Artigos

Introduction
Pregnancy is a complex event during which several factors contribute to changes in maternal and fetal outcomes. Antenatal care provides the opportunity for risk assessment and provision of support intending to improve perinatal results(1). Several measures may be implemented during prenatal visits, and there is evidence of reduced antenatal care visits associated with increased perinatal mortality(2).
The success of any clinical intervention requires access to health services. Barriers are numerous and depend on cultural, social, demographic, and economic contexts(3). Prenatal care access has been demonstrated to be responsive to public health measures aiming to increase adequate service usage, mainly when focusing on socially vulnerable groups(4,5).
Antenatal care usage is mainly measured through two variables: the month of initiation and number of visits over the pregnancy. Historically, the Kessner Index and the Kotelchuck Index have been the central classification systems to rank prenatal care usage adequacy(6). Both systems rely on the number of visits proposed by the American College of Obstetric and Gynecology (ACOG), thus eventually being modified when used outside the United States(7,8).
Inadequate routine healthcare service usage during pregnancy is not evenly distributed in the population. Feijen-de Jong et al. have identified usage determinants based on Andersen's behavioral model through a systematic literature review. Variables may be grouped into health behavior and predisposing, enabling, and need variables. Age, education, insurance status, parity, and availability of primary care physicians, among other determinants, were associated with adequacy of antenatal care usage(9).
In Brazil, a universal public healthcare system, the Unified Healthcare System (SUS), co-exists with private insurance. Primary care is offered mainly through the Family Health Strategy program, which has been increasing coverage over the years, reaching more than 131 million individuals in 2019, corresponding to over 62% of the country's population. These rates are higher among individuals with lower economic conditions and those with fewer years of education(10).
This research aims to assess the distribution of primary care funding among live birth patients in the State of Sao Paulo, Brazil, between 2015 and 2019. Furthermore, it aims to identify if there is a correlation between the type and level of funding of public primary healthcare and the adequacy of prenatal care usage. Additionally, this study intends to assess the potential impact of maternal and sociodemographic variables on the risk of inadequate antenatal care usage.

Material and Methods
This study is a retrospective population-based cohort study. Individual-level data was obtained for all live births through the Live Birth Information System (SINASC). At the same time, the independent variable of primary care public spending was ecologically attributed through information obtained from the Brazilian Office of the Comptroller General (CGU), specifically from the Transparency Portal website.
Anonymized patient data was obtained from the SINASC, which encompasses all information in all live birth certificates in the country. For each individual, data will be obtained regarding date of birth, city of residence, the Kotelchuck Index as provided in the birth certificate, the number of prenatal visits, the month antenatal care visits started, gestational age, pregnancy length in days, maternal age, ethnicity, partnership status, educational level, total number of previous deliveries, and number of fetuses.
Regarding the Kotelchuck Index, it is essential to mention Brazil's Ministry of Health utilizes a modified version, with the following criteria: absent if there are no prenatal visits; inadequate when prenatal care either starts after the first trimester or has less than three prenatal visits; intermediate when prenatal care starts during the first trimester, and there are 3 to 5 visits; adequate if prenatal care started during the first trimester and there are precisely six visits, and adequate plus if prenatal care started during the first trimester and there are seven or more visits(11).
Maternal age was converted to the maternal age group, defined as less than 20 years, from 20 to 35 years, and more than 35 years. Education level was categorized as less than eight years of study, from 8 to 11 years, and 12 or more years of education. Ethnicity was defined as used in official statistics in Brazil: white, mixed-race, black, Asian, or indigenous. Partnership status was defined as the presence of a partner if married or in a stable union, or absent if single, divorced, or widowed. The Kotelchuck Index was transformed to include all patients with adequate or adequate plus prenatal care utilization as "adequate" and all patients with absent, inadequate, or intermediate utilization as "inadequate."
All patients residing and having birth in the State of Sao Paulo from 2015 to 2019 were included in the study. Reasons for patient exclusion were as follows: missing or unknown city of residence; delivery occurred before 20 weeks; non-singleton pregnancy; absence of measurable outcome; the midpoint of pregnancy occurred before 2015.
Public primary healthcare funding data is available monthly through the Transparency Portal website. Transfers from the federal government to cities were included for all 645 cities in the State of Sao Paulo if categorized as Fixed Primary Healthcare Floor (FPHF) or Variable Primary Healthcare Floor (VPHF). Due to variations in official nomenclature throughout the years, FPHF included all spending classified as fixed primary care or funding for basic healthcare units; VPHF included variable primary care spending, temporary revenue to achieve primary healthcare goals, and spending directed to the Program of Improvement in Access and Quality (PMAQ), which is officially recognized as a portion of the VPHF. Population estimates per city were obtained as made available on the Ministry of Health website, known as DATASUS. This allowed the calculation of each financial transfer per city, year, and capita. Values for 2019 will be used nominally, and each previous year's values will be adjusted by the Extended National Consumer Price Index (IPCA), one of the country's primary monetary inflation indexes. IPCA adjustment is provided by the Brazilian Institute of Geography and Statistics (IBGE) website.
Once values were obtained for each of Sao Paulo's 645 cities during the five years of the study, the total 3225 combinations of town and year were classified in quartiles to determine the level of funding: low, medium-low, medium-high, and high. This procedure was carried out independently for FPHF and VPHF.
The expenditure variable was allocated ecologically. Since the very nature of primary public healthcare funding in Brazil is not individually directed, and potential policies can target a city’s overall funding, we consider the ecological approach valid for this study. The year of the midpoint of the pregnancy will be determined using each pregnancy date of delivery and pregnancy length. Each patient has a public primary healthcare funding level for the abovementioned categories utilizing the city of residence and year.
Descriptive statistics were used to display the overall cohort results. We used univariate and multivariate Poisson regressions with robust errors, as described by Zou(12), to assess the relative risk (RR) of inadequate prenatal care utilization. Variables used in the models included FPHF and VPHF funding level, maternal age group, partnership status, ethnicity, educational level group, and nulliparity.
Data manipulation and statistical analysis were conducted using the software RStudio version 4.2.2 (2022-10-31).

