A municipality can know how many older residents attended a health unit, how many families are registered with social assistance and how many places exist in local residential services, yet still struggle to answer a more important question: are older people maintaining function and receiving the support they actually need?
This is one of the defining evidence challenges within the Brazil Aging, Long-Term Care & Community Support Knowledge Hub. Brazil possesses extensive health, demographic and social-policy data, but long-term care crosses institutional boundaries that were not originally designed as one integrated care system. Information about disease, functional limitation, income, family circumstances, social assistance, home care and unpaid caregiving may therefore sit in different datasets governed by different parts of government.
The National Care Policy and its implementation through Brasil que Cuida make this increasingly important. Brazil is no longer dealing only with the question of how many older people exist or how many services operate. A care policy based on rights, shared responsibility and progressive implementation requires evidence about who needs care, who provides it, which needs remain unmet, whether territorial inequalities are narrowing and whether public intervention changes people's lives.
This creates a shift from administrative counting toward outcomes intelligence.
Brazil already has several foundations on which to build: the Instituto Brasileiro de Geografia e Estatística (IBGE), health information systems within SUS, e-SUS APS, multidimensional assessment through IVCF-20, SUAS information infrastructure, Censo SUAS, the modernization of Prontuário SUAS, and the emerging DataCuidados and Observatório Participativo dos Cuidados. The strategic task is to make these sources more coherent without erasing the different purposes for which they exist.
Brazil does not yet have one long-term care dataset because it does not yet have one long-term care system
The fragmentation of Brazilian care data reflects the architecture of Brazilian care itself.
SUS records health activity, clinical information, Primary Care contacts and other healthcare events. SUAS records social-assistance provision and family circumstances within its own statutory responsibilities. IBGE surveys reveal population characteristics, disability, functional limitation, household composition, work and unpaid activities. Social-security and income-support systems hold another layer of information. Residential institutions, private providers and families generate further information that may not enter any common national care record.
None of those datasets is inherently deficient because it does not describe the whole long-term care pathway. They were established for different purposes.
The problem arises when policymakers need to answer a cross-system question.
How many older people need assistance with activities of daily living? How many receive sufficient support? How much support is provided by relatives? What happens when a caregiver becomes unavailable? Which territories have day services or home support? Does functional decline lead to earlier Primary Care intervention or only become visible after hospitalization? How many families experience high-intensity care burden while receiving little formal assistance?
No single administrative system currently supplies all of these answers.
That is why data collection and data quality in Brazilian long-term care should not be reduced to building a larger database. The challenge is to define the questions that matter and identify which source, or combination of sources, can answer them reliably.
Population surveys reveal need that service records cannot see
Administrative data describe people who reach a service. Population surveys can reveal people who do not.
This distinction is particularly important in a care system where a substantial proportion of support remains within families.
The Pesquisa Nacional de Saúde (PNS) has provided important national evidence about functional limitation among older Brazilians. In the 2019 survey, 9.5% of people aged 60 and over were estimated to have difficulty with activities of daily living such as eating, bathing, dressing and moving around the home. Limitations in instrumental activities of daily living were substantially more common.
The significance of these measures goes beyond prevalence.
They show how need changes with age, gender, education, income and geography. They also expose a population whose long-term support needs cannot be inferred from diagnoses alone.
A health administrative dataset may record diabetes, hypertension and arthritis. A functional survey can reveal whether those conditions have translated into dependence in everyday life.
This distinction should influence planning. Long-term care demand is more closely related to combinations of function, cognition, social circumstances and available support than to chronological age alone.
Population evidence also helps prevent a common analytical error: treating service utilization as an estimate of need. Low use can mean low need, but it can also indicate unavailable services, financial constraints, transport barriers, lack of information or reliance on unpaid relatives.
For Brazil, measuring population need therefore requires both administrative and household evidence.
Functional outcomes are becoming more visible within Primary Care
The incorporation of the Índice de Vulnerabilidade Clínico-Funcional-20 into the electronic record used within e-SUS APS is an important development for older-person evidence.
IVCF-20 is designed to identify clinical-functional vulnerability rapidly. Its value is not simply that another assessment can be completed electronically. The Ministry of Health explicitly identifies longitudinal use as a way to support follow-up, evaluate interventions, improve team discussion, support coordination across the Rede de Atenção à Saúde and strengthen monitoring of the Política Nacional de Saúde da Pessoa Idosa.
