Performance of primary health care in São Paulo state, Brazil, during the period 2010-2019

This article presents the results of an analysis of the performance of primary health care in São Paulo state over the last decade against a backdrop of financial crisis and health funding cuts. We conducted a time series analysis (2010-2019) of performance indicators across the following dimensions based on an adapted conceptual framework: health service performance, health system, and determinants of health. Annual percentage change was calculated for each indicator using a log-linear model. Performance across the indicators was generally positive; however, there was a decline in performance across indicators of quality of care (congenital syphilis, cesarean section rate and cervical cancer screening). The findings also show a potential rise in demand for public services (due to population aging and a reduction in the percentage of the population with private health insurance) and increase in health expenditure against a backdrop of falling GDP per capita.


Introduction
Primary health care (PHC) can be defined as a point of care that offers a set of individual and collective actions coordinated via a health system aimed at delivering comprehensive care to the population 1 .In Brazil, significant advances in the organization of PHC services have been made since the implementation of the country's public health system, the Sistema Único de Saúde (SUS) or Unified Health System.The system adopted a new organizational model called primary care (PC), centered on territorialization, the affiliation of the population to health teams and the delivery of family and community-oriented care through the progressive implementation of the Family Health Strategy (FHS) 2 .
To implement this model at national level, three national primary care policies (PNAB, acronym in Portuguese) have been published since the creation of the SUS (2006, 2012 and 2017) [3][4][5] , emphasizing the centrality of the FHS and defining funding rules for the expansion of the strategy.In addition, the "More Doctors" law was created in 2013, aimed at expanding and improving the quality of medical training in the country and promoting a significant increase in the availability of doctors in primary care services 6 .
In 2017 there was a shift in policy, with PC losing its centrality and the introduction of changes to the funding model, undermining the capacity of the FHS to reorientate the health system toward PHC [5][6][7][8][9] .In addition, constitutional amendment 95/2016 -'the spending cap amendment' -which freezes health spending for 20 years, has adversely affected already inadequate funding of the SUS and consequently PC 10 .
PHC monitoring and evaluation is key to effective public health management [11][12][13][14] .Given the above changes to public policy and the political and economic context in the country over the last two decades, there is a pressing need to better understand the effectiveness of the implementation of PC in Brazil.
This study therefore sought to analyze the development of PHC actions over the last decade in São Paulo state (SPS) using a set of performance indicators.While SPS is the country's most populated and wealthy state 15 , it had the lowest FHS coverage rate in 2009, partly due to its program-based care model, characterized by the existence of a broad network of health centers prior to the implementation of the SUS 16 .In 2019, SPS had an estimated population of 45,919,049 inhabitants (22% of the national population) distributed across 645 municipalities, with 76% of the population living in 81 municipalities with over 100,000 inhabitants 15,17 .
The aim of this study was to examine temporal trends in PHC performance indicators in SPS during the period 2010-2019.

Materials and methods
This ecological time series study of PHC indicators is part of a research project titled "Participation of Social Organizations in the Management of Primary Health Care in Municipalities in São Paulo State" (FAPSPS-PPSUS, Nº 2019/03961-8).The study period is 2010-2019.

Indicators and data sources
The selected indicators make up a conceptual framework for monitoring PC adapted from criteria proposed by the Health System Performance Assessment Project (PROADESS) 18 .The framework consists of the following dimensions: health service performance, health system, and determinants of health.For each dimension we considered the following subdimensions: health service performance (access, effectiveness and adequacy of PHC); health system (funding); determinants of health (socioeconomic and demographic determinants).
The indicators were selected based on the data available in official open access databases linked to the SUS, adopting relevant validity, reliability and sensitivity criteria 19

Statistical analysis
Indicator trends were analyzed using a log-linear model, where the independent variable was the year and the dependent variables were the indicator rates.The annual percent change (APC) in rates was calculated using the Joinpoint Regression Program 4.8.0.0.Joinpoints in the trend curves were detected to identify statistically significant changes in the APC over the study period.
When joinpoints were detected, we calculated the APC corresponding to each segment of the curve.In addition, we calculated the average annual percent change (AAPC) over the study period.We also calculated 95% confidence intervals (95%CI).