Results
The initial data included 3,033,193 live births from patients with residence in the State of Sao Paulo from 2015 to 2019. From this point, 39 cases were excluded due to unspecific code for city of residence; 9,280 were removed, as gestational age upon delivery was less than 20 weeks; 75,805 were excluded due to non-singleton pregnancy; 68,584 live births were removed, as they did not have a Kotelchuck Index score; and 231,992 were excluded due to midpoint of pregnancy occurring before 2015. The final cohort encompassed 2,647,493 live births spanning 23,031,252 prenatal visitations or an average of 4,606,250 visitations per year.
Inadequate prenatal care usage was identified in 535,447 (20.2%) live births in the cohort. Among these, 215,624 (40.27%) were classified as inadequate care usage solely on beginning prenatal care after the third trimester; 161,546 (30.17%) were classified as inadequate exclusively due to lacking enough prenatal care visits; 142,124 (26.54%) individuals had both lack of enough visits and the start of care after the first three months of pregnancy; 15,689 (2.93%) had no prenatal care; and 464 (0.09%) lacked some degree of information regarding one of the two variables, but were still able to be classified as inadequate prenatal care usage. Considering the goal of 6 antenatal care visits in Brazil, 263,808 individuals (49.27%) would have been classified as having adequate antenatal care usage if they had received one additional visitation in the first trimester.
A higher proportion of inadequate prenatal care users were concentrated in cities with low primary care funding levels; this phenomenon was observed in both fixed and variable funding. Inadequate prenatal care users were more often younger, had fewer years of education, had an absence of a partner, were of non-white ethnicity, and were non-nulliparous. These findings are summarized in Table 1.
When grouping patients according to their city of residence's funding level, the higher funding strata overrepresent in both financing models patients younger than 20 years, with fewer than eight years of education, with a partner present, and of white ethnicity. Notably, patients without a partner of black and mixed-race ethnicity compose 41.3%, 6.0%, and 36.5% of the cohort, respectively. Yet, they are underrepresented in all funding levels except for the low FPHF and medium-low VPHF strata. These results are summarized in Table 2 for the fixed financing model and Table 3 for the variable funding strategy.
All studied variables demonstrated an association with inadequate prenatal care usage in the generalized linear models, except for Asian ethnicity in the multivariate model. Increases in FPHF were associated with a decrease in RR: both medium-low and medium-high FPHF levels had a RR of 0.93 (95% CI: 0.92-0.93 and 0.92-0.94, respectively, p < 0.001), while RR among high FPHF level residents increased to 0.98 (95% CI: 0.97-1.00, p = 0.01) in the multivariate model. Increases in VPHF level, on the other hand, displayed constant RR reduction in the multivariate model, going from 0.94 (95% CI: 0.94-0.95, p < 0.001) in medium-low level to 0.91 (95% CI: 0.90-0.91, p < 0.001) in medium-high levels and 0.88 (95% CI: 0.86-0.90, p < 0.001) in high VPHF level. These findings, as well as RR for age, education, partnership, ethnicity, and nulliparity groups, can be found in detail in Table 4.
We performed an exploratory analysis of the average monetary value of the FPHF and VPHF per capita transfers in each funding level over the years. Values are expressed in the Brazilian currency, the Real. In 2015, the mean value of low FPHF was 22.76 reais, while the mean high FPHF value was 29.08 reais, 27.77% higher than the lowest funding level. As for VPHF, the high VPHF mean values in 2015 were 108.73 reais, while the lowest VPHF level had an average of 6.36 reais per capita; the highest VPHF strata had more than 17 times more average transfers per capita than the lowest VPHF level. The high VPHF values are also 3.74 times higher than the transfers dedicated to high FPHF spending. These relationships remained relatively constant over the years, as can be seen in Figure 1.

