Building on the two previous reports published over the last decade, Brazil in now committed to fully institutionalise its national health accounts. This chapter discusses this new initiative as described in the new Brazilian health accounts manual and assesses whether the proposed implementation is line with international standards. The chapter reviews the general approach taken by Brazil to construct health accounts, including the comprehensiveness of the identified financing schemes. It discusses the availability of data sources in Brazil and their appropriate use in health accounts and provides an assessment of the methodology applied to estimate health spending in the new Brazilian health accounts.
Institutionalising Health Accounts in Brazil
A Review of Methods, Data and Policy Relevance
Abstract
As discussed in Chapter 2, after two rounds of pilot implementation of health accounts in Brazil over the last decade, a renewed initiative to comprehensively institutionalise the production of health accounts in line with international standards and based on SHA 2011 commenced in 2023. In this chapter, the general approach to estimate health spending data, the data sources used, and the estimation techniques applied under this new initiative as described in the methodological documentation of the Brazilian health accounts team (the “manual”) will be discussed.
The production of health accounts in line with international standards, that is, based on SHA 2011 definitions, classification, categories and recommended methodology, has by now been institutionalised in most OECD countries (OECD, 2025[1]). Based on the SHA 2011 manual (OECD/WHO/Eurostat, 2011[2]), key methodological features and high-level guidance that countries should consider when developing national health accounts include:
Current health expenditure – the key variable for international health spending comparisons – is defined as the final consumption of healthcare good and services by the resident population plus subsidies to resident health providers.
The decision whether an activity should be included within current health expenditure should be determined by whether four main criteria are met:
the primary intent of the activity is to improve, maintain or prevent the deterioration of the health status of individuals, groups of the population or the population as a whole as well as to mitigate the consequences of ill health;
qualified medical or healthcare knowledge and skills are needed in carrying out this function, or it can be executed under the supervision of those with such knowledge, or the function is governance and health system administration and its financing;
the consumption is for the final use of healthcare goods and services by residents;
there is a transaction of healthcare services or goods.
The valuation of final consumption expenditure should be in purchaser prices, the prices parties agreed to for the transaction. But since the concept of “sales” can be alien to the health sector in many countries an alternative valuation can be used. Overall, the approach to value the output of health providers (and hence the consumption of these services) depends on whether they can be considered as market producers or not. If they are market producers, then the sales prices (including value added tax, if applicable) of the transactions should be used. However, the output of non-market producers should be measured by the sum of the costs of production. This includes (i) intermediate consumption (e.g. electricity used in a health facility), (ii) compensation of employees (e.g. salaries and employer charges for healthcare staff), (iii) consumption of fixed capital (e.g. depreciation of assets used in the production of health services), and (iv) other taxes paid on production.
The accrual principle should be applied when recording health spending. This means a transaction should be recorded in the time period when the service was delivered and “consumed”, which is not necessarily the same time period as when the services were paid.
How National Health Accounts based on SHA 2011 are set up in practice across the OECD and how data sources are combined and health spending estimated is country-specific, largely depending on data source availability and particular health system characteristics. In very broad terms, countries follow three different approaches:
Focus on identifying appropriate and detailed health spending data for each of the numerous financing schemes existing in the country;
Rely mainly on output information (e.g. revenues and activity data) from the side of health providers;
(Ideally) use a mixed approach by combining and triangulating data from the side of financing schemes with those available at the provider level.
In their current implementation, Brazil follows the first approach. In order to do this, the key tasks to provide a general description of a financing system from a health accounting point of view are as follows:
Identification of the national healthcare financing schemes: This involves compiling an inventory of all financing schemes by which the resident population gains access to healthcare with key details, including the mode of participation (whether compulsory, automatic, or voluntary), benefit entitlement rules, and the mechanisms for fund-raising and pooling. Additional information on the benefit package covered by each scheme can further enhance the understanding of the system.
Mapping the structure of the health financing system: This step entails charting out the entire health financing landscape by clearly establishing the relationship between financing schemes and the financing agents responsible for managing these financial flows. Financing agents include government entities, private insurers, social security institutions, and non-governmental organisations. It is crucial to clearly define which agents are involved in administering, collecting, and distributing funds within each scheme.
Identification of the basic flows in health financing: This last step is essential for understanding how resources are mobilised and allocated within the health sector. It involves analysing two core financial flows: (i) revenue‑raising by each financing scheme, which includes contributions from taxes, social security payments, private insurance premiums, and out-of-pocket expenditures; and (ii) allocation of these resources by the financing schemes according to key dimensions such as health functions (e.g. inpatient care, outpatient care, pharmaceuticals), providers (hospitals, clinics, pharmacies), and beneficiaries. A clear mapping of these flows is fundamental for ensuring transparency, efficiency, and accountability in the financing of healthcare services and to identify the best data sources.
4.1.1. Identified financing schemes in the Brazilian health accounts manual
In SHA 2011, financing schemes are identified as the key financing blocks or “body of rules” through which people obtain health services, including direct payments for health service utilisation by private households themselves. When identifying financing arrangements that qualify as financing schemes and in the distinction between different financing schemes, SHA 2011 proposes a set of key criteria (OECD/WHO/Eurostat, 2011[2]):
The mode of participation refers to whether the participation in a scheme is automatic for the population (or parts thereof), or compulsory by law, or voluntary – that is, at the discretion of the individual, population or other entities.
Whether beneficiaries’ entitlement to benefits is based on contributory payments by beneficiaries or universal, or at the discretion of the financing agents implementing the scheme.
On the basic method for fund-raising of a scheme to mobilise the necessary financial resources to provide healthcare for beneficiaries- whether these resources stem from transfers of government budgets, compulsory income‑related social contributions, compulsory or voluntary risk-related insurance premiums, or other sources of revenue.
And whether, at the level of the financing scheme, revenues are pooled across individuals or not.
Applying these criteria and complementary considerations, the Brazilian health accounts team has identified the following schemes for Brazil (Table 4.1).
Source: Authors based on the Brazilian National Health Accounts Manual.
Sistema Único de Saúde (SUS)
As set out in the federal constitution and by virtue of laws 8 080 and 8 142, the entire resident population of Brazil has access to services under the Sistema Único de Saúde (SUS). Access to SUS services is hence automatic and does not require any contributory payments by individuals. SUS is financed by the federation, the states and the municipalities out of the revenues these levels of government can mobilise – tax income, social contributions and other transfers (see Chapter 3). Since access to care is automatic for the whole population and independent of any contribution payment, SUS should be considered as a government financing scheme (HF11).