Ethical considerations
The study protocol was approved by the Santa Casa de Misericórdia hospital's research ethics committee (reference code no.4.007.368).

Results
The annual rates for each of the 23 indicators are shown in Table 1.The following trends were observed in the dimension service performance: an increase in PHC coverage from 51.03% to 60.33%; a reduction in hospitalization rates (for asthma among children aged under 10; strokes among people aged 30-59, and ARI among children aged under 5); a slight reduction in the rate of infant mortality and its components; a reduction in babies born to teenage mothers (< 20 years) and with low birth weight; a reduction in the percentage of hospitalizations for ambulatory care sensitive conditions.Rates of congenital syphilis and syphilis detection in pregnant women increased.In addition, there was an increase in the cesarian section rate among SUS deliveries and a decrease in Pap testing rates in the 25-64 year age group.
In the dimension determinants of health, there was an increase in the proportion of older people in the overall population throughout the study period.
As regards the dimension health system, the findings reveal an increase in total health expenditure per capita between 2010 and 2014, and a reduction between 2015 and 2018 followed by a slight increase in 2019.
The results of the temporal trend analysis are presented in Table 2.The AAPC for PHC coverage was 1.9% (95%CI 1.4; 2.5), while estimated oral health team population coverage showed a constant reduction over the study period (APC = -0.4;95%CI: -0.8; -0.1) (service performance dimension).
Of the 11 indicators in the subdimension effectiveness, nine showed a downward trend during the study period (Table 2), including percentage of hospitalizations for ambulatory care sensitive conditions, which declined over the study period (APC = -1.4;95%CI: -1.5; -1.3).
In the subdimension adequacy, while the overall cesarian section rate in both public and private sector health services showed a statistically significant reduction during the period 2013-2019, the rate among deliveries performed on the SUS showed an upward trend over the study period (AAPC = 1.1%; 95%CI: 0.3; 1.8).Pap testing rates showed a constant reduction over the study period (APC = -3.3%;95%CI: -4.1; -2.5).There were no statistically significant trends in the screening mammography rate among women (Table 2).Joinpoints were detected in the indicators of the dimension determinants of health, revealing differing trends.The proportion of older people in the overall population showed an upward trend throughout the study period (AAPC = 2.8%; 95%CI: 2.8; 2.8).The percentage of the population with private health insurance showed a downward trend during the period 2013-2019, with an APC of -2.3% (95%CI: -3;3; -1.4).GDP per capita showed an upward trend up to 2014 (APC = 2.1%; 95% CI: 0.9; 3.,3), followed by a downward trend thereafter up to the end of the study period (APC = -3,8%; 95%CI: -4.9; -2.7).
In the dimension health system, joinpoints were detected in the indicators total health expenditure per capita and expenditure on health as a percentage of total government expenditure, with an initial statistically significant upward trend in both indicators followed by a period of stabilization in total health expenditure and decline in expenditure on health as a percentage of total expenditure, from 2014 and 2016, respectively.
Chart 2 presents a synthesis of the results by dimension, highlighting trends over the entire study period and joinpoints.

