A Smarter Way to Decide Which Newborns Need Antibiotics

What a large Dutch study of 1,830 babies (van der Weijden et al., EClinicalMedicine 2025) found when hospitals used a risk calculator instead of standard checklists

A large study across ten hospitals in the Netherlands tested whether an online "sepsis calculator" could safely reduce the number of newborns given antibiotics in their first hours of life. Compared with the usual checklist approach, hospitals using the calculator started antibiotics in far fewer babies — about 7 in 100 instead of about 27 in 100 — without any sign that babies were harmed by the more selective approach [1]. The finding matters because most newborns who receive precautionary antibiotics turn out not to have an infection, and treating fewer of them, safely, spares families and babies a stressful and not-entirely-harmless intervention.

Why This Question Matters

In the first days after birth, a small number of babies develop a serious bloodstream infection called early-onset sepsis. It is uncommon, but it can become life-threatening quickly, and its early warning signs — a baby who is a little too sleepy, breathing a little too fast — look much like the ordinary bumps of the newborn period. Faced with that uncertainty, doctors have long erred heavily on the side of caution: if a baby had any risk factor, such as the mother having a fever in labour, a long gap between the waters breaking and delivery, or a particular bacterium called group B streptococcus, the baby was often started on antibiotics through a drip while blood tests were checked. The instinct is understandable. The problem is arithmetic: because true infection is rare, this approach means many babies are treated for every one who is actually infected.

That over-treatment is not without cost. Being started on antibiotics usually means a needle for a blood culture, a drip line, time in a special-care area rather than skin-to-skin with a parent, and interruption of early breastfeeding and bonding. There are medical costs too. In the most fragile, very premature babies, longer courses of early antibiotics have been linked to a dangerous bowel condition called necrotising enterocolitis and to higher death rates [2], and broader reviews have connected early antibiotic exposure to a range of later problems [3]. Antibiotics also disturb the community of helpful bacteria a baby is just beginning to build, and every unnecessary course adds, in a small way, to the worldwide rise of antibiotic-resistant bugs.

How Families and Clinicians Used to Face This

For decades the decision rested on a checklist: count the risk factors, and if certain ones were present, treat. Researchers gradually asked whether the odds could be judged more precisely. They studied hundreds of thousands of births to work out how a mother's specific circumstances change a baby's actual chance of infection [4], and then how that starting risk should be combined with how the baby actually looks in the hours after birth [5]. Those insights were built into a free online tool, the early-onset sepsis calculator, which a clinician can use at the cot side: enter the details, and it suggests whether to simply watch the baby, check them more closely, or start antibiotics [6].

When hospitals began using the calculator, they reported giving antibiotics to roughly half as many babies as before, and safety looked reassuring [7]. But there was a catch that kept expert groups from fully recommending it. All of that evidence came from simply watching what happened when hospitals adopted the tool, rather than from a fair head-to-head experiment. Hospitals that chose to adopt a new calculator might differ in many ways from those that didn't, so the comparison could be misleading. National guidelines therefore listed the calculator as one acceptable option among several, without declaring it better [8]. To settle the matter properly, someone needed to run a real trial.

What the Researchers Did

That is what this study set out to do [1]. Instead of leaving it to each hospital to choose, the researchers randomly assigned ten Dutch hospitals to one of two approaches: five would use the sepsis calculator, and five would continue with the standard Dutch checklist. Assigning whole hospitals, rather than individual babies, makes sense for a tool that works as a ward-wide way of doing things, and it keeps the two approaches from getting mixed up on the same unit. The plan for the study, including exactly how safety would be measured, was written down and published in advance so it could not be adjusted after the fact [9].

Between April 2022 and March 2024 the study included 1,830 babies born at 34 weeks of pregnancy or later, each with at least one sepsis risk factor — about 915 in each approach. The researchers watched two things at the same time. First, safety: were babies any more likely to run into trouble, judged by whether they needed help with breathing, needed support for their circulation, were transferred to intensive care, or had a laboratory-confirmed infection? Second, effectiveness: were fewer babies started on antibiotics in the first 24 hours? They also recorded how long treated babies stayed on antibiotics.

What They Found

On safety, the calculator held up well. Signs of trouble were recorded in 7.0% of babies in the calculator hospitals and 14.6% in the checklist hospitals [1]. In other words, the more selective approach showed no evidence of putting babies at greater risk — the study's central safety worry did not materialise. (These particular measures partly reflect how sick babies happened to be at different hospitals, so the fairest way to read them is as reassurance that the calculator did no harm.)

On the main goal, the difference was large. Antibiotics were started in just 7.2% of babies in the calculator hospitals, compared with 26.6% under the checklist [1]. That is roughly one in fourteen babies treated instead of roughly one in four — a real, everyday difference in how many families face a drip, a blood test, and separation in the first hours after birth. Serious problems and hospital readmissions for suspected infection were rare and similar in both groups, and when a handful of babies were brought back for a check, their tests came back clear [1].

There was one result worth understanding honestly. Among the babies who were treated, those in the calculator hospitals stayed on antibiotics a bit longer on average — about five and a half days versus two [1]. This sounds surprising but makes sense. The checklist starts a lot of low-risk babies who are quickly stopped once tests come back clear, which pulls the average length down. The calculator, by contrast, tends to reserve treatment for the babies who genuinely look more concerning, and those babies are more likely to be treated for a full course. So "fewer babies treated" is the reliable headline; "fewer total days of antibiotics" is a separate question the study did not settle.