Discussion
Main findings
Individuals living in cities with higher primary healthcare public spending had a lower relative risk of inadequate prenatal care usage. FPHF had a lessened risk reduction in the highest fixed funding level. VPHF spending, on the other hand, displayed a more homogeneous result, as risk decreased gradually over each funding level increment. Variable funding presented a more unequal distribution.
The risk of inadequate prenatal care usage decreased when patients displayed more years of education, were older, nulliparous, and were either married or living in a stable union. Non-white ethnicity expressed an overall increasing impact over risk, except for Asians. The main limiting factor for adequate usage was the late beginning of antenatal visits.
Patients without a partner and of black and mixed-race ethnicity care were underrepresented in medium-high and high funding levels. In the fixed model, which is guided by social vulnerability criteria, these disadvantaged minorities were underrepresented in all groups but the lowest funding level.
Interpretation
The FPHF and VPHF represent two methods of public health financing: fixed financing for expected demand, based on a priori information, and variable funding, associated with criteria such as adherence to specific federal programs or achieving particular healthcare goals.
Brazil's Ministry of Health has specific fixed and variable primary healthcare spending criteria. The FPHF of a town is calculated based on five criteria: Gross Domestic Product per capita, percentage of the population with private healthcare insurance, percentage of the population receiving government financial aid, percentage of the people living in extreme poverty, and demographic density(13). As for VPHF, it considers the implementation of specific programs required by the Ministry of Health, such as family medicine programs(14), as well as achieving goals regarding prenatal care, oral healthcare, and chronic disease control, among others(15). While the Brazilian federal government proposed changes to this model at the end of 2019(16), taking effect in 2020, we consider that the available data should be analyzed to generate evidence to further the scientific knowledge regarding the issue.
FPHF displayed a non-linear risk reduction, as the highest strata displayed less impactful result. These are the cities with highest degree of social vulnerability. It is possible that, as multiple demands accumulate, pregnant patient’s needs might not be prioritized. Thus, this model leads to the underfunding of socially vulnerable groups(17).
Variable funding had a more homogeneous impact, with each funding level having better results than the previous. This could be explained by antenatal care usage being one of the predefined goals. VPHF, however, is more unequally distributed and more costly than FPHF. To some degree, these findings could be secondary to mean higher monetary values of VPHF. The unequal financial distribution of VPHF raises concerns regarding the self-sustaining cycle of disease and poverty, with the addition of political representation and funding obtention to this equation.
Our results demonstrate that public policies that are supposedly neutral regarding ethnicity or family structure can systematically underfinance already disadvantaged groups. While having an independent higher risk of inadequate antenatal care, patients without a partner, black, and of mixed-race ethnicity had reduced participation in higher funding levels. This raises the hypothesis regarding the potential role of an allegedly impartial public healthcare system in generating and perpetuating differences, contributing to processes of racialization(18). Furthermore, it should be noted the impact on black and mixed-race patients demonstrated by our findings is likely underestimated: the portion of patients residing in high-funding municipalities are likely to face implicit racial bias in healthcare units, impacting the patients' accessibility(19). Similar difficulties in access might be expected from patients without a partner(20,21).
The shortcomings of both models point towards the overarching problem of women's healthcare underfunding. Healthcare research and public policy have been marked by gender inequality(22,23). Inadequate prenatal care is responsive to higher funding, especially if governmental mechanisms bind financing to the specific initiative. Considering the magnitude of antenatal care, which accounted for over 4.6 million visitations per year in this study, a dedicated line of primary care funding should be considered. This would allow the opportunity to revise the current social vulnerability criteria, which are ineffective for the specific population of pregnant patients.
Previous ecological research about the aforementioned PMAQ program in Brazil, which is a part of the variable primary care funding, did not find significant improvement in the number of antenatal care visits or overall mortality(24). This divergence could arise from our research considering all components of the VPHF instead of specifically the PMAQ. Also, the ecological research did not consider the first-trimester timing criteria, which our results indicate account for the most significant proportion of inadequate care.
A study in Rwanda compared input-based funded health facilities with payment-for-performance facilities in a randomized trial(25). Among users of different facilities, there was no increased probability of having any prenatal care or four or more antenatal care visits. This finding diverges from our results. The authors consider payment-for-performance might be more effective in interventions less dependent on patient behavior. The difference between this research and ours might be linked to differences in health systems, the co-existence of fixed and variable funding in Brazil, and differences in the barriers to access in different sociocultural and economic contexts.
An American study found an association between a local primary care shortage and late prenatal care initiation; while not directly considering the role of funding, it demonstrates the relevance of primary care in antenatal care accessibility and usage(26).
When reviewing the impact of reward-based financial incentives associated with improving primary care, Scott et al. considered existing evidence insufficient to recommend this model(27). It should be noted the definition of quality in that review was heterogeneous, and access was not always considered a part of quality; it is possible our results may not be generalizable to the actual quality of the provided service. Furthermore, that study did not include antenatal care in the primary healthcare setting.
Our findings are fundamentally convergent with previous research regarding sociodemographic variables. Previous studies have demonstrated individuals who are younger, have fewer years of education, do not have a partner, and belong to ethical minorities have a higher risk of inadequate prenatal care utilization(1,9,28,29). These findings reinforce the notion that barriers to access beyond the healthcare facilities and government policy might require a focus on healthcare-seeking behavior.
Strengths and limitations
This study has several strengths. It is based on birth certificates, which are mandatory documents. It encompasses a large population over several years. It also considers public primary healthcare spending levels, a variable previously was not studied in this context, to the best of our knowledge. Limitations arise from the retrospective nature of the study, the use of administrative datasets, the lack of information regarding other cofounders, such as the presence of comorbidities and if the pregnancy was unwanted, and changes in official nomenclature for some types of funding over the years. These limitations, however, do not seem to invalidate the results.