Discussion
Our findings clearly show that, despite an improvement in access to PHC in SPS during the study period, performance across care quality indicators was poor, revealing that SPS faces major challenges in improving the quality of PC.It was observed that most of the performance indicators in the access and effectiveness subdimensions showed a favorable trend, with only the congenital syphilis incidence rate showing a downward trend.Performance across indicators of adequacy was poor or stable over the study period, indicating that while access to care in the state has improved, care quality remains a challenge.
With regard to access, PHC coverage increased up to 2016.This may be related to the introduction of the More Doctors Program 20 , which allocated 14,256 professionals to family health teams across the country during the first two years of the initiative 6 .The stabilization of coverage rates in the following period may be due to several factors, including the effects of financial constraints on health actions, the reorientation of the PNAB and a reduction in the number of PHC doctors, especially after 2018 10,21 .There was a downward trend in estimated oral health team population coverage over the study period.In this respect, a study using national data reported an increase in oral health team population coverage between 2008 and 2015, followed by a slight reduction in 2016 and stabilization thereafter.In the same study, the authors found an upward trend in the transfer of resources for oral health up to 2010, followed by stabilization 22 .Another study investigating oral health care coverage across Brazil's regions showed that expansion of coverage during the period 2001-2013 was lowest in the Southeast 23 .
All indicators in the effectiveness subdimension except for congenital syphilis showed favorable APC, with the highest reductions being found in rates of hospitalization for asthma, stroke and ARI.Hospitalization rates for these conditions, which are on Brazilian List of Hospitalizations due to Ambulatory Care Sensitive Conditions 24 , have been shown to decrease with improving PHC quality.The percentage of hospitalizations for ambulatory care sensitive conditions decreased throughout the study period.This may be explained by increased coverage of PHC and the FHS.In this respect, ecological studies have reported an association between increased FHS coverage and a decrease in hospi-talizations and mortality due to ambulatory care sensitive conditions [25][26][27][28] .
Although the considerable increase in syphilis detection in pregnant women indicates ad-equate screening among pregnant women, the upward trend in the rate of congenital syphilis, with an APC of 25% between 2010 and 2015, is worrying 29 , indicating persistent flaws in antenatal care and treatment of the disease during pregnancy.A study investigating trends in mean congenital syphilis rates in Brazil between 2010 and 2015 observed rising rates over the period and a correlation between congenital syphilis and screening for maternal syphilis and infant death, miscarriage and stillbirth rates, suggesting gaps in basic health care for pregnant women 30 .
The indicators of adequacy of care reveal little progress and that some backward steps were taken.The high cesarean section rate in the SUS and deliveries performed in the private sector reveal a care model that favors the interests of professionals and parturient women, often resulting in unnecessary interventions during labor and child birth 31 .Despite the measures proposed by the Rede Cegonha (Stork Network) 32 to improve labor and childbirth care on the SUS, the results obtained were unsatisfactory, suggesting gaps in the quality of PHC services.
Also regarding indicators of adequacy, the percentage of newborns whose mothers made at least seven antenatal visits oscillated between 75% and 80%, showing a slight increase during the period 2013-2019.However, this result was not statistically significant throughout the whole period.In a nationwide study conducted with PHC patients in 2012 and 2013, 89% of respondents reported having made at least six antenatal visits, with only 69% having access to all recommended tests, 60% receiving all recommended guidance, and 24% undertaking all physical examinations, thus raising questions about the quality of antenatal care 33 .A literature review of articles on antenatal care in Brazil reported an increase in coverage between 2005 and 2015, with variability in care quality, including not only the number of antenatal visits but also routine tests, health guidance and basic technical procedures 34 .
Cervical and breast cancer screening are key prevention strategies aimed at reducing mortality due to these diseases.The related indicators in this study suggest low screening coverage, with a downward trend for cervical screening and stationary breast cancer screening rates.However, studies reveal a reduction in mortality from breast and cervical cancer in state capitals in the Southeast, due mainly to early diagnosis and treatment.This reduction was more pronounced for cervical cancer 35,36 .A study evaluating cervical and breast cancer screening in Brazil using data from the Surveillance of Risk and Protection Factors for Chronic Diseases by Telephone Survey (VIGITEL) for the period 2007-2017 found an upward trend in breast cancer screening rates and stabilization in cervical screening rates.Coverage of cervical and breast cancer screening in SPS was above 90% and 80%, respectively, in 2018 37 .It is important to note that the VIGITEL methodology consisted of phone interviews conducted with people living in state capitals.The difference in indicator values between the VIGITEL study and our study may be partially due to variations in calculation methodology -with the present study considering only screening performed on the SUS -and bias arising from phone interviews.
Our findings suggest the need to step up cervical cancer screening efforts in PHC services to increase the coverage of testing, which is generally performed in health centers.With regard to breast cancer, efforts need to be intensified to increase the referral of women to services that perform mammograms.In this respect, PHC services need to develop strategies to promote mammography screening among target women affiliated to family health teams and other PHC teams.
In the dimension determinants of health, the selected indicators reveal a trend towards population aging and consequently growing demand for related services and actions, such as the treatment of neoplastic and circulatory system diseases and neurological complications resulting from strokes.The so-called "demographic transition" and population aging give rise to changes in the morbidity and mortality profile 38 that need to be recognized to adapt public health policy to address the challenges of population aging 39 .
In addition to population aging, declining GDP per capita and the reduction in the percentage of the population with private health insurance put increasing pressure on the SUS, and consequently PHC services.
It is important to highlight that there was an initial upward trend in both health expenditure per capita and expenditure on health as a percentage of total government expenditure.However, the joinpoints observed reveal a decline in funding in recent years against a backdrop of falling GDP per capita and increased demand for public health services, pointing to a potentially unfavorable scenario for the SUS and PHC.This is probably due to the influence of changes made during the same period, including the reorientation of primary care policy (review of the PNAB in 2017) and recent modifications to fund-ing rules, especially Constitutional Amendment 95/2016.
In the dimension health system, the funding indicators were certainly impacted by the financial crisis that emerged in 2015.This inflection in health funding adds to the strain on the SUS and PHC services 40 .
One of the limitations of this study is the fact that the data were not disaggregated by region or municipal characteristics, meaning that it was not possible to capture more pronounced trends in PHC performance in more homogeneous groups of municipalities.