This creates the possibility of observing change rather than recording only state.
An older person can move from relative independence to emerging vulnerability before becoming dependent on extensive support. Repeated functional information can help services identify that trajectory earlier.
For long-term care, this may ultimately be more useful than measuring whether the person received a specific consultation.
Good outcome intelligence would distinguish among several questions:
- Did functional ability remain stable?
- Was deterioration recognized early?
- Did intervention restore any lost capacity?
- Was preventable decline reduced?
- Did the person remain safely at home where that was their preference?
- Did increasing dependency lead to an appropriate change in support?
No single score should determine these judgments. Function is influenced by illness, environment, rehabilitation, housing, caregiver availability and personal goals. But longitudinal functional information provides a much stronger basis for aging policy than service counts alone.
Operational scenario: the same service volume produces different outcomes
Two municipalities report broadly similar numbers of Primary Care contacts with residents aged over 75. At first sight, their older-person provision appears comparable.
A deeper review tells a different story.
In the first municipality, teams use multidimensional assessment consistently and repeat functional assessment for higher-risk older people. The data show that people identified with emerging vulnerability are more likely to receive multidisciplinary review and planned follow-up.
In the second municipality, activity levels are high but functional information is recorded inconsistently. Older residents often appear in performance data only through diagnoses, consultations and acute episodes.
Both municipalities can demonstrate activity. Only one can begin to understand whether intervention is changing trajectories.
This does not prove that the first municipality provides better care. Outcome analysis would still need to account for population differences, socioeconomic context, service capacity and case mix. But it allows a more useful governance discussion: not simply how much care was delivered, but whether needs were identified, responses occurred and outcomes changed.
This is the difference between administrative reporting and performance intelligence.
Organizations developing comparable approaches can use the Quality Dashboard Builder to structure balanced measures of access, quality, function, safety and outcomes. It is not a Brazilian government reporting instrument, but the underlying discipline of combining activity and outcome measures is directly relevant.
SUS data are essential but cannot describe long-term care on their own
Brazil's public health system generates enormous volumes of information. These data are indispensable for understanding older-person health, disease burden, Primary Care activity, hospital use, vaccination, home healthcare and other parts of the care pathway.
Yet long-term care is not synonymous with healthcare.
A person may be medically stable while needing help with bathing, meals, mobility, supervision or community participation. Another may have substantial caregiver support and therefore use relatively few formal services despite significant dependence.
Health data can consequently underestimate social-care need if interpreted without functional and household context.
The opposite problem also occurs. A social-assistance record may identify family vulnerability but not contain the clinical detail needed to understand why an older person's function has changed.
This is why stronger coordination across health and social care requires information relationships rather than wholesale merger of records.
SUS and SUAS perform distinct roles. Their data should retain appropriate purpose, professional confidentiality and governance. Integration should focus on the minimum relevant information required to coordinate support, identify risk and understand population needs.
SUAS provides a different view of care need
The Sistema Único de Assistência Social sees dimensions of older people's lives that clinical systems may not.
CRAS and CREAS services work with families experiencing different forms of vulnerability and social risk. Specialized provision can include Centro-Dia and support for dependent older people and their families. Basic Social Protection may identify isolation, fragile family networks and difficulties accessing rights before those circumstances become visible within healthcare.
Censo SUAS provides national information about the structure of the social-assistance system, including CRAS, CREAS, Centros de Convivência, Centros-Dia, institutional reception services, management arrangements and workforce. The 2026 Censo SUAS cycle is being prepared across these service types, continuing an established national mechanism for observing SUAS capacity.
The strength of Censo SUAS is structural visibility. It can help answer where services exist, what type of units operate, what workforce and provision are available, and how the system is organized territorially.
Its limitation for long-term care is equally important: the existence of a service does not establish that all older people needing it can access it or that the service has changed outcomes.
A municipality can report a Centro-Dia without demonstrating whether demand exceeds capacity, whether caregiver strain has fallen or whether admission to institutional care has been delayed.
Those questions require outcome measures linked to the purpose of the service.
Prontuário SUAS can strengthen continuity while raising legitimate privacy questions
Prontuário SUAS provides another layer of evidence. Its purpose is to support professional recording, continuity and planning in work with individuals and families receiving social-assistance services.