Conclusion
Higher levels of public primary healthcare spending were associated with a reduced risk of inadequate usage of prenatal care services. Broad social vulnerability criteria are inadequate for the pregnant patient population, which might need a specific financing model and criteria. Addressing the late beginning of prenatal care should be the focus of future healthcare policies.
Declarations
Ethics
The study followed the Declaration of Helsinki. According to Resolution 510/2016 of the National Health Council of Brazil, research that uses publicly available data without patient identification does not require approval from an Ethics Committee.

Transparency
We declare this manuscript is an honest, accurate, and transparent account of the study conducted and that no important aspects of this study have been omitted.

Funding
There are no sources of funding to disclose.
Disclosures and Funding Sources
The authors report no conflict of interest regarding financial issues relevant to the research reported in this manuscript.

Data availability statement
The data used in this research is made publicly available by Brazil's Ministry of Health and can be downloaded from https://datasus.saude.gov.br/transferencia-de-arquivos/
Data from the Brazilian Office of the Comptroller-General can be downloaded from https://portaldatransparencia.gov.br/download-de-dados/transferencias
The inflation calculator from the Brazilian Institute of Statistics and Geography can be found at https://www.ibge.gov.br/explica/inflacao.php
Original data files, the cohort file, and scripts used to analyze data are available at https://drive.google.com/drive/folders/1UAzuWxzyT9kJk264e8QKSlm1X-G3cbNf?usp=sharing

Acknowledgments
We want to thank all members of Brazil's Ministry of Health, the Brazilian Office of the Comptroller-General, and the Brazilian Institute of Statistics and Geography for making such data publicly available.

Competing interests
The authors report no conflict of interest relevant to the research reported in this manuscript.

