Hip axis length (HAL), a simple measure obtained from dual X-ray absorptiometry (DXA) scans of the proximal femur, predicts hip fracture, independent of bone mineral density and age. Owing to a magnification error associated with newer fan-beam DXA, automatic calculation of HAL, which is available in pencil-beam DXA, is inaccurate. The purpose of this study was to model HAL, measured from a pencil-beam DXA, as a function of HAL from a fan-beam DXA combined with other anthropometric measures. Proximal femur scans were performed using pencil- and fan-beam DXA in 21 women (ages 24-60). Height, weight, hip circumference, subject thickness, height of the greater trochanter and anterior superior iliac spine from the scanning table, and HAL as measured by a ruler from the scan printouts were recorded. Anthropometric measurements were taken by two researchers; all but the greater trochanter were reproducible (r > or = 0.92). A simple linear model using the manual measurement of HAL from the fan-beam scan, height, weight, body mass index, and hip circumference was able to predict HAL measured using the pencil-beam DXA with a high degree of accuracy (R(2) > or = 0.96). The fan-beam-acquired values of HAL using our model are nonbiased and accurate estimates of the "gold standard" pencil-beam method. This model may provide researchers and clinicians with a simple method of calculating HAL using fan-beam DXA.