Schemes for Civil Servants
Separate public schemes to access healthcare exist for civil servants, at a federal, state and municipality level.1 Active civil servants and their dependants, as well as retired civil servants are entitled to a range of services under these “closed” schemes. Beneficiaries of the federal scheme can receive medical, hospital, dental, psychological and pharmaceutical care and these services may be provided by the public sector, or service delivery under this scheme may be organised differently by the entity (e.g. ministry) to which the civil servant is linked.2 Compulsory contribution payments by beneficiaries appear to be required.3
Yet, in practice there seem to exist considerable differences across the various schemes at the federal level (including across different federal entities), and at the state and municipality levels. This concerns not only the scope of entitlement to services but also the extent to which civil servants are mandated to contribute to and participate in the scheme or not. The various levels of government also provide funding for these schemes.
In the Brazilian NHA, the schemes for civil servants have been collectively classified as government financing schemes (HF11). Yet, in terms of scope, only spending by schemes at the federal and state level is currently recorded; spending by schemes on the municipality level is so far missing.
Schemes for Armed Forces and Public Security Forces
In Brazil, another type of “closed” public scheme exists, which covers personnel in the armed forces and public security forces (and their dependents). Again, these schemes exist at the federal, state and municipality level. They are financed by compulsory contributory payments by the beneficiary as well as complementary government funding.
In the Brazilian health accounts these schemes have been categorised as social health insurance schemes (HF121) based on the fact that they all require compulsory contribution payments by beneficiaries. Currently, only health spending by the federal scheme of the armed forces can be identified in the Brazilian health accounts. Expenditure by similar schemes at other levels of government are currently unaccounted for.
Voluntary health insurance
As discussed in Chapter 3, around a quarter of the population generally choose to forgo their entitlement to healthcare services under the public SUS and instead use private health insurance coverage. These insurance policies can either be purchased individually or are offered to employees as part of their employment benefits. The numerous private insurance plans are regulated by the ANS. What they have in common is that the uptake is voluntary (meaning there is no obligation to purchase private coverage) and dependent on contributory premium payments, either by or on behalf of the beneficiary. As such, coverage under these types of insurance contracts should be classified as voluntary health insurance (HF21) according to the SHA 2011 methodology.
Out-of-pocket payments
In all health systems around the world, people in need of care are required to finance some healthcare goods and services out-of-pocket. This can take the form of cost-sharing as part of a publicly or privately financed benefit package (HF32) or can take the form of direct purchases for goods and services without involvement of a third-party payer (HF31). Brazil is no exception to this.
Overall assessment of the identification and classification of health financing schemes
The new Brazilian health accounts identify health spending for the key financing schemes that exist in Brazil. There are two types of schemes for “closed” clientele for which the spending is currently not comprehensively accounted. This refers to the schemes for civil servants at the municipality level and schemes for public security forces beyond the armed forces. The associated reporting gaps are difficult to quantify but efforts should be made in the next iteration of the health accounts to bridge these gaps by appropriately estimating spending of those schemes that are currently not included.
Regarding the schemes for civil servants, there is potentially one conceptual problem with the current implementation in Brazil: they have been allocated to category HF11 (government financing schemes) regardless of whether contributory payments by the beneficiaries are compulsory or not. This is not necessarily in line with international best practice. Ideally, these schemes would be allocated to the various financing scheme categories on a case‑by-case basis – and informed by the existing legislation at the federal, state and municipal level. The schemes should be allocated to:
HF11 (government financing schemes), in case coverage is automatic by virtue of the law and beneficiaries do not have to make any contributory payments;
HF121 (social health insurance schemes), in case any contributory payments by beneficiaries are compulsory;
HF21 (voluntary health insurance schemes), in case any contributory payments by beneficiaries are voluntary.
Given the potentially high number of civil servant schemes, it might be difficult in practice to make this type of assessment on the level of the individual schemes (which could go into the thousands). In that case, it could be admissible to collectively allocate all schemes of civil servants to one single category based on the majority principle.
Related to this, a footnote in the methodological section of the Brazilian NHA manual raises the question as to whether health workers fit into the closed clientele plans, like other public servants, or whether, they are part of the SUS administration and as such form part of the expenses of the management of the system within SUS (HF111). In the manual it is indicated that these civil servants are allocated to the SUS financing regime. In reality, SUS workers in several government entities have coverage under closed clientele plans or private plans subsidised by governments regardless of their access to SUS services. This should be further evaluated to ensure the appropriate allocation.
To close the discussion on public/compulsory schemes, one question is whether spending on SUS fully captures the healthcare purchases by federal, state and municipal entities. Experience from OECD countries suggest that other avenues to finance healthcare may exist beyond the “main” financing scheme, for example, regarding the financing of health services in schools, in prisons or for long-term care. An example of the latter would be in England, where the NHS finances a vast array of health services for the entire population but much of long-term care spending is financed by local authorities (outside of the NHS). In the case of Brazil, while the vast majority of public spending is being captured it might be worthwhile to further explore whether some public spending for specific health services exists outside of SUS.
In the area of voluntary schemes, health expenditure by non-profit institutions financing schemes (HF22), while existing in Brazil could not be identified. Generally, this refers to healthcare goods and services that non-governmental organisations (NGOs) or charities (acting as a financing scheme) finance for the population or selected groups of the population. To purchase these services, these institutions may receive funding from government entities, from domestic sources via donations or membership fees, or from abroad. This may refer, for example, to healthcare activities organised by the Red Cross and financed via donations but also government support. Across the OECD, 28 countries were able to identify spending by this type of scheme in 2022. On average, it accounted for less than 1% of total health expenditure. In low- and medium-income countries charities tend to play a bigger role as a lot of international aid flows can be funnelled through domestic NGOs. In the case of Brazil, it may be worthwhile to scrutinise whether a small amount of the funding from government entities currently identified as SUS spending (HF111) would more appropriately refer to transfers to some non-profit financing schemes.
Likewise, spending from enterprise financing schemes (HF23) is also not reported in Brazil. The most important element in this category relates to health services provided by employers to their employees (e.g. health check-ups at the workplace by company doctors). Some OECD countries use information included as intermediate consumption in the National Accounts data (input-output table) or cost structure statistics of industries as sources to identify this type of spending. Another approach could be to access occupational medicine records that may be available in the Ministry of Labour. In addition, this category is used to report spending by “healthcare providers financing schemes”. This generally refer to situations when providers need to finance part of the services they deliver to their patients from their own revenues (which are additional sources to the funding they receive from the financing schemes) – for example, if private providers operate at a financial loss. In 2022, 27 OECD countries reported spending by this type of scheme, albeit representing less than 1% of overall health spending, on average.
In line with most OECD countries, spending by non-domestic financing schemes (HF4) are also not included in the Brazilian health accounts. These schemes can be important in countries where a significant part of the population is covered for healthcare costs through non-domestic financing schemes or where external aid agencies play a decisive role in financing healthcare programmes without domestic intermediaries. None of these cases seem to be relevant in Brazil.