The information can include circumstances directly relevant to long-term care, including whether a family member requires constant care because of aging or illness and who is responsible for providing that support.
In 2025, the Comissão Intergestores Tripartite issued updated national directions for the electronic Prontuário SUAS. The framework reinforces the record as a right of people and families using SUAS, recognizes the presence of personal and potentially sensitive data and places responsibilities around confidentiality, security and appropriate professional use within the public unit and reference team.
This is strategically important as Brazil develops more integrated care policy.
There will be understandable pressure to connect information. But greater integration should not result in unrestricted visibility across government.
A daughter's financial stress, an older person's health condition, family conflict and a safeguarding concern may all be relevant to care planning, yet not every professional in every system requires access to every detail.
Effective data governance and information accountability therefore depend on purpose, proportionality and role-based access.
The objective is enough information to support continuity without turning vulnerability into unnecessarily shareable data.
The National Care Policy changes what Brazil needs to measure
Brasil que Cuida creates a broader evidence requirement because it defines care as a social, economic and rights issue rather than only a service activity.
Axis 5 of the National Care Plan concerns governance and management. Its strategies specifically include intersectoral, federative and participatory monitoring and evaluation; improvement of official statistics and administrative records; and further research.
This is an important policy signal.
Brazil's emerging care framework recognizes that inadequate evidence is itself an implementation problem. If unpaid care remains statistically invisible, policy can underestimate its economic and gender consequences. If administrative systems record provision but not unmet need, expansion decisions may follow existing infrastructure rather than population need. If national monitoring lacks territorial detail, improvements in major cities can hide weak coverage elsewhere.
The care policy therefore needs an evidence architecture capable of showing at least four dimensions:
- need — who requires care and at what level of functional support;
- supply — what formal and informal support is actually available;
- experience — whether people and caregivers can navigate and use that support;
- outcomes — whether care protects rights, function, wellbeing and sustainable family life.
These dimensions are related but cannot be substituted for one another.
DataCuidados represents a significant move toward a care-specific evidence layer
The creation of the Observatório Participativo dos Cuidados and its DataCuidados platform is therefore particularly significant.
DataCuidados brings together indicators relating to the organization of care in Brazil using official sources including the Demographic Census, PNAD Contínua, DATASUS and administrative records. It is designed for public access and periodic review.
This matters because care cannot be understood from one ministry's records alone.
A care observatory can combine information about population structure, work, unpaid activity, social inequality and public services while preserving the provenance of the underlying datasets.
The participatory element also matters. The Observatório is intended not merely as a statistical portal but as a space connecting government, civil society and academia, supporting accessible information, public participation and social control.
That creates a different accountability model from internal management reporting.
Evidence becomes something citizens can interrogate.
The strongest future development would be for DataCuidados to make territorial inequalities increasingly visible without creating false precision where source data are incomplete. A national estimate may support policy strategy; a municipal manager may need much more granular evidence to decide where a day service, home-care team or caregiver-support intervention is required.
Operational scenario: a municipality discovers that demand is not where existing services are
A municipality begins planning implementation of its territorial care strategy under Brasil que Cuida. Existing provision has developed incrementally over many years, and local leaders assume that neighborhoods with the greatest service activity represent areas of highest need.
A combined needs assessment challenges that assumption.
IBGE demographic information identifies neighborhoods with rapidly increasing numbers of older residents. Primary Care data show substantial clinical-functional vulnerability in some of those areas. SUAS information identifies concentrations of low-income households and families supporting dependent relatives. Existing service-use data, however, remain relatively low.
The apparent contradiction is operationally important.
Low use in those neighborhoods may not indicate low demand. Transport, lack of services, insufficient information or reliance on relatives may be suppressing access.
The municipality does not immediately conclude that a new facility is required. It undertakes further local engagement, examines waiting and referral patterns and speaks with older residents and caregivers.
The eventual response combines outreach, better referral coordination and expansion of community support rather than simply enlarging the busiest existing service.
Data have changed the planning question from "where are services already used?" to "where is support most needed?"
That is the practical value of data-led equity planning.
Caregiver outcomes need to become part of long-term care performance
Brazil cannot measure long-term care effectively while treating unpaid caregivers as an invisible input.
Families provide an enormous share of daily assistance. Women carry a disproportionate part of that work, and the National Care Policy explicitly seeks to recognize and redistribute care responsibility.
This means the outcome of an intervention cannot be judged only from the perspective of the person receiving care.
