Electronic Fluid Balance Calculator Shows 96.6 Percent Accuracy in Critically Ill Children

Fluid overload is one of the most quietly dangerous complications in pediatric intensive care. When a critically ill child accumulates more fluid than the body can safely manage, the consequences can ripple through the kidneys, lungs, heart, and brain, yet the arithmetic that would reveal this danger is often buried in the daily noise of intensive care. Now a team of pediatric critical care researchers led by Denise C. Hasson of NYU Langone Health’s Hassenfeld Children’s Hospital reports in Pediatric Research that a real-time cumulative fluid balance calculator, built directly into the electronic health record, matched a rigorously validated retrospective calculation on 96.6 percent of ICU patient-days, offering the strongest evidence yet that automated fluid accounting can be trusted at the bedside.

The study, published on 29 September 2026 and conducted on behalf of the EDEMA Collaborative, addresses a deceptively simple problem. Cumulative fluid balance, often abbreviated CFB, is the running net total of every milliliter a patient receives and every milliliter the patient loses. Expressed as a percentage of body weight, percent cumulative fluid balance, or %CFB, has emerged in recent years as one of the most powerful predictors of outcomes in critically ill children. Prior meta-analyses, including a 2024 systematic review in EClinicalMedicine and a 2018 analysis in JAMA Pediatrics, have linked positive fluid balance to longer ventilation, acute kidney injury, and increased mortality. The AWARE study, a multicenter international investigation, further demonstrated that both the magnitude and the timing of fluid overload shape outcomes in children.

Despite this growing evidence base, the researchers argue that fluid overload remains chronically under-recognized in practice. Calculating %CFB by hand requires tallying dozens of inputs: intravenous medications, blood products, nutrition, insensible losses, urine output, drains, and dialysis effluent, all indexed to a weight that may itself change daily. In a busy pediatric intensive care unit, that arithmetic is rarely performed in real time, and clinicians often rely on intuition or spot checks. The new study set out to remove that barrier by embedding the calculation inside the electronic health record itself, so that the number appears automatically wherever clinicians already work.

The technical approach is notable for what it does not require. Rather than building a standalone application or a separate dashboard, the team worked with information-technology expertise, acknowledged in the paper to Denise Dauterman at NYU, to create the calculator within the EHR and then migrate it into the University of Rochester Medical Center system. The calculator indexes net fluid balance to the patient’s weight and updates continuously, presenting a real-time %CFB value drawn from the same documentation flows that clinicians already use. This design choice matters because tools that live outside the clinical workflow are notoriously abandoned, while those woven into the record itself become part of the ambient information environment of the unit.

To validate the tool, the researchers performed a retrospective cohort study in both the pediatric cardiac intensive care unit and the general medical-surgical pediatric intensive care unit of a single center. They included every hospital encounter between June and November 2025, ultimately capturing 468 unique encounters that spanned 154 calendar days and totaled 2,041 ICU patient-days. That scale is meaningful: cardiac patients and general medical-surgical patients have very different fluid landscapes, with the former often receiving large volumes of cardiopulmonary bypass-related fluid and the latter spanning everything from respiratory failure to sepsis, so a calculator that performs across both populations is more likely to generalize.

The validation logic was straightforward but demanding. For each of the first fourteen ICU days of every encounter, the team compared the real-time %CFB produced by the embedded calculator against a documented %CFB computed from de-identified retrospective data extracted from the same electronic health record. The threshold for agreement was strict: the researchers counted any day on which the two values differed by more than 0.1 percent cumulative fluid balance as a discrepancy. On 1,971 of 2,041 patient-days, or 96.6 percent, the real-time calculator matched the documented value within that narrow margin. Across the entire cohort, the correlation between the two measures exceeded 0.99, an R-squared value indicating that the embedded calculation reproduces the reference calculation almost perfectly.

The discrepancies that did arise were concentrated in a small subset of patients. Thirty-four encounters, twenty from the pediatric cardiac ICU and fourteen from the pediatric medical-surgical ICU, showed differences greater than 0.1 percent at some point during their stay. The paper does not identify a single cause for these outliers, but the pattern is familiar to anyone who has audited fluid data: complex patients accumulate documentation edge cases, from unrecorded insensible losses to medications charted in unexpected units, and the highest-acuity children generate the most intricate fluid histories. The near-perfect overall correlation suggests these are data-quality events rather than flaws in the calculator’s logic, but they define exactly where future attention should focus.

What makes the result clinically significant is the framing of reliability as evidence. The authors state plainly that they developed the real-time %CFB calculator within the EHR and validated its accuracy in an attempt to provide evidence of its reliability for clinical use. That evidentiary step is often skipped when institutions deploy internal clinical tools, leaving clinicians to trust software on faith. By quantifying agreement against a retrospective gold standard across more than two thousand patient-days, the EDEMA Collaborative has produced the kind of validation data that can support adoption at other centers, and the funding trail, a grant from the George M. O’Brien Kidney Resource Alliance, signals that the National Institutes of Health sees kidney-focused informatics as a priority area.

The broader context is a field in transition. A 2025 fluid management bundle study in Pediatric Critical Care Medicine found that structured protocols were associated with reduced excess fluid accumulation in children with respiratory failure, and a state-of-the-art review published the same year in Intensive Care Medicine Paediatric and Neonatal catalogued how fluid balance shapes outcomes across pediatric critical illness. The Pediatric Acute Disease Quality Initiative has convened experts to standardize how fluid assessment should be performed in sick children. What has been missing is infrastructure: a way to make the fluid balance number as continuously visible as a heart rate or an oxygen saturation. The new calculator is a direct answer to that gap, and its authors suggest that a %CFB calculator that indexes net fluid balance to weight and is readily available may bring greater attention to clinical fluid overload.

There are, of course, limits to what a single-center retrospective study can establish. The validation took place within one institution’s electronic health record configuration, and EHR implementations vary widely across health systems, meaning the calculator’s performance elsewhere will depend on local data practices and the fidelity of the migration. The study also measured agreement between calculations, not whether displaying the number in real time actually changes clinician behavior or improves patient outcomes, a question that will require prospective interventional trials. Still, the arithmetic is unambiguous: a tool that agrees with a rigorous reference standard on 96.6 percent of patient-days, with correlation above 0.99, has cleared the threshold of reliability that any behavioral study would demand. For the hundreds of thousands of children who pass through intensive care units each year, the study suggests that the single most important number in fluid management may soon be computed for them, continuously, before anyone has to reach for a calculator.

Subject of Research: Development and validation of an electronic health record-based cumulative fluid balance calculator for critically ill children

Article Title: Development and Validation of an Electronic Fluid Balance Calculator for Critically Ill Children

Article References: Hasson, D. C., Odum, J. D., Kothari, U., Shah, A. J., Braun, C. G., Dixon, C. G., Fitzgerald, J. C., Dziorny, A. C., & on behalf of the EDEMA Collaborative (2026). Development and Validation of an Electronic Fluid Balance Calculator for Critically Ill Children. Pediatric Research. https://doi.org/10.1038/s41390-026-05496-1

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05496-1

Keywords: pediatric intensive care, fluid balance, fluid overload, electronic health record, cumulative fluid balance, clinical calculator, validation study, pediatric research, acute kidney injury, critical care informatics, retrospective cohort, EDEMA Collaborative