Study objective. To obtain reliable and accurate predictions of adult stature in white American children who are free of disease without using skeletal age.

Design. Apply a modification of the Roche-Wainer-Thissen stature prediction model to a sample of white American children.

Setting. Longitudinal data (every 6 months) from participants in the Fels Longitudinal Study were used for the development of the stature prediction model.

Participants. Residents of Southwest Ohio, 223 white males and 210 white females, at the time of their enrollment into the Fels Longitudinal Study.

Measurements and results. The errors of the proposed method, which does not use skeletal age as a predictor variable, are only slightly larger than those for the Roche-Wainer-Thissen method which uses skeletal age as a predictor variable.

Conclusions. Adult stature predictions are needed commonly but the current methods are difficult to apply because they require a skeletal age assessed by a modern method. The Khamis-Roche method predicts adult stature in the absence of skeletal age with only a slight deterioration in accuracy and reliability. The applicability of the Khamis-Roche method is limited to white American children without pathologic conditions that alter the potential for growth in stature, but it should be useful for white children who are unusual in stature or in levels of maturity for age.