Brazilian hospitals became more efficient in 2025. Margins fell anyway.

I spend a good part of my weeks talking with hospital CFOs. The conversation almost always starts with the clinical dashboard, and the numbers there look good. Then someone opens the financial results, and the mood changes. I have seen this scene in the South, the Southeast and the Northeast, at institutions of very different sizes. The question left on the table is always similar: if operations improved, why didn't the bottom line improve with them?

The Anahp Observatory 2026 answers that with considerable precision.

Start with operations. In 2025, across the 138 hospitals that feed Anahp's Hospital Indicators System, average length of stay fell to 3.76 days, the lowest level in the historical series. Bed turnover rose to 6.09. Bed turnover interval stood at 1.19 days. Anahp itself reads this set as one of the main signs of efficiency in bed use. And demand kept pace: the country closed the year with 53.18 million health plan members.

Now the results. In the same year, hospitals' average EBITDA margin fell to 11.05%, from 11.68% in 2024.

The math behind that gap lies in revenue and expense per patient-day. Adjusted for inflation, net revenue per patient-day rose 2.49% in 2025. Total expense per patient-day rose 10.88%. In real terms, expenses grew more than four times faster than revenue, precisely in the year hospitals started making better use of every bed they have.

It is worth noting how Anahp builds this net revenue figure. It already comes after taxes and after amounts denied and not received. The 2.49% is what was left. What was held back along the way shows up in another indicator: the initial denial rate from payers, which measures what was denied and is still under negotiation, stood at 15.93% in 2025. It was 11.89% in 2023. Over the same period, the collection rate fell from 91.27% to 85.80%.

It is not the cycle

The most common reading is that the sector is going through a rough patch and that the tide will eventually lift everyone. The data does not support that. In the first quarter of 2026, more than 80% of health plan operators posted positive results. Hospitals did not follow. Two links in the same chain, poring over the same medical bill, with results moving in opposite directions.

The explanation I see in practice has little to do with the economic cycle and a lot to do with processing capacity.

On the payer side, claims review has become an industrial process. Codified rules, automatic application, scale. Generating a denial costs little and happens in batches.

On the hospital side, defending that same claim is still handcrafted. One person reads the medical record. One person finds the authorization. One person builds the appeal. One claim at a time.

When one side automates the denial and the other responds by hand, the outcome stops depending on who is right. It starts depending on who processes faster. Much of what is lost to denials is not lost for lack of an argument. It is lost for lack of hands to sustain the argument before the deadline.

The deadline closes the account

In 2025, hospitals were paid by health plans in 77.35 days, about seven days longer than in 2024. And they paid their suppliers in 48.30 days. That is a 29-day gap someone has to finance, and that someone is the hospital. Add the 76.47% of gross revenue that comes from health plans, and working capital stops being a treasury topic and becomes a board agenda item.

That is why price adjustments alone do not solve the problem. An adjustment improves the price of a claim that is still reviewed by machines on one side and defended by people on the other. As long as that asymmetry exists, part of what is won at the negotiating table comes back as denials the following year.

What changes the game is hospitals coming to the table with equivalent processing capacity. In practice, that means treating clinical documentation as evidence from the very first entry, not as paperwork to be dug up after the denial arrives. It means knowing, on the same day, which claims are at risk and why. It means having the authorization trail retrievable in seconds, not weeks.

None of this depends on technology that does not exist yet. It depends on deciding that the revenue cycle deserves the same management rigor that clinical operations received over the past decade.

Brazilian hospitals have already proven they know how to operate better. The 2025 bed indicators show it clearly. Clinical efficiency has been achieved. It still has to reach the cash register in full.

Rafael Elustondo de Souza is Director of Go-To-Market at Rivio. Before that, he spent fifteen years in the medical conference industry and working with the pharmaceutical industry.

Originally published in Portuguese in Saúde Digital News.