In summary, the key financing schemes in Brazil have all been identified. There are some reporting gaps and potential classification issues for public schemes with a closed clientele. There are also two voluntary scheme categories that are missing although typically both only account for a minimal share of overall health spending. Yet, going forward, efforts should be made to detect data sources that could allow for a comprehensive reporting of all existing financing schemes, even if it does not significantly change the overall picture of health spending in Brazil.
Good practice in implementing and eventually institutionalising health accounts in countries calls for a systematic screening of all potentially available data source that can be of use in the estimation of health spending. This includes, of course, any data on financial transactions in the health sector but can also cover secondary data on activities or physical resources, which may be helpful to generate “allocation keys” to distribute health spending across different healthcare functions (HC) or providers (HP). When identifying potential data sources, ideally, all health system dimensions should be considered, that means including data sources that refer to spending by financing agents and schemes, but also provider – or health industry-specific data sources (e.g. covering the hospital sector or the pharmaceutical industries). Reviewing information on health expenditure potentially included in other statistical systems such as the System of National Accounts is also recommended.
4.2.1. Brazil has access to comprehensive data sources from the financing perspective
Brazil can tap into a rich health data infrastructure, in particular when it comes to measuring of expenditure and activity related to SUS. The Brazilian NHA manual suggests that available data sources from financing schemes and agents have been comprehensively explored. Figure 4.1 provides a visualisation of identified data sources that are potentially appropriate for the calculation of health spending. Table 4.2 summarises some key characteristics of each of these data sources.
Note: CNES: National Registry of Health Establishments, DIOPS: Periodic Information Document of Health Plan Operators, FNS: National Health Fund, IRPF: Individual Income Tax, PAIC: Annual Civil Construction Survey, PIA: Annual Industry Survey, PMC: Monthly Commerce Survey, POF: Household Budget Survey, SIA: Ambulatory Information System, SIAFI: Integrated System of Federal Government Financial Administration, SIGA Brasil: Advanced Managerial Budgetary Information System Brazil, SIGTAP: Management System of the Procedures Table of SUS, SIOPS: Information System on Public Health Budgets, SIP: Product Information System, SIH: Hospital Information System, SISAB: Primary Care Health Information System, SI-PNI: Information System of the National Immunisation Programme, TISS: Supplementary Health Information Exchange Standard. FORÇAS: Armed Forces.
Source: Authors based on Brazilian Health Accounts Manual.
Source: Authors based on Brazilian Health Accounts Manual.
Below are the key characteristics of the essential data sources that the Brazilian team has considered in the health accounts development phase. A number of data sources or data portals are available to retrieve data on public spending but also spending by private insurance and households. In addition, various data sources exist to measure the utilisation of different health services.
Sistema Integrado de Administração Financeira do Governo Federal (SIAFI)
The Integrated System of Federal Government Financial Administration (SIAFI) is a tool used to monitor the budget and the financial management of entities belonging to the federal government. The tool is not publicly available and SIAFI is used in particular by the National Treasury of Brazil, which is responsible for the administration and monitoring of federal government financial resources.
The granularity of the data in the system is focused on the level of the federal government, but also makes it possible to identify the amounts transferred to other federative entities, such as states, the federal district, and municipalities (https://siafi.tesouro.gov.br/).
SIGA Brasil
The objective of the online portal SIGA Brasil is to integrate data from different budget systems, such as SIAFI and the Integrated System of General Services Administration (SIASG), into a single repository, and to make information of the federal budget publicly accessible.
SIGA Brasil provides detailed data on budget planning and execution, including expenses, revenues, parliamentary amendments and transfers. The level of data granularity allows for detailed analysis of expenditures according to (sub)functions, including for health spending, type of transaction and executing agency, among others (www12.senado.leg.br/orcamento/sigabrasil). Figure 4.2 showcases some high-level analysis on health spending that can be carried out using this data portal.
Note: Data refers to spending at the federal level. Refers to executed payments including settled payables from previous years.
Source: SIGA Brasil (n.d.[3]), SIGA Brasil, www12.senado.leg.br/orcamento/sigabrasil, (accessed on 4 March 2025).
Sistema de Informações sobre Orçamentos Públicos em Saúde (SIOPS)
The Information System on Public Health Budgets (SIOPS) was created to monitor whether the municipalities, the states and the federal district, and the federal Government of Brazil meet their minimum spending obligation in the area of health. SIOPS includes information on revenues and expenses for health services, in accordance with the classification defined by the National Treasury Secretariat. The use of the system is mandatory, and it provides transparency and control over the minimum application of health resources by each federative entity.
Data in SIOPS is organised according to the health budget subfunctions and can be traced back to the origin of the resources (in case of transfer across levels of government). However, the health subfunctions (primary care, outpatient and hospital care, prophylactic and therapeutic support, among others) do not have a direct correspondence with the SHA care functions (ICHA-HC). Information included in SIOPS is publicly available (http://siops.datasus.gov.br).
Transparency portals of the states and the Federal District (portais de transparência)
In accordance with requirements to ensure transparency of administrative acts at all levels of government, state‑specific transparency portals provide publicly available data on budget expenses and revenues, among other things. The data covers, for example, public spending on all human resources including for health plans by civil servants of the states.
Sistema de Informações em Saúde para a Atenção Básica (SISAB-SUS)
The Health Information System for Primary Healthcare (SISAB-SUS) records and provides detailed information on primary healthcare services that are delivered within the scope of SUS. The system uses medical records as a basis for data collection and contains information on consultations, home visits, dental procedures and collective activities. It covers non-financial data on health acts as well as on the teams managing primary healthcare in municipalities, states and the federation. SISAB-SUS is a crucial tool used to monitor the progress in access to primary healthcare throughout the country.
The data is publicly available through the SISAB portal (https://sisab.saude.gov.br/).
Sistemas de Informações Ambulatoriais (SIA-SUS)
The Outpatient Information System (SIA-SUS) was created to record and process data on all outpatient care (but excluding primary healthcare) provided in public and private facilities contracted by SUS. The system uses the Authorization of Outpatient Procedures (APAC) as a basis for data collection, covering information on medical consultations, exams, procedures and their costs according to the SUS Table (SIGTAP).
The maintenance of SIA-SUS is the responsibility of the Ministry of Health. Data is collected monthly and made available on the DATASUS portal with a time lag of around three months. Detailed data is available at the level of the state, municipality and even facility (https://datasus.saude.gov.br/acesso-a-informacao/producao-ambulatorial-sia-sus/).
