 (2.8)
Yellow 31 (0.5)
Indigenous 4 (<0.1)*
Education
Missing data 4.116 (60.7)
High School 1.375 (20.3)
Middle School (Grades 6-9) 638 (9.4)
Completed Higher Education 357 (5.3)
Elementary School I 181 (2,7)
Incomplete Higher Education 100 (1.5)
No Schooling 14 (0.2)
Marital status
Single 3.205 (47.3)
Married 2.177 (32,1)
Common-law marriage 484 (7.1)
Single 236 (3.5)
Missing data 218 (3,2)
Legally Separated/Divorced 201 (3.0)
Married 182 (2,7)
Widow 59 (0.9)
Legally Separated/Divorced 16 (0.2)
Widower 3 (<0.1)*
Extracted from the Restructuring and Redesign of the Maternal and Neonatal Care Network (ReMaP) panel, integrated into the Brazilian Obstetric Observatory (OOBr), 2000–2025. *Percentages calculated in relation to the total number of maternal deaths that occurred during the period (n=6,781). Values ​​less than 0.1% were presented as “<0.1”.
Table 4. Proportional distribution of women’s sociodemographic characteristics, according to race/color, education level, and marital status, with respective 95% confidence intervals and statistical significance values.
Table 4. Proportional distribution of women’s sociodemographic characteristics, according to race/color, education level, and marital status, with respective 95% confidence intervals and statistical significance values.
Variable/Category Proportion (Average indicator) 95% CI p-value
Race/Color
White 0.565 (56.5%) 55.31%; 57.67% < 0.001
Brown 0.307 (30.7%) 29.62%; 31.81% < 0.001
Black 0.094 (9.4%) 8.77%; 10.16% < 0.001
Missing data 0.028 (2.8%) 2.48%; 3.27% < 0.001
Yellow 0.005 (0.5%) 0.32%; 0.65% < 0.001
Indigenous <0.001 (<0.1%) 0.02%; 0.15% < 0.001
Education
Missing data 0.607 (60.7%) 59.53%; 61.86% < 0.001
High School 0.203 (20.3%) 19.34%; 21.25% < 0.001
Middle School (Grades 6-9) 0.094 (9.4%) 8.74%; 10.13% < 0.001
Completed Higher Education 0.053 (5.3%) 4.76%; 5.82% < 0.001
Elementary School I 0.027 (2.7%) 2.31%; 3.08% < 0.001
Incomplete Higher Education 0.015 (1.5%) 1.21%; 1.79% < 0.001
No Schooling 0.002 (0.2%) 0.12%; 0.35% < 0.001
Marital status
Single 0.473 (47.3%) 46.08%; 48.45% < 0.001
Married 0.321 (32.1%) 31.00%; 33.23% < 0.001
Common-law marriage 0.071 (7.1%) 6.55%; 7.77% < 0.001
Single 0.035 (3.5%) 3.07%; 3.94% < 0.001
Missing data 0.032 (3.2%) 2.82%; 3.66% < 0.001
Separated/Divorced 0.030 (3.0%) 2.59%; 3.40% < 0.001
Married 0.027 (2.7%) 2.33%; 3.10% < 0.001
Widow 0.009 (0.9%) 0.68%; 1.12% < 0.001
Separated Jud./Divorced 0.002 (0.2%) 0.15%; 0.38% < 0.001
Widower <0.001 (<0.1%) 0.02%; 0.13% < 0.001
Table 5. Proportional distribution of obstetric and care characteristics of maternal deaths, according to type of death, period of death, ICD-10 chapter, medical care provided, and place of occurrence.
Table 5. Proportional distribution of obstetric and care characteristics of maternal deaths, according to type of death, period of death, ICD-10 chapter, medical care provided, and place of occurrence.
Variable/Category Proportion 95% CI p-value
Type of maternal death
Direct 0.612 (61.2%) 59.99%; 62.31% < 0.001
Indirect 0.360 (36.0%) 34.82%; 37.10% < 0.001
Not specified 0.029 (2.9%) 2.52%; 3.32% < 0.001
Period of death
During the postpartum period, up to 42 days 0.499 (49.9%) 48.70%; 51.08% < 0.001
During pregnancy, childbirth, or abortion. 0.269 (26.9%) 25.86%; 27.97% < 0.001
Not informed 0.105 (10.5%) 9.75%; 11.21% < 0.001
Inconsistent period 0.056 (5.6%) 5.08%; 6.18% < 0.001
During the postpartum period, from 43 days to less than 1 year. 0.054 (5.4%) 4.87%; 5.95% < 0.001
Not during pregnancy or the postpartum period 0.018 (1.8%) 1.48%; 2.11% < 0.001
Death during childbearing age
Yes 1,000 (100.0%) 99.94%; 100.00% < 1.000
ICD-10 Chapter
XV – Pregnancy, childbirth and the postpartum period 0.965 (96.5%) 96.09%; 96.96% < 0.001
I – Some infectious and parasitic diseases 0.033 (3.3%) 2.95%; 3.80% < 0.001
V – Mental and behavioral disorders <0.001 (<0.1%) 0.02%; 0.15% < 0.001
II – Neoplasms (tumors) <0.001 (<0.1%) 0.01%; 0.11% < 0.001
IV – Endocrine, nutritional and metabolic diseases <0.001 (<0.1%) 0.00%; 0.08% < 0.001
Medical assistance
With assistance 0.688 (68.8%) 67.73%; 69.93% < 0.001
Missing data 0.286 (28.6%) 27.56%; 29.71% < 0.001
Without assistance 0.025 (2.5%) 2.19%; 2.94% < 0.001
Location of occurrence
Hospital 0.932 (93.2%) 92.55%; 93.75% < 0.001
Another health facility 0.032 (3.2%) 2.85%; 3.69% < 0.001
Home 0.026 (2.6%) 2.26%; 3.02% < 0.001
Others 0.008 (0.8%) 0.60%; 1.02% < 0.001
Public road 0.002 (0.2%) 0.10%; 0.31% < 0.001
Missing data <0.001 (<0.1%) 0.00%; 0.08% < 0.001
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.