Breathing is one of the first adaptations a baby makes shortly after birth. In some newborns, however, this process may be accompanied by difficulties, with signs that the baby needs to exert more effort to breathe. This is known as neonatal respiratory distress, a condition that can have various causes, is frequently associated with prematurity, and, in some cases, may require hospitalization and intensive care.
In this edition of “Interview with Authors,” Everton Nunes da Silva, Associate Editor of Cadernos de Saúde Pública (CSP), speaks with Julia Raffin, a co-author of the study, about the hospital costs of neonatal respiratory distress for the Unified Health System (SUS). The interview addresses the frequency of hospitalizations, the use of Intensive Care Units (ICUs), regional inequalities, and the implications of the findings for health management and economic evaluation.
Everton Nunes:
What is neonatal respiratory distress? You found a relatively high rate of 15.7% among neonatal hospitalizations. We’d like to know if this is specific to Brazil and if this rate tends to be high in other contexts as well?
Julia Raffin:
Neonatal respiratory distress is a condition characterized by increased respiratory effort in the newborn. This can be recognized through signs and symptoms such as flaring of the nostrils, grunting, and decreased arterial oxygen saturation. These are signs and symptoms that indicate increased respiratory effort, which can have various causes. To explore this further, we need to discuss specific diagnoses. For example, transient tachypnea is a form of neonatal respiratory distress resulting from delayed reabsorption of pulmonary fluid after birth, which impairs breathing and gas exchange. Another specific diagnosis, such as neonatal respiratory distress syndrome—formerly known as hyaline membrane disease—occurs when the baby is premature; the baby’s lungs are immature, and the baby is also unable to produce enough surfactant to facilitate adequate gas exchange. Consequently, the baby experiences respiratory distress resulting from this primary complication.
Respiratory distress, then, is extremely significant and carries a very heavy burden, not only in Brazil but worldwide, because it is primarily associated with prematurity. So, when we look at Brazil, we see a prematurity rate of around 11%. So, if we consider the nearly 3 million live births in Brazil each year, this means, on average, 320,000 live births per year are at risk of developing some form of neonatal respiratory distress. Globally, this rate varies, but studies indicate that, in pediatric hospitalizations, respiratory distress and respiratory diseases in newborns are among the leading conditions in terms of both the economic burden of these hospitalizations and their prevalence. So, they’re not only costly for the healthcare system; they’re also highly prevalent, and there are many hospitalizations resulting from this condition.
Everton Nunes:
One finding that drew a lot of attention in your study is that half of the hospitalizations for neonatal respiratory distress were classified as nonspecific according to ICD-10 coding. Is this also expected to be in this condition?
Julia Raffin:
It’s not expected, because we have well-defined classifications regarding what the diagnoses are, the signs, and what defines a specific diagnosis. To explain it a little better, in the International Classification of Diseases (ICD), there are specific ICD codes that, for example, include this one: transient tachypnea of the newborn. And there are unspecified diagnoses, where you’re not yet certain, or the case is still under investigation, and you classify the condition as a symptom—because you know it’s a symptom—but you haven’t yet pinpointed the exact cause or specific diagnosis.
So, we expect the diagnoses to be very well defined. However, we have been observing—not only in this study we just conducted but also in other studies by our group—that unspecified, undefined diagnoses have a higher prevalence. So, typically, most of the hospital records we’ve been analyzing consist of undefined, unspecified diagnoses. We even had the recent study by Emilie Freire Batista, one of the co-authors, on cerebral palsy. Cerebral palsy has a very well-defined clinical presentation, so it would be unusual to have an undefined diagnosis. This occurred in her study and now also in ours, with respiratory distress, which raised concerns among us. So much so that we sought to speak with some neonatologists to determine why this has been happening. Among the many hypotheses we can raise, this also opens the door to questions for future studies. The people we contacted to identify the cause of this said that when babies are admitted to the hospital, a mandatory diagnosis is recorded in the admission record. However, at that time, no adequate or timely investigation was conducted to determine the specific diagnosis for that baby. So, they list an unspecified diagnosis as a placeholder until clinical investigation can be conducted later. Upon the baby’s discharge, there is no requirement to provide a definitive, closed diagnosis, nor to change that diagnosis. Consequently, what usually happens is that the initial diagnosis remains, and we are left with this limitation. So, it’s a matter of the quality of our records: we’ve been finding that the majority are unspecified and that, unfortunately, this limits our ability to provide a slightly more refined analysis of which health conditions are costing us the most or are most prevalent; they remain in the realm of the “undefined.”
Everton Nunes:
When it comes to real-world data, right? We don’t always have that specific focus in our research, and sometimes we end up losing that later when we try to conduct more specific analyses. So, taking these issues into account, Julia, I’d like to know what the main findings were, since your article focused specifically on the analysis of direct hospital costs. What did you find regarding this health condition?
Julia Raffin:
Quite simply and directly, our findings are strongly focused on the public benefit this will bring. So, regarding the cost, what did we find? Over the four-year period we analyzed, 1.4 billion reais were spent on hospital admissions for neonatal respiratory distress. And we saw an alarming statistic: 70% of that cost goes toward admissions to Intensive Care Units (ICUs). Half of all hospitalizations are in the ICU, yet more than half of the costs are for ICU stays.
There is also the issue of regional disparities. We observed, for example, that in the South and Southeast, there are more hospitalizations, greater use of ICUs, lower mortality rates, and a higher standardized investment per 1,000 live births in these regions. Meanwhile, in the North, ICU use is lower, mortality is higher, and this cost per 1,000 live births is lower. We see that there is a very large discrepancy between the country’s regions in terms of what is being spent. In the North, an average of 112,000 reais is invested per 1,000 live births, while in the South, 770,000 reais are invested per 1,000 live births.