References
1. Gadson A, Akpovi E, Mehta PK. Exploring the social determinants of racial/ethnic disparities in prenatal care utilization and maternal outcome. Semin Perinatol. 2017 Aug 1;41(5):308–17.
2. Ota E, da Silva Lopes K, Middleton P, Flenady V, Wariki WM, Rahman MdO, et al. Antenatal interventions for preventing stillbirth, fetal loss and perinatal death: an overview of Cochrane systematic reviews. Cochrane Database Syst Rev. 2020 Dec 18;2020(12):CD009599.
3. Thaddeus S, Maine D. Too far to walk: Maternal mortality in context. Soc Sci Med. 1994 Apr;38(8):1091–110.
4. Eliason EL, Daw JR. Presumptive eligibility for pregnancy Medicaid and timely prenatal care access. Health Serv Res. 2022;57(6):1288–94.
5. Shah JS, Revere FL, Toy EC. Improving Rates of Early Entry Prenatal Care in an Underserved Population. Matern Child Health J. 2018 Dec 1;22(12):1738–42.
6. Kotelchuck M. An evaluation of the Kessner Adequacy of Prenatal Care Index and a proposed Adequacy of Prenatal Care Utilization Index. Am J Public Health. 1994 Sep;84(9):1414–20.
7. Santos Neto ETD, Oliveira AE, Zandonade E, Leal MDC. Access to prenatal care: assessment of the adequacy of different indices. Cad Saúde Pública. 2013 Aug;29(8):1664–74.
8. Jarvis C, Munoz M, Graves L, Stephenson R, D’Souza V, Jimenez V. Retrospective Review of Prenatal Care and Perinatal Outcomes in a Group of Uninsured Pregnant Women. J Obstet Gynaecol Can. 2011 Mar 1;33(3):235–43.
9. Feijen-de Jong EI, Jansen DE, Baarveld F, Van Der Schans CP, Schellevis FG, Reijneveld SA. Determinants of late and/or inadequate use of prenatal healthcare in high-income countries: a systematic review. Eur J Public Health. 2012 Dec 1;22(6):904–13.
10. Giovanella L, Bousquat A, Schenkman S, Almeida PFD, Sardinha LMV, Vieira MLFP. Cobertura da Estratégia Saúde da Família no Brasil: o que nos mostram as Pesquisas Nacionais de Saúde 2013 e 2019. Ciênc Saúde Coletiva. 2021 Jun;26(suppl 1):2543–56.
11. BRASIL M da S. Saúde Brasil 2017: uma análise da situação de saúde e os desafios para o alcance dos Objetivos de Desenvolvimento Sustentável. [Internet]. Brasilia; 2018 [cited 2023 Aug 8]. Available from: http://tabnet.datasus.gov.br/cgi/sinasc/saude_brasil_2017_analise_situacao_saude_desafios_objetivos_desenvolvimento_sustetantavel.pdf
12. Zou G. A Modified Poisson Regression Approach to Prospective Studies with Binary Data. Am J Epidemiol. 2004 Apr 1;159(7):702–6.
13. Mendes Á, Marques RM. O financiamento da Atenção Básica e da Estratégia Saúde da Família no Sistema Único de Saúde. Saúde Em Debate. 2014 Dec;38:900–16.
14. Ferreira MPS, Rosa MRR, Coelho TCB, Teles AS. FINANCIAMENTO DA ATENÇÃO BÁSICA EM UMA MACRORREGIÃO DA BAHIA. Rev Saúde Coletiva UEFS. 2015;5(1):27–32.
15. Silva P de S e, Souza EC de. Análise dos indicadores de desempenho de uma equipe de atenção básica à saúde. Rev Saúdecom. 2016 May 17;12(1):470–6.
