Low back pain (LBP) during pregnancy afflicts about half of all pregnant women. Pregnant women leaving work early because of LBP cost employers in sick leave pay, replacement and training costs.

Anterior weight gain and awkward postures may result in higher loads on the low back, increasing the risk of injury. The changing biomechanics of the pregnant woman's body have been suggested as a cause of LBP during pregnancy as a result of increased loads on the spine. An investigation of the loads at the low back of pregnant women requires the development and application of a biomechanical model of pregnancy using anthropometric data that was representative of the ongoing changes in the body of a pregnant woman.

A 3-D dynamic link segment model composed of eleven body segments was developed by Pei Lai Cheng and Delphine Perie to determine the loads on L5/S1 (5th Lumbar/1st Sacral Vertebrae) during lifting tasks. Hands-down and feet-up versions of the biomechanical model determined loads at L5/S1. The models used position and orientation data of the subject's body segments colleted by the Motion Star®, ground reaction loads collected by a force platform, and individualized anthropometric data.

The purpose of this work was to determine the validity of the model. In pursuit of the generalization of the model validation, the gender and body mass index of the participant, the pregnant state of the participant and the tasks performed were varied. Participants were asked to perform several lifting tasks with three parameters varied:​ symmetry of task, duration, and lifting technique. Loads were 3.5kg for women and 12 kg for men.

The forces and moments were compared at L5/S1 from the hands-down and the feet-up models, and the measured ground reaction forces and moments were compared to those predicted using a hands-down model composed of all 11 body segments and the applied load.

The model performance was comparable to the literature of similar validation studies for all force and moment components at both L5/S1 and the ground except for the flexion/extension moment and the moment about the transverse axis at the ground where errors were large, particularly at the beginning of the lift. Model performance for female and pregnant participants was similar.