It also helps to know what this study could and could not show, so the results are neither dismissed nor oversold. Because the researchers compared whole hospitals rather than individual babies, and because only ten hospitals took part, small differences between the hospitals themselves — the kinds of families they serve, how quickly they admit babies to intensive care — could account for part of the safety difference. That is why the honest summary is not "the calculator makes babies safer," but rather "the calculator let doctors treat far fewer babies without any sign that safety suffered." It is also important that the study only included babies born at 34 weeks or later; it does not tell us how the tool would perform in very premature babies, who are the most vulnerable of all and for whom these decisions are hardest. Careful, large studies like this one are how a promising idea earns its place in everyday care rather than being adopted on hope alone.

What This Means for Families and What Comes Next

For parents, the practical meaning is encouraging. Used carefully, the calculator lets clinicians say to more families, "Your baby's individual risk is low enough that we can watch closely rather than start antibiotics right away" — keeping more mothers and babies together and skin-to-skin, with fewer needles and drips, and without the study finding a safety cost [1]. It does not replace clinical judgement: the tool relies on a nurse or doctor repeatedly checking the baby, and any baby who looks unwell is treated regardless of what the calculator says.

Researchers are now working on the next questions. They want to confirm the safety findings in even larger numbers, to be sure the rare truly-infected baby is not missed; to work out how to also shorten treatment in the babies who do need it, perhaps by pairing the calculator with a blood test that signals when it is safe to stop [10]; and to keep the tool accurate as infections and hospital practices change over time [11]. Ultimately the goal behind all of this is to give each newborn exactly the care they need and no more — protecting the few who are truly infected while sparing the many who are not from antibiotics, and helping slow the wider rise of drug-resistant bacteria that unnecessary treatment feeds [12].

References

  1. van der Weijden BM, Janssen SWCM, van der Weide MC, et al. Safety and effectiveness of the early-onset sepsis calculator to reduce antibiotic exposure in at-risk newborns: a cluster-randomised controlled trial. EClinicalMedicine. 2025;87:103419. doi:10.1016/j.eclinm.2025.103419 ↩
  2. Cotten CM, Taylor S, Stoll B, et al. Prolonged Duration of Initial Empirical Antibiotic Treatment Is Associated With Increased Rates of Necrotizing Enterocolitis and Death for Extremely Low Birth Weight Infants. Pediatrics. 2009;123(1):58–66. doi:10.1542/peds.2007-3423 ↩
  3. Esaiassen E, Fjalstad JW, Juvet LK, van den Anker JN, Klingenberg C. Antibiotic exposure in neonates and early adverse outcomes: a systematic review and meta-analysis. J Antimicrob Chemother. 2017;72(7):1858–1870. doi:10.1093/jac/dkx088 ↩
  4. Puopolo KM, Draper D, Wi S, et al. Estimating the Probability of Neonatal Early-Onset Infection on the Basis of Maternal Risk Factors. Pediatrics. 2011;128(5):e1155–e1163. doi:10.1542/peds.2010-3464 ↩
  5. Escobar GJ, Puopolo KM, Wi S, et al. Stratification of Risk of Early-Onset Sepsis in Newborns ≥34 Weeks' Gestation. Pediatrics. 2014;133(1):30–36. doi:10.1542/peds.2013-1689 ↩
  6. Kuzniewicz MW, Puopolo KM, Fischer A, et al. A Quantitative, Risk-Based Approach to the Management of Neonatal Early-Onset Sepsis. JAMA Pediatr. 2017;171(4):365–371. doi:10.1001/jamapediatrics.2016.4678 ↩
  7. Achten NB, Klingenberg C, Benitz WE, et al. Association of Use of the Neonatal Early-Onset Sepsis Calculator With Reduction in Antibiotic Therapy and Safety: A Systematic Review and Meta-analysis. JAMA Pediatr. 2019;173(11):1032–1040. doi:10.1001/jamapediatrics.2019.2825 ↩
  8. Puopolo KM, Benitz WE, Zaoutis TE; Committee on Fetus and Newborn; Committee on Infectious Diseases. Management of Neonates Born at ≥35 0/7 Weeks' Gestation With Suspected or Proven Early-Onset Bacterial Sepsis. Pediatrics. 2018;142(6):e20182894. doi:10.1542/peds.2018-2894 ↩
  9. van der Weijden BM, van der Weide MC, Plötz FB, Achten NB, et al. Evaluating safety and effectiveness of the early-onset sepsis calculator to reduce antibiotic exposure in Dutch at-risk newborns: a protocol for a cluster randomised controlled trial. BMJ Open. 2023;13(2):e069253. doi:10.1136/bmjopen-2022-069253 ↩
  10. Stocker M, van Herk W, el Helou S, et al. Procalcitonin-guided decision making for duration of antibiotic therapy in neonates with suspected early-onset sepsis (NeoPIns): a multicentre, randomised controlled trial. Lancet. 2017;390(10097):871–881. doi:10.1016/S0140-6736(17)31444-731444-7) ↩
  11. Kuzniewicz MW, Escobar GJ, Forquer H, et al. Update to the Neonatal Early-Onset Sepsis Calculator Utilizing a Contemporary Cohort. Pediatrics. 2024;154(4):e2023065267. doi:10.1542/peds.2023-065267 ↩
  12. Fjalstad JW, Esaiassen E, Juvet LK, van den Anker JN, Klingenberg C. Antibiotic therapy in neonates and impact on gut microbiota and antibiotic resistance development: a systematic review. J Antimicrob Chemother. 2018;73(3):569–580. doi:10.1093/jac/dkx426 ↩