## Notes from 27 February 2026 [[2026-02-26|← Previous note]] ┃ [[2026-02-28|Next note →]] [[André Medici]] is one of the sharpest commentators on public health policy in Brazil. A health economist with decades of experience, he writes the _Monitor de Saúde_ blog, where he regularly produces the kind of rigorous, politically unafraid analysis that Brazilian health policy rarely gets from inside the sector. Early this month, Medici published a detailed institutional archaeology of the [Previne Brasil program (2019–2023)]([https://media.graphassets.com/EqiiIT0yRqmhNFwBzIva](https://www.linkedin.com/pulse/heran%C3%A7a-do-previne-na-aten%C3%A7%C3%A3o-primaria-sa%C3%BAde-brasil-medici-andre-szbre/?trackingId=me8cQ3K5nLmAQiYf2BRZHA%3D%3D)). The piece traces the program's origins, design, implementation, and partial survival through the 2024 restructuring, making a careful case that Previne's core innovations were not destroyed by the incoming Lula administration but repackaged into a new, if more complex, financing architecture. It is both a defense of the program's legacy and an honest accounting of its limits. ## The piece Medici's text is, at its core, an act of institutional archaeology. The Previne Brasil program (2019–2023) was the deepest structural reform that primary health care (in Portuguese, APS) financing underwent since the creation of Brazil's SUS (_[[Public Healthcare in Brazil (SUS)|Sistema Único de Saúde]]_), the country's universal public health system established by the 1988 Constitution. For context, Brazil is a highly decentralized federation where municipalities are the primary providers of primary care, but the federal government co-finances APS through transfers — making the design of those transfers a powerful lever over how care is actually organized on the ground. [Previne replaced the inertial logic of the financing framework it replaced](https://media.graphassets.com/EqiiIT0yRqmhNFwBzIva), with a model built on weighted capitation tied to performance-based payments linked to tracer indicators and strategic incentives. For the first time, APS financing explicitly stated that money should follow minimally observable results. The counterfactual matters here. Bolsonaro administration (2019–2022) has its health policy overwhelmingly narrated through the lens of the pandemic — the denialism, the vaccine delays, the institutional chaos at the Ministry of Health. That narrative is accurate but incomplete. Had the pandemic not happened, the health policy legacy of that period might look very different, because Previne was designed and launched in November 2019, before COVID-19 arrived. ## Previne as part of the "Index as Lever" family The most interesting analytical move is to read Previne not as an isolated health sector reform but as part of a broader trend: the use of metrics and formulas to induce behavior in intergovernmental transfers in Brazil. Three federal instruments now share this DNA, even if they differ in institutional architecture: - The "[famous](https://open.substack.com/pub/rainerkattel/p/week-5-structural-capacity-the-hidden?r=10o1&selection=681adb23-fd9c-4189-95fa-1f5dcfdaccc0&utm_campaign=post-share-selection&utm_medium=web&aspectRatio=instagram&textColor=%23ffffff&bgImage=true)" **IGD** (_Índice de Gestão Descentralizada_) in [[Programa Bolsa Família|Bolsa Família]] (Brazil's flagship conditional cash transfer program, one of the largest in the world) measures municipal management results: cadastral maintenance of poor families, routines, coordination. It converts those results into financial transfers that sustain the decentralized administrative machinery. It buys _means_: administrative and informational capacity to operate a targeted program. Since the main benefit flow goes directly to families (not through municipal budgets), the IGD exists as a separate instrument to keep the local management engine running and reward those who do it well. - The [**VAAR** (_Valor Aluno Ano Resultado_)](https://todospelaeducacao.org.br/wordpress/wp-content/uploads/2025/08/nota-tecnica-sobre-vaar.pdf) in the Fundeb (Brazil's main intergovernmental fund for basic education financing) is a slice of the Union's complementary funding distributed by results. It buys _ends_: improvements in learning outcomes and equity reduction. **Previne** did something more ambitious: it tried to become the _regime itself_. Not an add-on or a slice, but the core architecture of APS co-financing, with capitation, performance, and strategic incentives as its three pillars. Inside that architecture, the performance component functioned exactly like IGD/VAAR (indicator becomes money!) but the scope was larger. Previne bought a _mix_: it bought means (enrollment, territorial linkage, data registration) and proxies of ends (traceable delivery in priority care lines like prenatal care, chronic disease management, vaccination). This distinction matters. IGD and VAAR are levers attached to larger policies. Previne was, for a period, the payment regime itself. The analogy with IGD/VAAR holds when you look at the component where indicator turns into money, but breaks down when you consider that Previne put the formula at the heart of APS financing rather than at its margin. The clinical object is also more granular — individual-level enrollment and care delivery, not aggregate network results or administrative routines. And the fiscal risk for municipalities was more immediate: variation in enrollment or production could shift monthly cash flows, whereas VAAR operates annually and IGD finances management rather than the main benefit. Still, the family resemblance is real and underappreciated. Before Previne, APS financing was closer to a per-capita-plus-adhesion model — you got money for having teams, not for what they did. After Previne, APS joined the club of Brazilian intergovernmental transfer mechanisms [[Digital Public Infrastructure (DPI)|where data management is central]], measurable performance has financial consequences, and the Union "purchases" something specific through formula rather than distributing resources by inertia. In a federation with over 5,500 municipalities of wildly varying capacity that shift is not trivial. ## What survived the transition Medici is careful to show that the 2024 restructuring did not destroy Previne. The [[Lula Administration|Lula's third administration]] repackaged it, but the main DNA survived: measurement, linkage, and quality induction became institutionalized parts of how APS is financed and evaluated. This is worth noting for the broader record of Brazilian policy continuity. Lula (center-left) returned to the presidency in January 2023 after defeating Bolsonaro (right-wing), and the rhetorical frame of "reconstruction" that characterized the new administration suggested a clean break with predecessor policies. In practice, what happened in APS was more calibrated: the new government recognized that transferring financial risk to municipalities via performance had outrun local capacity in many contexts, especially in vulnerable territories, but it did not abandon the principle that financing should induce measurable quality. It rebalanced the mix between predictability and performance rather than reverting to the pre-Previne inertial model. Medici does flag real losses. The original Previne was simpler, more legible to municipal managers, and more directly tied to clinical delivery. The 2024 methodology introduces multiple sub-components, intermediate indices, and transition rules that weaken the causal link between managerial action and financial result. National comparability suffers. The performance signal gets diluted among administrative and structural criteria. The risk, in Medici's reading, is that the system shifts from active management of financing toward reactive compliance with federal calculations — a bureaucratization of what was designed as a clinical incentive. But the structural point holds: the 2024 reform was possible _because_ Previne exposed both the limits of the previous model and its own internal limits. It taught the system how to measure, link, and compare. ## The Previne results window (2022–2024) Medici presents data showing clear aggregate improvement in APS delivery capacity during the period when performance-based payment was fully operational. The Synthetic Final Indicator (ISF), which consolidates seven tracer indicators (prenatal access, syphilis/HIV testing, dental care in pregnancy, cervical screening, childhood vaccination, and chronic disease follow-up for hypertension and diabetes) into a 0–10 score, shows a sharp upward trajectory. The average ISF for the 100 large municipalities with the lowest per-capita revenue rose from 3.18 in 2020 to 8.33 in 2023. The analytical caution is appropriate: improvement in "monitored delivery capacity" (teams producing, recording, and demonstrating delivery) is not identical to improvement in final health outcomes. The system got better at measuring and proving care faster than at transforming clinical endpoints.