Sistema de Informações Hospitalares (SIH-SUS)
The Hospital Information System (SIH-SUS) records and processes data related to hospital services provided within the scope of the SUS covering both, public and private facilities. Based on the hospital admission authorisation, for each hospitalisation information requirements cover diagnoses, procedures performed, length of stay and costs of procedures according to the SUS Table (SIGTAP).
The Ministry of Health is responsible for maintaining SIH-SUS and data is available via DATASUS on a facility, municipality and state level. (http://datasus.saude.gov.br/acesso-a-informacao/producao-hospitalar-sih-sus/).
Sistema de Gerenciamento da Tabela de Procedimentos, Medicamentos e OPM do SUS (SIGTAP)
SIGTAP is a management system to manage more than 4 600 items on procedures, medications, medical devices such as orthosis and prothesis and special materials used in SUS (Table 4.3). Developed by the Ministry of Health, its aim is to increase transparency in resource use across the various levels of government. It is used for resource management and billing and serves as a basis for financial transfers between the federal government, the states and federal district, and municipalities.
The information is publicly accessible on the system’s website (https://sigtap.datasus.gov.br/). SIGTAP can be combined with the activity data includes in SISAB-SUS and SIA-SUS to estimate health expenditure for those services.
Note: Classification only shows the activity groups at the highest level and examples of subgroups.
Source: DATASUS (n.d.[4]), SIGTAP – Sistema de Gerenciamento da Tabela de Procedimentos, Medicamentos e OPM do SUS http://sigtap.datasus.gov.br/tabela-unificada/app/sec/inicio.jsp, (accessed on 4 March 2025).
Documento de Informações Periódicas das Operadoras de Planos de Assistência à Saúde (DIOPS/ANS)
The Periodic Information Documents of Private Health Insurance operators (DIOPS) are a set of standard information that all entities operating voluntary health insurance need to submit quarterly to the National Supplementary Health Agency (ANS), the federal agency regulating private health insurance. DIOPS allows the ANS to monitor the economic and financial situation of all health insurance operators, including financial statements on revenues and expenses (Table 4.4).
Data is available two months after the end of each quarter and can be publicly accessed on the ANS web portal (Dados do Setor – Agência Nacional de Saúde Suplementar).
Revenues and expenses by all private health insurance operators, in million BRL, 2019‑24
Source: ANS (2025[5]), Dados Gerais – Receita de contraprestações das operadoras (em Reais) (Brasil – 2015-2024), https://www.gov.br/ans/pt-br/acesso-a-informacao/perfil-do-setor/dados-gerais (accessed on 31 March 2025).
Troca de Informações na Saúde Suplementar (TISS/ANS)
The Supplementary Health Information Exchange Standard (TISS) is hosted by ANS. Since 2012, all health insurance operators are required to use a standard format to submit data on health services performed within the scope of plans regulated by ANS. Data is very granular and includes information on the beneficiaries receiving services, providers of the services, procedures performed, associated costs and quantity of services. It includes several thousand procedures.
Data is available quarterly and can be analysed via an analytical IT tool on a data portal, and allows, for example, a comparison of average costs for the same procedure across states or the total number of procedures performed (https://www.gov.br/ans/pt-br/acesso-a-informacao/perfil-do-setor/dados-e-indicadores-do-setor/d-tiss-painel-dos-dados-do-tiss).
Sistema de Informações sobre Produtos (SIP/ANS)
The Product Information System (SIP) is hosted and monitored by the ANS. All private health insurance operators have to report the number of expense claims by category, such as preventive procedures, consultations, tests, hospitalisations, etc. They also need to report costs for groups of activity at a highly aggregated level (e.g. for hospitalisations, imaging services).
In SIP, private health insurance operators also must report the cost-sharing of beneficiaries, in case private insurance is not covering the full cost of a specific service.
Pesquisa de Orçamentos Familiares (POF/IBGE)
The household budget survey (POF) is carried out by the Brazilian Institute of Geography and Statistics (IBGE) on a regular but infrequent basis. The last available data stems from the 2018 wave. In household budget surveys, a representative sample of households is required to provide information about their income as well as their spending and consumption patterns, including on healthcare goods and services. Household budget surveys are an important source to estimate out-of-pocket spending in national health accounts in many OECD countries and beyond.
Cadastro Nacional de Estabelecimentos de Saúde (CNES)
The National Registry of Health Establishments (CNES) is of essential importance for a possible extension of the application of health accounts to also identify health spending by health provider (HP). CNES covers all licensed outpatient and inpatient health establishments in Brazil and includes information on location, workforce, equipment, and services provided among other things. SUS databases on activities requires using CNES codes which would allow for a mapping between services and providers. There are 45 provider categories in CNES that could be mapped into the international HP classification.
Assessment of data sources
For the production of health accounts, Brazil can rely on numerous budgetary and financial data sources as well as sources to monitor primary healthcare, outpatient specialist and hospital activity in SUS. Data sources that reflect reference prices for products and activity also exist. Compared to a number of OECD countries, Brazil appears to be in privileged position when it comes to data availability of private health insurance. It is understood that for 2024 onwards it may be possible to produce tri-axial data for HF.2.1. Yet, data sources on health spending (or revenues) from an industry or health provider perspective seem to be missing in Brazil. This type of data is generally extremely useful to triangulate with data obtained from the financing side and can validate health accounts results and further calibrate them.
4.3.1. Overall health spending is estimated by financing schemes before allocating it to the various services
As mentioned in Section 4.1, the general approach taken by Brazil to estimate health spending, as documented in their National Health Accounts manual, prioritises data sources from the financing schemes’ perspective. This is an approach common with several countries across Latin America, but also often in countries with large government-funded health systems, such as the United Kingdom. What this means in practice is that for a particular financing scheme, one or more data sources are used to establish the value of overall health expenditure by financing scheme. Then in a second step, this spending is distributed to the various healthcare functions (HC).4 In future iterations of the health accounts in Brazil, this second step should also include allocating health spending to health providers (HP).
SUS spending
As discussed in Chapter 3, the Unified Health System (Sistema Único de Saúde, SUS) is the key financing scheme in the Brazilian health system providing universal rights to access a set of comprehensive healthcare goods and services for the population from primary healthcare through to tertiary care, and financed through a complex mix of federal, state and municipal funding. As such, there is clear importance in adopting a robust methodology to determine accurate estimates of both the overall expenditures of SUS, as well as, in the second step, the comprehensive allocation of spending across the functions of care (and/or providers) through the integration of more disaggregated financial data or secondary information, such as activities and unit costs.
Constructing a “control total”
The starting point is to derive a “control total” (or spending estimate on the highest level of aggregation) using aggregated data sources that offer sufficient detail to adhere closely to the health boundary defined under SHA. The funding for SUS comes from the three levels of government (federal, states and municipalities), complicating the process and requiring consistent application of concepts and boundaries to determine an overall spending estimate.