An older man may remain at home, apparently demonstrating successful community living, while his daughter leaves employment because she provides continuous unpaid supervision. The placement outcome may be positive for him and economically damaging for her.
Similarly, a day service may appear relatively modest in clinical terms but create substantial value if it allows a family caregiver to continue paid work or obtain predictable periods of rest.
Measures of caregiver support and family navigation therefore need to include burden, sustainability, access to respite, employment impact and confidence in managing care.
This is not an argument for making families accountable for service outcomes. It is the opposite. It recognizes caregiver capacity as an outcome that public policy can either protect or exhaust.
Outcome measurement must respect individual goals
Standardization is necessary for population monitoring, but long-term care remains deeply personal.
Two people with similar levels of functional impairment may define a successful outcome differently. One may prioritize remaining in her own home. Another may value greater social contact. A person living with dementia may value familiar routines and relationships more than improvement on a narrow clinical indicator.
Good measurement therefore needs both standardized and person-centered evidence.
Standard measures allow comparison, stratification and planning. Individual goals show whether support matters to the person receiving it.
This principle is particularly important when care systems begin using performance dashboards. Indicators that are easy to count can gradually displace outcomes that are difficult to quantify.
For example, the number of home visits can be measured immediately. Whether those visits allowed someone to retain confidence in bathing independently may require qualitative or functional evidence.
Organizations working through similar questions can use the Community Impact Report Builder to structure a mixture of quantitative outcomes and human evidence. It does not replace Brazilian public reporting requirements, but it illustrates why service impact should be demonstrated through more than activity counts.
Equity requires disaggregated evidence
National averages can conceal very different experiences across Brazil.
Age, gender, race, income, disability, education, geography and access to family support can all affect care need and service access. Regional differences are substantial, while rural and Amazonian communities face distinct access conditions from metropolitan populations.
Outcome data therefore need disaggregation wherever statistical quality and privacy permit.
A national increase in service coverage may coexist with widening inequality if growth occurs primarily in better-resourced territories. Improved average functional outcomes may hide poorer results for low-income populations. High digital registration can obscure exclusion among older people unable to use online services.
This does not mean every indicator should be fragmented into dozens of categories. Small numbers can become statistically unstable and create privacy risks.
The governing principle is material inequality: disaggregate where doing so reveals a meaningful difference that could influence policy or resource allocation.
Brazil's care policy explicitly recognizes intersecting inequalities of gender, race, ethnicity, income, disability and territory. Its evidence framework therefore needs to be capable of testing whether implementation is reducing them.
Operational scenario: a successful average hides a poorer outcome for rural older people
A state reviews a new older-person support initiative and finds that follow-up rates have improved significantly. The headline indicator suggests successful implementation.
When analysts examine the data geographically, the improvement is concentrated in larger municipalities. Smaller rural areas show slower referral completion and substantially longer journeys to specialist services.
The overall average is technically accurate but operationally incomplete.
The state examines whether telehealth, regional support arrangements and stronger Primary Care pathways could reduce unnecessary travel. It also reviews whether municipalities need technical support to improve local information recording.
Importantly, leaders do not assume that poor recorded performance reflects poor professional practice. Some municipalities have weaker connectivity and smaller teams, while others face genuine service-capacity constraints.
The data therefore trigger inquiry rather than blame.
This is a critical feature of mature performance systems. Indicators should identify variation requiring explanation; they should not automatically explain why that variation exists.
Where persistent differences remain after adjusting for relevant context, the evidence can then support targeted investment, redesigned pathways or stronger regional coordination.
Measurement can become harmful if targets replace judgment
Any system that links performance to visible indicators creates behavioral incentives.
This is especially relevant in Brazilian Primary Care, where national cofinancing increasingly includes performance measures relating to older people. Current indicators include professional consultations, anthropometric assessment, Community Health Agent or technician home visits and influenza vaccination.
These measures can encourage valuable preventive activity. They also illustrate why indicator design must be handled carefully.
If performance attention becomes too narrowly focused on specified actions, services may optimize what is measured while neglecting less visible needs.
An older person could technically satisfy several indicators while experiencing worsening cognition, caregiver breakdown or escalating social isolation.
Good performance frameworks therefore combine process measures with outcome and balancing measures.