Conclusion
In the last decade, national primary health care policy [4][5][6] has expanded coverage and improved access to health services.However, recent years have seen a reorientation of this policy 8,10 .In SPS, the expansion of PHC coverage up to 2016, followed by a period of stabilization reflect this shift in policy.
The indicators in the dimensions health system and determinants of health reveal important changes to the structure of the health system and health demands in the second half of the decade.The decline in health funding and economic slowdown (fall in GDP per capita) have potentially weakened the response of PHC to the challenges posed by population aging and the growing proportion of the population who depend exclusively on SUS.
Finally, the negative or stationary trends in indicators of adequacy of PHC and motherto-child transmission of syphilis are worrying, pointing to persistent gaps in care quality.These gaps need to be addressed by the state's public health managers.

Collaborations
A Sala, CG Luppi, GA Wagner e N Carneiro Junior: substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; and drafting the work or reviewing it critically for important intellectual content; and final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.RVB Pinheiro Junior: drafting the work or reviewing it critically for important intellectual content; and final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.All authors: agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Chart 1 .
Indicators and respective numerators, denominators, calculation methods and data sources.The indicators were obtained from the following health information systems: the Mortality Information System (SIM); the Live Birth Information System (SINASC); the National Health Facility Registry (CNES); the National Notifiable Diseases Information System (SINAN); the National Immunization Program Information System (SI-PNI); and the Public Health Budget Information System (SIOPS).They were extracted from the following platforms: Primary Care Information and Management (e-Gestor AB); São Paulo Department of Health (SES-SP); the SUS's Department of Informatics (DATASUS); and the State Data Foundation (SEADE).Annual GDP per capita and health expenditure per capita were inflation adjusted based the national consumer price index (IPCA) using the IPCA calculator.The databases used in this study are online and open access and provide annual indicator rates.
Indicators and respective numerators, denominators, calculation methods and data sources.
Sources: Primary Care Information and Management (e-Gestor AB); Mortality Information System (SIM); Hospital Information System (SIH); Newborn Information System (SINASC); National Notifiable Diseases Information System (SINAN); Ambulatory Care Information System (SIA); State Data Foundation (SEADE); Public Health Budget Information System (SIOPS); National Supplementary Health Agency (ANS); National Health Facility Registry (CNES); e Brazilian Institute of Geography and Statistics (IBGE).

Table 2 .
Annual percentage change (APC) for each indicator and average annual percent change (AAPC) for indicators where joinpoints were detected on the trend curve and respective 95% confidence intervals (95%CI).

Table 2 .
Annual percentage change (APC) for each indicator and average annual percent change (AAPC) for indicators where joinpoints were detected on the trend curve and respective 95% confidence intervals (95%CI).Synthesis of the results by dimension and subdimension, showing mean trends during the period, presence of joinpoints, and trend in each segment from the joinpoint.