For the federal budget, expenses data for the Ministry of Health and Ministry of Education5 are extracted in SIAFI according to health and education functions as defined under the Brazilian public budget monitoring system (MCASP). For the health-related sub-functions, constructing and applying filters to align with the SHA boundaries and concepts is a crucial step for correctly determining the inclusion or exclusion of specific expense items. For example, certain health sub-functions could in theory be combined with the overall education function and vice versa. A key item in Brazil is health spending covered under the important aggregate ASPS (“public health actions and services”) but adjustments are made. For example, capital expenses included under ASPS in SIAFI are correctly noted as being excluded for the purposes of health accounts. Non-ASPS budgetary units linked to the Ministry of Health (e.g. National Health Fund, Fiocruz, Funasa, National Supplementary Health Agency, National Health Surveillance Agency) are identified and included. Including these transactions under SUS spending (HF111) is appropriate from a financing scheme perspective if these agencies can be considered as financing agents implementing SUS. They would need to be classified differently if otherwise. In that case, the transfer from SUS would need to be considered as a “revenue flow” to another financing schemes. This distinction could be important for analysing SUS financing/expenditure in later stages of health accounts development in Brazil.
The health sub-functions in the Brazilian public budget monitoring system are highlighted in (Table 4.5).
Source: Ministério do Planejamento, Orçamento e Gestão (n.d.[6]), Orçamento Federal, www.orcamentofederal.gov.br/orcamentos-anuais/orcamento‑1999/Portaria_Ministerial_42_de_140499.pdf, (accessed on 20 March 2025).
The relevant health expenditures under the education function (financed from the budget of the Ministry of Education) are limited to those incurred by university hospitals. Identification can also be made by identifying the institutions that execute the budget e.g. in the case of health institutions linked to the Ministry of Education. The methodology in the manual explicitly refers to the inclusion of “current expenses in actions related to the modernisation of these institutions”, but it is unclear if these should be considered capital expenditures and therefore excluded. The methodology correctly identifies the Ministry of Education expenditures as separate transfers in SIAFI from regular MAC ceiling funding, ensuring additional expenses are clearly accounted for.
An update in the new methodology (compared to the health accounts pilot implementation) aligns the timing of budget execution at the federal level with that at the state, district, and municipal levels, using “liquidated expenses” as the expenditure item. While this provides consistency across sources and better reflects the expenditures of the federal government, it may not completely align with the SHA principle of accrual accounting, which measures expenditure based on consumption time rather than payment or liquidation time. Including commitments or payments to be made might be more appropriate if there are significant time lags.
The expenditures of states, the federal district, and municipalities are identified using SIOPS. SIOPS classifies and reports the origin of each resource, ensuring resources transferred from the federal government to subnational entities are not counted as part of the subnational entities’ own resources. These transfers are categorised separately and considered federal expenses, even when used by states or municipalities. This distinction eliminates the need to subtract federal transfers from total expenditures, as they are already recorded as federal expenditures, preventing double counting, and ensuring funds are correctly categorised (e.g. federal vs. state or municipal spending).
The manual adjusts the ASPS current expenses total identified in SIOPS to exclude payments related to “Retirees and Pensioners”, as they are no longer part of the productive workforce. Clarifying the types of payments that are excluded could help determine if the exclusion is correct, as pension payments to retired medical staff are typically classified under social security spending rather than health spending.
Thus, total SUS spending is determined by adding up designated health spending from the federal government (using SIAFI), and from states, the federal district and municipalities (using SIOPS). Aside from the points raised above, the methodology for calculating the control total for SUS appears to be robust, based on the most accurate sources of information available linked to budget execution, even if these sources themselves do not provide detailed production data.
The SUS expenditures are classified according to different financing mechanisms, such as Primary Health Care (PAB), Medium and High Complexity Care (MAC), and the Strategic Actions and Compensation Fund (FAEC), based on the existing budget structures. The method of allocation of these subcomponents is further assessed in the next section.
Allocating across functions
After determining the overall spending of SUS this amount needs to be allocated to healthcare functions (HC) to reflect how these financial resources are used. To do this, one or more “allocation keys” need to be calculated. This approach involves linking detailed health production data (i.e. volumes of services) to aggregated budget data (Table 4.3). This allows for a granular breakdown of health expenditures by functions and funding agents (federal, state, or municipal). However, reconciling the adjusted totals relies on expansion factors derived from SUS production and procedure reimbursement data.
Source: Brazilian National Health Accounts Manual.
The key issue for Brazil is that the budget data used to define total SUS spending only has a limited number of functional subcategories which impedes a comprehensive allocation of SUS spending to the SHA functional categories. To amend this, databases that include information on utilisation and costs are leveraged to provide information at a much more detailed service level.
The approach uses the health information systems – SIA/SUS, SIH/SUS, SISAB, and SI-PNI – as activity databases in conjunction with SIGTAP, which includes price information for services and procedures. The detailed procedure list in SIGTAP is mapped to the HC functional classification. This ensures health expenditures are correctly categorised by service type, financing entity, and geographic level. The result is a detailed table linking financial data with service production, providing a more accurate representation of healthcare spending across different government levels.
There are some challenges. SIGTAP does not exactly map with the functional classification. While it classifies procedures, medications, and medical devices within SUS, it does not inherently categorise them by healthcare setting (e.g. hospital vs. ambulatory care). Additional data processing may be required to allocate SIGTAP procedures accurately within the SHA framework. However, distinctions can be inferred indirectly using other SUS databases, such as SIA/SUS (ambulatory care) and SIH/SUS (hospital care). Also, SIGTAP lacks direct financial information for Primary Health Care (PAB) procedures, necessitating the use of proxies to estimate expenditures. Specifically, these are based on a 2002 report that attributed values to these procedures and which are then updated based on annual inflation.
Once the mapping and the detailed expenditure structure based on combining activity databases with SIGTAP is completed, the aggregates must be aligned with the aggregate budget expenditures using expansion factors which vary by healthcare service type. These expansion factors adjust the detailed breakdown of healthcare expenditures to ensure that when aggregated together are in line with the predefined “control totals” for subcomponent.
For each of the subcomponents, the expansion factor and process is different. For MAC, expenditures are calculated using an expenditure weight approach by multiplying recorded financial values in health information systems (SIA, SIH and SISAB) by the MAC expansion factor. Since primary care procedure values funded by PAB are not directly available in SIGTAP, they must be estimated using cost proxies. Expenditures are calculated on a unit weight basis by multiplying the number of procedures performed by the unit cost and then inflated using the PAB expansion factor. FAEC expenditures, since fully funded by the federal government, do not require expansion factors.