A process measure asks whether an assessment occurred. An outcome measure asks whether the person's health or function improved or stabilized. A balancing measure asks whether the intervention created an unintended consequence elsewhere.
Organizations seeking to strengthen this approach can use the Quality Improvement Action Plan Builder to translate performance findings into structured improvement actions. The value lies not in treating every variance as non-compliance, but in connecting evidence to investigation, ownership, action and review.
Longitudinal evidence matters because aging is a trajectory
Cross-sectional measures tell policymakers what is happening at a moment in time. Long-term care is fundamentally longitudinal.
A person's needs develop over months and years. Function may decline gradually, recover after rehabilitation, deteriorate after hospitalization or change suddenly when a caregiver dies.
A mature evidence system therefore needs to understand trajectories.
This is one reason longitudinal Primary Care records and repeated IVCF-20 assessments are potentially valuable. They can help distinguish between a person who has remained stable and someone whose vulnerability is increasing even though both currently receive similar service intensity.
At population level, longitudinal evidence can also reveal whether earlier intervention changes later demand.
If stronger community support is associated with fewer avoidable hospital admissions or delayed institutionalization, those effects may take time to become visible. A program evaluated only during its first months could therefore underestimate impact.
Conversely, short-term improvements should not automatically be projected indefinitely.
Long-term care evaluation needs realistic time horizons matched to the outcome being measured.
Linking datasets creates analytical power and governance risk
Brazil's ability to understand care will improve as datasets become more interoperable and analytical methods become more sophisticated.
Linkage can reveal patterns that separate systems cannot. Health events might be examined alongside socioeconomic circumstances, functional vulnerability and territorial service supply. Researchers could potentially explore whether particular pathways are associated with better outcomes.
But linkage also increases sensitivity.
A combined dataset describing someone's health, income, disability, family relationships and service use can reveal far more than any single record.
This means stronger analytics require stronger governance.
LGPD principles, professional confidentiality, access controls and clear purposes remain essential. Identifiable information needed for direct care should not automatically be reused for every analytical purpose. Population planning can often use de-identified or aggregated information.
The key distinction is between data availability and legitimate data use.
Brazil can become more evidence-driven without assuming that every useful analytical possibility should automatically be pursued.
Artificial intelligence will increase the importance of evidence quality
As Brazil expands digital health and data infrastructure, artificial intelligence and predictive analytics are likely to play a growing role in identifying risk and prioritizing care.
That makes foundational data quality even more important.
An algorithm trained predominantly on people already reaching formal services may underestimate hidden need among populations with poor access. Incomplete functional records can create misleading predictions. Historical patterns of unequal provision can be reproduced as apparently objective recommendations.
AI therefore does not solve weak evidence. It can amplify it.
Before predictive tools are used to influence high-impact long-term care decisions, authorities need to know what populations are represented, what outcomes the model predicts, how accuracy varies across groups and what professional review surrounds its use.
The strongest future application may be decision support rather than automated allocation: identifying possible deterioration, highlighting people lost to follow-up or revealing patterns requiring professional attention.
Human accountability should remain visible.
Brazil needs an outcomes framework that connects people, services and systems
A national long-term care outcomes framework does not need to reduce the diversity of Brazilian services to one score.
It should instead create a common line of sight between national policy and local experience.
At person level, measures could consider function, autonomy, safety, participation, continuity and experience. At caregiver level, burden and sustainability matter. At service level, access, responsiveness, workforce capability and quality are important. At system level, equity, avoidable acute-care use, territorial coverage and financing sustainability become relevant.
The framework should also distinguish what each actor can reasonably influence.
A municipal Centro-Dia cannot control national poverty rates. A Primary Care team cannot compensate for the absence of every specialist service. A federal ministry cannot directly determine the quality of every daily interaction.
Accountability becomes more credible when indicators align with responsibility.
This is central to outcomes frameworks and indicators: measures should inform decisions at the level where action is possible.
Participatory evidence can expose what administrative systems overlook
The participatory ambition behind the Observatório dos Cuidados is therefore more than an engagement exercise.
People using care and those providing unpaid care possess information that administrative systems often cannot capture.
A dataset may show that a service was accessed. A caregiver may explain that arranging transport took four hours. A record may show that a home visit occurred. An older person may explain that different professionals repeatedly asked the same questions because information had not followed them.
Qualitative evidence helps explain the mechanism behind quantitative outcomes.
