Methodology

Growth-Engine

All growth references run through a common engine. A dataset defines the reference, axis (age in days or months, postmenstrual age in weeks, length or height in cm), sex, measure, unit, and the LMS parameters for each anchor point. Linear interpolation is used between anchor points; no value is output outside the reference range.

Z-score according to Cole & Green (LMS method): z = ((value/M)^L − 1) / (L·S) for L ≠ 0, otherwise z = ln(value/M)/S. Percentile = Φ(z)·100. The inverse yields the measured value for a z-score. For the extended CDC BMI, above P95 the half-normal distribution of Wei et al. 2020 applies.

Age

Chronological age in days from date of birth and date of measurement; months = days / 30.4375 (CDC definition). Corrected age for preterm infants = chronological age − (280 days − gestational age at birth), applied until month 24 of life. Postmenstrual age = gestational age + chronological age.

Data source

The origin of each dataset is given in the Data set directory and in each file under data/growth/. Only published reference values are used; where authors do not release their curve data, the portal shows the gap instead of estimating.

Bilirubin

AAP 2022: hour-specific thresholds from Supplemental Tables 1–4 of the guideline. AAP 2004: documented digitization of the curves (origin stated in the calculator). AWMF 024-007: formula set of the German guideline.

Validation

php tests/run.php checks anchor values (the median of the source tables gives z = 0), round trips z → value → z on all datasets, the normal distribution functions, the extended BMI and the remaining calculation cores.