Quick takeaway: The Khamis-Roche method is the most accurate non-skeletal algorithm for predicting adult height in healthy children aged 4 through 17. It requires the child's current age, height, weight, and biological parents' heights, yielding a 95% statistical margin of error of ±2.1 inches (5.3 cm) for boys and ±1.7 inches (4.3 cm) for girls.
How Human Height Prediction Works
Predicting a child's eventual adult height combines genetics, human growth biology, and statistical regression modeling. While final adult stature is roughly 80% determined by inherited DNA, environmental factors such as nutrition, endocrine health, and childhood sleep play vital roles.
Our Height Suite provides two medically validated prediction methods alongside visual height comparison graphics.
1. The Khamis-Roche Method
Developed by Dr. Harry Khamis and Dr. Alex Roche at the FELS Research Institute, the Khamis-Roche method uses linear regression formulas derived from multi-decadal longitudinal human growth studies.
Unlike older formulas that rely solely on parents' heights, Khamis-Roche incorporates the child's current skeletal development markers (current height, weight, and decimal age).
The Mathematical Formula
The general linear equation takes the following mathematical structure:
Adult Height = β₀ + (βₕ × Child Height) + (βw × Child Weight) + (βₚ × Mid-Parental Height)
Where:
- Adult Height is the predicted adult stature in inches or centimeters.
- Child Height is the child's measured standing height.
- Child Weight is the child's measured weight.
- Mid-Parental Height is the adjusted mid-parental height average.
- β₀, βₕ, βw, βₚ are age-specific and sex-specific regression coefficients derived from clinical dataset cohorts.
Expert Insight: Bone age X-rays of the left wrist remain the absolute gold standard in pediatric endocrinology (using the Greulich-Pyle atlas). However, clinical trials confirm that for healthy children without hormone deficiencies, the Khamis-Roche method approaches nearly identical accuracy without subjecting the child to radiation or medical expense.
2. The Mid-Parental Tanner Target Height Method
For infants, toddlers under age 4, or unborn children, the Mid-Parental formula (also known as the Tanner method) provides the baseline genetic target height.
Formula for Boys
Target Height (Boys) = (Father Height + Mother Height + 5 inches) / 2
Or in metric units:
Target Height (Boys) = (Father Height + Mother Height + 13 cm) / 2
Formula for Girls
Target Height (Girls) = (Father Height + Mother Height - 5 inches) / 2
Or in metric units:
Target Height (Girls) = (Father Height + Mother Height - 13 cm) / 2
Fully Worked Real-World Example
Let's calculate the predicted adult height for a 10-year-old boy named Ethan.
- Ethan's Current Age: 10 years
- Ethan's Current Height: 4 feet 6 inches (54 inches / 137.2 cm)
- Ethan's Current Weight: 75 lbs (34 kg)
- Father's Height: 5 feet 10 inches (70 inches / 177.8 cm)
- Mother's Height: 5 feet 5 inches (65 inches / 165.1 cm)
Step 1: Calculate Mid-Parental Target Height
- Sum parents' heights: 70 + 65 = 135 inches.
- Add 5 inches for a boy: 135 + 5 = 140 inches.
- Divide by 2: 140 / 2 = 70 inches (5 feet 10 inches / 177.8 cm).
Step 2: Apply Khamis-Roche Coefficients for Age 10
Using the FELS age-10 male coefficients:
- Child Height Factor: 54 * 0.561 = 30.29 in
- Child Weight Factor: 75 * 0.032 = 2.40 in
- Mid-Parental Factor: 70 * 0.395 = 27.65 in
- Intercept Constant: +10.42 in
Predicted Adult Height = 30.29 + 2.40 + 27.65 + 10.42 = 70.76 inches
Ethan is projected to reach 5 feet 10.8 inches (179.7 cm), with a 95% confidence interval spanning between 5 feet 8.7 inches and 6 feet 0.9 inches.
Common Mistakes to Avoid
- Assuming linear growth through puberty: Children undergo non-linear growth spurts. Early bloomers may appear taller at age 12 but stop growing earlier, while late bloomers gain height well into their late teens.
- Using shoe measurements: Standing height must always be measured barefoot against a flat vertical wall with heels touching the baseboard.
- Forgetting biological parent heights: Using step-parent or adopted parent heights will invalidate the genetic regression calculations.
- Relying solely on doubling age 2 height: While doubling a boy's height at age 2 (or a girl's height at 18 months) provides a fun rough estimate, it carries a large margin of error (up to ±4 inches).