Every newborn admitted to a hospital is quickly assigned a label that shapes the care they receive: small for gestational age, appropriate, or large. That label depends not just on the baby’s weight and gestational age, but on which growth reference chart a clinician happens to use. A new study from a tertiary referral hospital in Yogyakarta, Indonesia, suggests that this choice matters far more than many clinicians assume, because the international charts widely used around the world classify Indonesian babies differently from a chart built specifically from Indonesian birth data.
The research, published in BMC Pediatrics, was conducted by a team from Universitas Gadjah Mada led by Ade Febrina Lestari and Ekawaty Lutfia Haksari. Their goal was straightforward but clinically important: to test how well the Haksari Newborn Curve, a growth reference developed from Indonesian birth data, agrees with two of the most influential international references, the Fenton Growth Chart and the INTERGROWTH-21st Newborn Size Standards, when classifying birthweight relative to gestational age. The answer, in short, was that they do not agree well enough to be used interchangeably.
The study was a cross-sectional analysis of medical records from 1,395 singleton live-born infants delivered at Academic Hospital Universitas Gadjah Mada between 2018 and 2025. The researchers deliberately excluded babies with major congenital anomalies, since such conditions can distort birthweight independent of gestational growth. Just over half of the newborns, 54.1 percent, were boys, and 16.6 percent were born preterm, meaning before 37 completed weeks of gestation. Each infant’s birthweight-for-gestational-age was classified as small, appropriate, or large using all three references, and the researchers then measured how often the classifications matched.
The statistical backbone of the study was Cohen’s kappa, a measure that quantifies agreement between classification systems beyond what would be expected by chance alone. Kappa values range from zero, indicating no agreement beyond chance, to one, indicating perfect agreement. The researchers calculated both unweighted and weighted versions of kappa, with weighted kappa giving partial credit when two charts disagreed only slightly, for example when one labeled a baby appropriate and the other labeled the same baby large. They also reported observed agreement, the raw percentage of babies classified identically, and examined discordance according to gestational age and sex. A sensitivity analysis using exact gestational age rather than completed weeks confirmed the robustness of the findings.
The design included an important methodological nuance. The Fenton Growth Chart was originally constructed to describe growth in preterm infants, so the researchers treated the preterm-restricted comparison between Haksari and Fenton as their main analysis for that reference, covering only babies born before 37 weeks. A full-population comparison with Fenton was presented as exploratory. For INTERGROWTH-21st, which covers the entire gestational-age range, the main comparison included all 1,395 newborns. This careful matching of chart to intended clinical population strengthens the study’s conclusions, because it avoids penalizing Fenton for being applied outside its design purpose.
Even under these favorable conditions, the agreement was only moderate. In the main preterm-restricted comparison, Haksari and Fenton showed a weighted kappa of 0.490, with a 95 percent confidence interval of 0.391 to 0.589, and discordance in 28.1 percent of cases. When the full population was included in the exploratory analysis, agreement dropped to fair, with a weighted kappa of 0.245 and 33.8 percent discordance. The main comparison between Haksari and INTERGROWTH-21st yielded a weighted kappa of 0.405, with a confidence interval of 0.355 to 0.455, and 22.6 percent discordance. In practical terms, roughly one in four or one in three babies received a different label depending on which chart was used.
Why do the charts diverge so much? The study found a clear structural reason: within the overlapping gestational-age ranges, the Haksari curve showed lower 10th and 90th percentile boundaries than the international references. Percentile boundaries are the operational heart of any growth chart. A baby below the 10th percentile is typically classified as small for gestational age, while a baby above the 90th percentile is classified as large. If the Indonesian curve sits lower than the international curves, then a baby who appears appropriately grown on the Haksari chart may fall below the 10th percentile on Fenton or INTERGROWTH-21st, flipping the label from appropriate to small. Conversely, some babies classified as large by international standards would fall within the normal range on the Indonesian curve.
The clinical consequences of such misclassification are not trivial. Small-for-gestational-age babies are at elevated risk of hypoglycemia, hypothermia, feeding difficulties, infections, and long-term metabolic and neurodevelopmental problems, so this label often triggers closer monitoring, additional investigations, or admission to a neonatal intensive care unit. Large-for-gestational-age classification raises suspicion of maternal diabetes and prompts screening for birth injury and metabolic complications. A chart that systematically shifts these thresholds relative to the local population will therefore generate both false alarms and missed cases, straining resources in one direction and leaving vulnerable infants under-observed in the other.
The study’s authors were careful about the limits of their evidence. This was a single-centre study at one tertiary referral hospital, which may not represent the full diversity of Indonesian births, including deliveries in rural settings, lower-level facilities, or different regions of the archipelago. Because the analysis used secondary data from medical records, it depended on the accuracy of recorded birthweights and gestational ages. Most importantly, the study measured agreement between classification systems, not whether one system predicted adverse outcomes better than another. The authors explicitly concluded that further outcome-based studies in diverse Indonesian populations are needed before any growth reference can be recommended for routine clinical use.
Nevertheless, the message for global neonatal medicine is significant. The Fenton chart, derived from populations in the United States, Canada, and Australia, and the INTERGROWTH-21st standards, drawn from eight geographically diverse but predominantly well-nourished populations, are both used in countries where newborns may have substantially different anthropometry. This Indonesian study provides concrete quantitative evidence that externally derived references are not interchangeable with locally derived ones, and that the differences are large enough to change clinical decisions for a substantial fraction of newborns. As countries worldwide work to localize growth standards, the Yogyakarta findings offer a template for how to evaluate them: compare the local curve against international references using kappa statistics, quantify discordance at both the small and large thresholds, and then test which classification best predicts real outcomes. Until that outcome evidence exists, clinicians using imported charts in Indonesian and similar populations should be aware that the labels they assign may depend as much on the chart in the drawer as on the baby in the bassinet.
Subject of Research: Agreement between Indonesian and international newborn growth references for birthweight-for-gestational-age classification
Article Title: Comparison of the Haksari newborn curve vs. the Fenton growth chart and INTERGROWTH-21st newborn size standards for classifying birthweight relative to gestational age: a report from a tertiary referral hospital in Indonesia
Article References: Lestari, A. F., Nurmadina, A. M., Haksari, E. L., Sitaresmi, M. N., & Julia, M. (2026). Comparison of the Haksari newborn curve vs. the Fenton growth chart and INTERGROWTH-21st newborn size standards for classifying birthweight relative to gestational age: a report from a tertiary referral hospital in Indonesia. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07774-7
Image Credits: AI Generated
DOI: 10.1186/s12887-026-07774-7
Keywords: birthweight classification, gestational age, Haksari newborn curve, Fenton Growth Chart, INTERGROWTH-21st, small for gestational age, large for gestational age, Cohen's kappa, neonatal growth standards, Indonesia, BMC Pediatrics, growth charts
News Source: Harold Sullivan. (October 19, 2026). Indonesian Newborn Growth Chart Reveals Global Standards May Misclassify Babies. Scienmag.