This does not mean policy should be based on individual anecdotes. It means experience should be systematically gathered, analyzed and considered alongside statistical evidence.
Participation is particularly valuable when indicators look positive but lived experience suggests otherwise.
It can reveal hidden waiting, administrative burden, fear of services, cultural barriers or support being provided informally because formal pathways are unavailable.
For a care policy explicitly founded on rights, social participation is therefore part of evidence quality.
Governance should turn data into learning rather than another reporting burden
Brazil already asks frontline organizations and public bodies to record substantial amounts of information. The answer to every evidence gap cannot be another form.
Poorly designed reporting can consume professional time while producing data that nobody uses.
Stronger governance begins by identifying which decisions the information is expected to support.
Municipal teams need actionable local intelligence. States need to identify territorial variation and coordinate capacity. Federal ministries need evidence about implementation, inequality and national policy outcomes. Citizens need understandable public information. Researchers need sufficiently robust data to evaluate what works.
The same measure will not always serve all five purposes.
A mature evidence cycle therefore asks:
- What decision will this information inform?
- Who is responsible for interpreting it?
- What variation requires escalation?
- What additional context is needed before judgment?
- What action follows the finding?
- How will the system know whether the action worked?
Without this cycle, dashboards become repositories rather than governance tools.
The future evidence challenge is measuring whether Brazil is actually redistributing care
Brasil que Cuida creates a particularly demanding evaluation question.
Its ambition is not only to expand isolated services. The policy seeks to recognize care as a right and as work, reduce unequal responsibility within families, strengthen public provision, improve paid-care employment and build greater social co-responsibility.
Those objectives require measures beyond conventional service activity.
Brazil will need to know whether unpaid care hours are changing, whether women's employment is being protected, whether access to day and home-based services is expanding, whether territorial disparities are narrowing and whether people requiring support experience greater autonomy and continuity.
Some outcomes will take years to demonstrate.
Others may conflict. Expanded formal services may initially increase recorded demand because previously hidden need becomes visible. A rise in service use should not automatically be interpreted as deterioration in population wellbeing.
This is why implementation evidence needs interpretation alongside outcome trends.
A national care system in development should expect better measurement to reveal unmet need before it demonstrates that unmet need has fallen.
International learning: count what the system is trying to achieve
Brazil's experience is relevant internationally because many countries face the same mismatch between extensive administrative data and limited understanding of long-term care outcomes.
The institutional arrangements differ. Some countries operate dedicated long-term care insurance systems; others finance provision through municipalities, taxation or mixed public-private structures. Brazil's combination of SUS, SUAS, family provision and an emerging National Care Policy cannot simply be transferred elsewhere.
The transferable lesson lies in measurement design.
Countries often know how many services they purchased long before they know whether those services maintained independence, protected caregivers or reduced inequality.
Brazil's developing approach suggests the value of combining household surveys, functional assessment, service records, territorial capacity data, public dashboards and participatory evidence.
No single source provides the truth. Each illuminates a different part of the system.
The strongest evidence architecture is therefore one capable of connecting those perspectives while preserving their methodological limits and governance boundaries.
Conclusion
Brazil's long-term care evidence challenge is not a shortage of data. It is the need to turn multiple forms of information into a coherent understanding of care.
IBGE surveys reveal population need and unpaid support that service records cannot see. SUS data describe healthcare pathways. e-SUS APS and IVCF-20 create opportunities to monitor functional trajectories. Censo SUAS and Prontuário SUAS illuminate social-assistance infrastructure and family circumstances. DataCuidados and the Observatório Participativo dos Cuidados are beginning to create a more explicit national evidence layer around the organization of care itself.
The next step is to connect measurement more closely with purpose. Brazil needs to know not only how many consultations, home visits, facilities or benefits exist, but whether older people maintain autonomy, whether deterioration is identified early, whether caregivers can sustain their own lives, whether territorial inequalities are narrowing and whether public investment changes outcomes.
That requires careful governance. More integrated data must remain proportionate, secure and intelligible. Performance measures must encourage learning rather than target compliance alone. National averages must not obscure regional inequality. Quantitative evidence must be complemented by the experience of people and families.
If Brazil can build that evidence architecture alongside its emerging National Care Policy, measurement can become more than a retrospective reporting function. It can become one of the mechanisms through which a fragmented landscape of health, social assistance, community support and unpaid care gradually becomes a more accountable national care system.