In summary, despite its overall strengths, the allocation approach still has some limitations. The reliance on historical proxy cost estimates (in SIGTAP) in particular may introduce inaccuracies, especially given regional variations and changes in the relative healthcare costs between procedures and services. Discrepancies between different information systems could impact data reliability. To improve accuracy, greater harmonisation between health and financial databases is needed, along with periodic updates to cost proxies to reflect current healthcare expenditures more precisely. Future refinements in data integration and classification techniques could further enhance the reliability of SUS expenditure estimates.
Estimating pharmaceutical expenditures under SUS
The identification of pharmaceutical spending under SUS is independent of the methods using expansion factors. The methodology to estimate pharmaceutical spending by SUS uses the same approach as in the Health Satellite Account with an important difference. At the subnational level (states, federal district, and municipalities) the estimation is derived from the sum of the expenditure categories related to pharmaceuticals in SIOPS database. As with other spending the focus is on liquidated expenditures. At the federal level data from SIAFI is used. The estimation includes expenditures classified under Function 10 (Health), with a focus on current expenses and budgetary actions linked to pharmaceutical assistance programmes. The “Programa Farmácia Popular do Brasil” (PFPB), which provides free or subsidised medications for chronic diseases, is a key component of pharmaceutical spending. The estimation does not rely on pharmaceutical assistance procedures recorded in the SIA/SUS system. This could lead to a potential underestimation of pharmaceutical (retail) spending in SHA in case the pharmaceutical assistance procedures recorded in SIA/SUS also refer the direct dispensing of medication to patients (and not used as part of a hospital inpatient treatment).
Unlike the Health Satellite Account, pharmaceutical spending includes the federal transfers to the PFPB. This ensures that the SHA captures the expenditures related to pharmaceutical consumption while maintaining consistency with national health accounting methodologies.
Spending by schemes for civil servants
The closed clientele financing regime covers health plans for public servants, but its mapping remains incomplete, particularly at the municipal level. Currently, only the federal government and most state governments are included in the estimates. This limitation means that potentially a portion of public servant health expenditures across Brazil remains unaccounted for. In the case of state employees, a prorate increase based on the number of public servants in the missing states could be a simple interim measure to avoid any significant underestimation. The extent of the use of the closed plans by municipal civil servants is something that should be explored. They comprise a potentially large group of employees although they may be perhaps less inclined to be part of the civil servant scheme than federal and state employees. Furthermore, as mentioned in the section of financing schemes in Section 4.1, the treatment of health workers as civil servants and their associated medical expenses with regard to the closed scheme should also be reviewed.
To estimate these expenditures, data sources vary by government level. For the federal government, information is obtained from SIAFI, considering only current (liquidated) expenses while excluding expenditures from the Ministry of Health (included under SUS) as well as from the Ministry of Defence (included separately as a scheme for military). For each of the states, data is collected from the Transparency Portals, and legal frameworks are reviewed to determine whether health plans for public servants exist. Reports are generated to quantify spending by institutions responsible for managing these plans. There appears to be no data available to identify functional categories of the health spending by these closed schemes. The distribution of spending by service category estimated for health spending by voluntary health insurance (HF21) is used as a proxy to break down health spending by type of healthcare.
In summary, the methodology has several limitations. The municipal-level health plans remain unmapped, and there is a need for further research to clarify the legal basis for these expenditures. Additionally, it is still unclear which branches of government (executive, legislative, judicial) are covered in the estimates. Expanding the scope of data collection and harmonising methodologies across different government levels would improve the accuracy of these estimates in future revisions. Lastly, deriving a specific functional breakdown instead of relying on a proxy distribution to allocate total health spending could be explored in the future.
Spending by schemes for military and armed forces
The estimation of government compulsory contributory schemes (HF.1.2) focuses on federal health expenditures for the Armed Forces. Similar spending by schemes for public security forces at the state and municipal levels are so far missing. In the current approach, the methodology primarily relies on SIAFI data to extract liquidated expenses under the national defence function, filtering them to isolate health-related expenditures. This filtering process considers three key elements: (i) expenditures from management units executing the Ministry of Defence’s health budget, (ii) health-related expenditures in other military budget units, and (iii) spending on budgetary actions directly tied to healthcare services and legally required health benefits. This structured approach ensures that only relevant expenditures are classified under HF.1.2.
Once these expenditures are estimated, they are allocated to healthcare functions (HC). However, due to the lack of public, disaggregated data on health services provided within the military system, the methodology relies on a proxy – the production data of federal hospitals from SIA and SIH – to distribute these expenses across health functions. This assumption is based on the idea that federal hospital production patterns can approximate the utilisation of military health services. While this provides a practical method for estimating service distribution, it may not fully capture the distinct characteristics of military healthcare, such as specialised care for active‑duty personnel, different funding structures, and military-specific treatment facilities.
Despite the methodological coherence, there are areas for improvement. Options should be explored to comprehensively capture spending for these schemes beyond the Armed Forces. For the functional spending breakdown, the reliance on proxies introduces potential inaccuracies, as federal hospital data may not accurately reflect military health service utilisation. Additionally, the absence of detailed, public data on military health production limits the precision of cost allocations. To enhance accuracy, more direct data sources – such as internal Armed Forces health reports or audits– should be integrated into the estimation process. Greater transparency and standardisation in recording military health expenditures would also help refine future analyses and align them more closely with SHA functional classifications.
Spending by voluntary health insurance
The approach used to estimate health spending of voluntary health insurance overall and by healthcare function is similar to the one applied for SUS but less complex.
The overall spending by voluntary health insurance is generated from DIOPS, the monitoring tool that allows the ANS to assess the financial situation of all health insurance operators and includes financial statements on revenues and expenses. To determine the total spending of voluntary health insurance, all the high-level cost items by all health insurance operators included in DIOPS are aggregated: This includes insurance expenses for claims (or indemnifiable events), other operational expenses, marketing expenses, administrative expenses, and compensation for benefit administrators.
A first step to distribute this overall health expenditure figure for HF21 to the functional categories of SHA can be done with information included in DIOPS: All marketing expenses, administrative expenses, and the compensation for benefit administrators fulfil the description of “health systems administration” under SHA and are thus categorised as HC.7. The remaining costs are consequently considered under categories HC.1‑HC.6 and need to be allocated based on another method.