So, we looked at how much is spent on hospital admissions for neonatal respiratory distress, in addition to the regional discrepancies we presented in the article.
Everton Nunes:
So, if I understand correctly, you’re saying that the structure and distribution of neonatal ICU capacity will also influence the cost structure, right? The greater the neonatal ICU capacity, the higher the costs will be, since you’re providing this service to this specific population. Is that right, Julia?
Julia Raffin:
We can’t establish causality in our type of study, but the observations we’ve been making—not only in this study but also in others—are that where there is a greater number of ICU beds, ICU use is higher, and higher costs are associated with that. This is partly because ICU admissions are more expensive due to the technologies used. However, we must be very careful when interpreting these results, because using more and spending more are not always synonymous with improved healthcare quality or patient satisfaction and improved population health. So, these are two different things that we need to analyze with some caution.
Everton Nunes:
And does this use of ICUs apply to all these classifications of neonatal respiratory distress, or do some classifications account for a higher proportion, based on the data you collected?
Julia Raffin:
There is a difference in ICU utilization across diagnoses and, more importantly, a temporal variation in this proportion. We even found it somewhat curious that neonatal respiratory distress syndrome—which is the diagnosis related to prematurity—has been leading to a decrease in ICU use, while other diagnoses have been increasing over time. So, this means that diagnoses that aren’t clinically well-defined and those considered clinically somewhat milder and simpler are seeing an increase in ICU admissions, while more complex diagnoses—which require a bit more technology and care—are decreasing. In our study, we can’t explain or understand why this is happening. However, based on what has been published in the literature, we propose some hypotheses: perhaps clinical management is changing, or—specifically regarding acute respiratory distress syndrome—we have made further progress in prevention and in the technologies used, which may reduce the severity and progression of these cases within hospitals.
Everton Nunes:
Very interesting. Even though we’re limited by the DATASUS data, at least we can get a broader picture of what the pattern looks like and how to track it over time. So, in terms of perspectives or implications for management, based on the data you’ve gathered, what would you recommend if you had to craft a specific message for managers involved in this area of neonatal care? What could we suggest to them?
Julia Raffin:
We always focus on preventable conditions. For everything we can prevent, we need to take a more sensitive approach and always prioritize health equity. Whether we like it or not, we work here in public health; we are part of the SUS, so we must always prioritize equity and prevention. Therefore, we consistently convey the message to carefully consider not only the cost—the raw numbers—but also where demand is highest and where supply is insufficient, seeking strategies for prevention and improving service delivery where it’s needed. So, the Alyne Network, for example, has been, since 2024, addressing these issues of promoting greater equity in maternal and child health and, for instance, the redistribution of hospital beds. So, the Alyne Network, for example, has been addressing these issues since 2024—such as promoting greater equity in maternal and child health and, for instance, the redistribution of hospital beds. So, looking at this issue of cost and ICU usage as a way to identify where there are gaps—what we can do in that location—and not viewing higher spending as a bad thing, but rather understanding that in that other under-resourced location, we also need to invest more. Investing, at the end of the day, isn’t a bad thing. We need to take care of our population.
And then we also turn our attention to academia. So, we still have a very large gap in the field of health economic evaluation, particularly in maternal and child health. Little is known about the cost-effectiveness of the interventions we’ve been using in our health care system. So, we have to look into this as well—to understand the cost-effectiveness of what we’re implementing on the ground. We need the cost data to build models. Our study is providing the foundational data necessary for comprehensive economic evaluation studies, so that health system managers can make informed decisions based on the more complex information that will be generated.
Another point I’d like to address—one that’s more complex for the general public but relevant to managers at the Ministry of Health—is an issue we’ve consistently emphasized in our studies: the rates in the SIGTAP table (Management System for the SUS Table of Procedures, Medications, and OPM) are sometimes a bit out of step with what is actually spent on healthcare. So, we sometimes even have a bit of trouble explaining—for example, to people outside Brazil—how this fixed value has remained unchanged over time, with many years without an adjustment, and that sometimes the figure we’re presenting doesn’t reflect the actual amount spent on that hospital stay. And the issue of data quality, which I’ll bring up again: if we don’t invest in improving quality and ensuring that the data we’re using to generate this study is adequate, we won’t produce data as reliable as we’d like. So, a focus on improving data quality is also important in terms of the message I wanted to convey.
Everton Nunes:
Great. Julia, I think this final figure that you analyzed specifically in relation to the SIGATAP fixed rates — when retrieving data on hospital admission authorizations — is only one part of the funding picture. From this perspective, we do not take into account the contributions from states and municipalities. So, these figures are likely much higher than what you’ve discussed, but they already give an idea of just how significant the impact is on the budget. Therefore, this is a situation that requires substantial resources, and any improvement will yield benefits or free more funds from the budget. So, I think you made a very good point that we need clinical protocols based on the best evidence of cost-effectiveness and other available evidence. We’ve only covered a small part of the article’s findings here, but there are several other interesting points. So, I’ll leave the final word to you, Julia: if there’s anything from the article, you’d like to highlight that we haven’t discussed yet and that might pique our readers’ curiosity even more.
Julia Raffin:
I think this study is very interesting because it raises more questions than anything else. So, when we read it — especially the discussion section — I think it raises a very important concern, particularly for researchers. We’ve already touched on some of these questions here in our conversation: why half of the records still involve unspecified diagnoses; why does the South spend nearly seven times more per 1,000 live births than the North, even though the North has the highest mortality rate? And why is acute respiratory distress syndrome — which is the most complex diagnosis — resulting in fewer ICU admissions, while milder diagnoses and unspecified cases are leading to more? So, I hope this sparks people’s interest, curiosity, and concern, so that we can make progress.