16. Melo EA, Almeida PF de, Lima LD de, Giovanella L. Reflexões sobre as mudanças no modelo de financiamento federal da Atenção Básica à Saúde no Brasil. Saúde Em Debate. 2020 Jun 19;43:137–44.
17. Ashworth M, L’Esperance V, Round T. Primary care funding entrenches health inequalities: time for a rethink. Br J Gen Pract. 2021 Feb 26;71(704):102–4.
18. Barot R, Bird J. Racialization: the genealogy and critique of a concept. Ethn Racial Stud. 2001 Jan 1;24(4):601–18.
19. Hamed S, Bradby H, Ahlberg BM, Thapar-Björkert S. Racism in healthcare: a scoping review. BMC Public Health. 2022 May 16;22(1):988.
20. Amroussia N, Hernandez A, Vives-Cases C, Goicolea I. ‘Is the doctor God to punish me?!’ An intersectional examination of disrespectful and abusive care during childbirth against single mothers in Tunisia. Reprod Health. 2017 Mar 4;14(1):32.
21. Kim JE, Lee JY, Lee SH. Single Mothers’ Experiences with Pregnancy and Child Rearing in Korea: Discrepancy between Social Services/Policies and Single Mothers’ Needs. Int J Environ Res Public Health. 2018 May 10;15(5):955.
22. Schubert KG, Bird CE, Kozhimmanil K, Wood SF. To Address Women’s Health Inequity, It Must First Be Measured. Health Equity. 2022 Dec 2;6(1):881–6.
23. de Oliveira ALM, Alloatti MN. Gendering the crisis: austerity and the Covid-19 pandemic in Brazil. Econ Polit Bologna Italy. 2022;39(1):203–24.
24. Kovacs R, Barreto J, Nunes Da Silva E, Xander Russo L, Hone T, Powell-Jackson T. Paying Our Way to Better Health? The Impact of Public Spending on Health Outcomes in Brazil [Internet]. SSRN; 2023 [cited 2023 Aug 10]. Available from: https://www.ssrn.com/abstract=4427528
25. Basinga P, Gertler PJ, Binagwaho A, Soucat AL, Sturdy J, Vermeersch CM. Effect on maternal and child health services in Rwanda of payment to primary health-care providers for performance: an impact evaluation. The Lancet. 2011 Apr;377(9775):1421–8.
26. Perloff JD, Jaffee KD. Late entry into prenatal care: the neighborhood context. Soc Work. 1999 Mar;44(2):116–28.
27. Scott A, Sivey P, Ait Ouakrim D, Willenberg L, Naccarella L, Furler J, et al. The effect of financial incentives on the quality of health care provided by primary care physicians. Cochrane Effective Practice and Organisation of Care Group, editor. Cochrane Database Syst Rev [Internet]. 2011 Sep 7 [cited 2023 Aug 10]; Available from: https://doi.wiley.com/10.1002/14651858.CD008451.pub2
28. Uwimana G, Elhoumed M, Gebremedhin MA, Nan L, Zeng L. Determinants of timing, adequacy and quality of antenatal care in Rwanda: a cross-sectional study using demographic and health surveys data. BMC Health Serv Res. 2023 Mar 6;23(1):217.
29. Feijen-de Jong EI, Jansen DEMC, Baarveld F, Boerleider AW, Spelten E, Schellevis F, et al. Determinants of prenatal health care utilisation by low-risk women: A prospective cohort study. Women Birth. 2015 Jun 1;28(2):87–94.