For this allocation, the Brazilian health accounts team combines the data information systems SIP and TISS, both managed by ANS. For all private plan operators, the SIP records the number and type of activity financed for private insurance beneficiaries and reports costs for activity groups on a highly level of aggregation. TISS includes more detailed information on the procedures performed for outpatient care and hospitals and can also generate average costs per activity. Thus, for the use in the Brazilian health accounts, SIP and TISS complement each other. Due to the lack of granularity, the sole use of SIP would not be sufficient to allocate voluntary health insurance spending to the detailed HC categories in SHA. To obtain the additional level of detail, TISS information needs to be added.
The approach used in Brazil to identify and distribute spending by voluntary health insurance looks generally robust. Compared to several OECD countries, Brazil appears to be in a privileged position when it comes to data availability to estimate spending by private health insurance. Yet, one issue that Brazil could address for the next iteration of the production of health accounts is the determination of administrative spending of private health insurance. While a good number of OECD countries also calculate total administrative spending for private insurance by the sum of several cost items, the actual recommendation in the SHA 2011 Manual is to measure administrative spending as “insurance output”, defined as total premiums earned plus premiums supplements less adjusted claims incurred. This method typically increases the estimated administrative costs as this will also include profits of health insurers.
Out-of-pocket spending
For the measurement of direct out-of-pocket spending excluding cost-sharing (HF31), the Brazilian health accounts team relies on the Household Budget Survey (Pesquisa de Orçamentos Familiares – POF) which is conducted regularly by IBGE. The main advantage of incorporating HBS data into health accounts lies in its widespread availability and applicability. Sample sizes of HBS are generally large enough to be statistically robust and reflect the entire population.6 The survey covers all health expenses borne privately, including cost-sharing with third-party payers as well as purchases initiated by the patient. Care must be exercised to exclude any premium payments to voluntary health insurance – these payments are not considered as out-of-pocket payments (HF3) but should be included as revenues of the financing scheme voluntary health insurance (HF21) – in case the HFxFS table is produced. One of the main disadvantages of using HBS is that they are carried out infrequently, typically only every five years.
In Brazil, the latest available data from POF stems from the 2018 survey7 in which ten health spending categories were included (Table 4.6). These categories can be mapped relatively clearly into the corresponding HC classifications.
Average monthly household monetary and non-monetary expenditure – value and distribution – by total income classes and monthly household wealth variation, according to types of expenditure, average monthly household monetary and non-monetary expenditure (in BRL), total, Brazil
Source: IBGE (2018[7]), “Pesquisa de Orçamentos Familiares”, https://sidra.ibge.gov.br/tabela/6715 (accessed on 22 March 2025).
In order to be of use in the health accounts for the reporting year 2022, these figures need to be extrapolated to take into account how out-of-pocket spending has evolved since 2018.8 These extrapolations reflect the development of prices and volumes consumed and are carried out using a variety of different data sources:
Data from the Product Information System (SIP/ANS) to create a single volume index that is used for all health services (except medications);
Data from a Monthly Survey of Commerce (PMC) to create a volume index for medicines based on the activity “Pharmaceutical, medical, orthopedic, perfumery and cosmetic articles”; and
The extended Consumer Price Index (IPCA)
Combining these data sources, the evolution of the individual out-of-pocket health spending estimates can be generated. Moreover, other data sources to verify these trends are also used. They include the Outpatient Information System (SIA/SUS) and Hospital Information System (SIH/SUS), as well as data on the individual income tax (IRPF).
For the estimation of cost-sharing with third-party financing schemes (HF32), the Brazilian health accounts teams has identified a data source managed by ANS. The SIP includes information on how much beneficiaries of voluntary health insurance have to pay out-of-pocket for services financed by private health plan operators on an aggregated level.
Overall, the Brazilian health accounts team resorts to standard data sources to estimate out-of-pocket spending. However, when it comes to using household budget surveys, caution must be exercised. Typically, household budget surveys are designed in a way that they are used to define the total amount of out-of-pocket spending HF.3 (either per services or overall) since they combine direct purchases of households (HF.3.1) and payments as a part of cost-sharing arrangements (HF3.2). The Brazilian manual suggests that the household budget survey is only used to measure direct purchases – this could potentially lead to double‑counting for payments under cost-sharing arrangements if they are not explicitly excluded from household budget survey. The exclusion of premium payments to health insurance from out-of-pocket payments is also essential if these are expenses covered in household budget surveys. Finally, in some instances, for example, out-of-pocket payments for long-term care household budget surveys are of limited use, since the target group using these services may be underrepresented in the budget surveys. Overall, the manual could include some more detail on the adjustments carried out to the two data sources used and how double‑counting issues were avoided.
When it comes to the extrapolation of OOP spending creating indices that incorporate annual price and volume changes is standard practice also in other countries. Yet, it would be advisable to create different indices for all identified health spending categories and not use a single index for all functions apart from medication. Moreover, continuing the exploration whether data sources from a provider perspective exist would help in validating out-of-pocket estimates generated solely using household budget surveys.
4.3.2. Mapping to healthcare services
In summary, based on the various approaches described above for each financing scheme, the Brazilian health accounts team can allocate overall health spending to a wide range of healthcare functions (Table 4.7) – a greater level of detail than what is reported in many OECD countries. In some instances, they go beyond the proposed international ICHA-HC categories to reflect country-specific information needs or necessary adaptations to existing categories. This refers, for example to spending on primary healthcare (according to the Brazilian definition), emergency care, or spending by the “Farmacia Popular” programme.
Note: SHA-BR refers to additional categories for information needs in Brazil, or adjustments made to existing categories within HC.
Source: Brazilian NHA manual.
The OECD assessment of the new approach to estimate health spending was based on the new Brazilian health accounts manual9 and detailed exchanges with the Brazilian health accounts team.
The general approach to construct health accounts in Brazil is in line with standard practice in a number of OECD countries. Within this approach, the focus lies on identifying the most appropriate data sources to estimate overall health expenditure for each financing scheme before allocating this spending aggregate to healthcare functions (HC) and providers (HP). This is generally done either by using information included in the primary data sources or constructing an “allocation key” from a secondary data source to distribute health spending to the most appropriate HC and HP categories. Total health spending is then determined by summing up health spending for each of the relevant financing schemes. While this is also practice in several OECD countries, it would be desirable to triangulate these results with data sources from the health provider perspective, such as hospital statistics or other industry statistics that can determine the economic output and production of a specific sector, or information on the production side of the health sector obtained from the National Accounts. It is understood that these type of health provider-specific data source are generally not available in the Brazilian context.
The choice of data sources used to establish total health expenditure per financing scheme in Brazil appears robust. Yet, there are some gaps that should be addressed in future iterations of the health accounts production. This refers, for example, to spending by the specific schemes for civil servants and security forces at state and municipal level, where spending has so far not been identified, as well as estimating the spending by charities and employer financing schemes. However, the latter are expected to play a minimal health financing role in Brazil.
When it comes to the detailed calculation of health spending on a service level the process becomes complex – as is the case in most OECD countries.
The method to determine overall SUS spending by combining data extractions from SIAFI and SIOPS appears robust as double counting of spending is ruled out. Regarding the allocation of total SUS spending by functional category a two‑tier approach appears to be used. First, spending is allocated to the key budget titles given that these are considered the most accurate source of aggregate spending at each level of government. However, since the level of disaggregation is not sufficient to allocate spending to the detailed functional categories, a number of “allocation keys” are generated based on volume information included in the SIA, SIH and SISAB and an indication of prices of services included in SIGTAP for a further breakdown of spending. This approach appears to be a practical solution to add the needed granularity although the precise calculations are not sufficiently detailed in the manual itself to allow a full assessment for those not overly familiar with the specifics of the data sources. In this respect, some illustrative examples in the annex to the manual could be beneficial. Regardless of the practical calculation method, one issue with basing an allocation on the SIGTAP price list is the fact that the list is outdated, and the application of these prices may lead to an over/underestimation of several spending components, particularly in the case where price inflation has varied significantly for different activities. However, there do not appear to exist any alternative approaches to using SIGTAP.
The approach to calculate and distribute spending by voluntary health insurance follows a similar approach to that for SUS spending but seems less complex in terms of fewer funding flows and thus information sources. The total expenses for administration and health services are determined by DIOPS and a breakdown on the service level is done combining SIP for highly aggregated categories and the price and quantity information included in TISS for the spending distribution on the more granular level. This approach appears to be sound in general, but the manual could include more specific information on how the different information included in SIP and TISS are combined in practice. That said, overall data source availability to measure and distribute private health insurance spending in Brazil appears to be more advanced than in a sizeable number of OECD countries.
The estimation of out-of-pocket spending should be reviewed for future iterations of the health accounts production. It is not fully clear in the manual whether some potential issues have been addressed. The manual mentions using the household budget survey to measure direct payment of households and SIP for copayment with private health insurance. However, household budget surveys normally already incorporate co-payments by households, so there the risk of double counting if these sources are used side‑by-side. It is also not fully clear which health spending elements from the household budget surveys are used. Moreover, any possible tax return due to elevated out-of-pocket spending should be considered when determining out-of-pocket payments according to the SHA methodology. Relying on household budget surveys and extrapolating spending with appropriate instruments is common practice to estimate overall out-of-pocket spending in several OECD countries. Yet, it appears that only two different indices are created to extrapolate the various health spending items. This should be reviewed. Improvements in the disaggregation of out-of-pocket spending into functions (and providers) will require further work to identify at least some supplementary primary sources or indexes.
Finally, there are data gaps that prevent a scheme‑specific distribution of health spending for the schemes of civil servants and the armed forces. For the former, the functional distribution of voluntary health insurance is used. It is difficult to ascertain whether spending by the schemes mirrors that of private health plans in reality but it is understood that no reliable data sources are currently identifiable to produce refined allocation schemes for these schemes. This could be an area to explore if progress is possible in future rounds of health accounts production.
Regarding the documentation, the Brazilian health accounts manual is seen as a good starting point for anyone who wants to learn more about how the Brazilian health accounts results are derived. It also provides a good introduction to understanding the complexities of the Brazilian health system. People unfamiliar with the System of Health Accounts are initiated to the general concept of health accounts. The manual also includes a very useful description of the available data sources, in particular their content, where they can be retrieved, the updating frequency and when data is available to be used in the health accounts production. It also attempts to explain the complex procedure how several data sources are combined to estimate health spending on a service level.
Moving forward, however, there is scope to expand some aspects of the manual. Regarding data sources, expert users may be interested to get a practical idea how the data structure actually looks – some data source extractions, could, for example, be included in an annex. When it comes to the detailed calculation process, in some areas, for example when it comes to the creating “expansion factors”, the manual could provide more detail. In general, the manual should be a “living document”, regularly updated if methods are revised and expanded if new demand for information arises. As such, the manual serves its purpose for the outside user. To serve as an internal handbook for data compilers within the health accounts team who need to be able to replicate all estimations, adding more detail would be necessary – in particular when it comes to methodology how health spending is distributed to the various functional categories.
References
[5] ANS (2025), Dados Gerais - Receita de contraprestações das operadoras (em Reais) (Brasil - 2015-2024), https://www.gov.br/ans/pt-br/acesso-a-informacao/perfil-do-setor/dados-gerais (accessed on 31 March 2025).
[4] DATASUS (n.d.), SIGTAP- Sistema de Gerenciamento da Tabela de Procedimentos, Medicamentos e OPM do SUS, http://sigtap.datasus.gov.br/tabela-unificada/app/sec/inicio.jsp (accessed on 4 March 2025).
[7] IBGE (2018), “Pesquisa de Orçamentos Familiares”, Tabela 6715 - Despesa monetária e não monetária média mensal familiar - valor e distribuição - por classes de rendimento total e variação patrimonial mensal familiar, segundo os tipos de despesa, https://sidra.ibge.gov.br/tabela/6715 (accessed on 22 March 2025).
[6] Ministério do Planejamento, Orçamento e Gestão (n.d.), Orçamento Federal, http://www.orcamentofederal.gov.br/orcamentos-anuais/orcamento-1999/Portaria_Ministerial_42_de_140499.pdf (accessed on 20 March 2025).
[1] OECD (2025), Best Practice in Institutionalising Health Accounts: Learning from Experiences in 13 OECD Countries, OECD Publishing, Paris, https://doi.org/10.1787/cf997130-en.
[2] OECD/WHO/Eurostat (2011), A System of Health Accounts: 2011 Edition, OECD Publishing, Paris, https://doi.org/10.1787/9789264116016-en.
[3] SIGA Brasil (n.d.), SIGA Brasil, https://www12.senado.leg.br/orcamento/sigabrasil (accessed on 4 March 2025).
← 1. These schemes predate the introduction of SUS in 1990, but for Federal civil servants they have been confirmed in Law 8112/1990.
← 2. Article 230 of Law 8112/1990.
← 3. Article 231 of Law 8112/1990.
← 4. In instances where the prime data sources to estimate overall health spending also includes sufficiently detailed on health services, the first and second step are essentially carried simultaneously.
← 5. The Ministry of Education is responsible for some of the funding of the federal university hospitals which form part of SUS.
← 6. Although some population groups, e.g. those living in institutions such as nursing homes are typically not covered.
← 7. The most recent POF was initiated in late 2024.
← 8. The Household Budget Survey (“POF”) 2024-2025 will allow for better estimates for the period between the two surveys, such as the year 2022. In other words, the data may be improved in future reports.
← 9. Provided to the OECD by end of January 2025.