IRSST - Institut de recherche Robert-Sauvé en santé et en sécurité du travail

Development of a Personalized Decision Support Tool to Minimize Manual Handling Risks Involving Symmetrical and Asymmetrical Postures


Excessive loads on the lumbar spine are a major risk factor for back pain. Biomechanical modeling of the spinal column is the only non-invasive approach to estimate these loads, but the complexity of these models limits their use by ergonomists. To help them identify situations in which there is a risk of back injuries in order to limit loading on the lumbar structures, there are decision support tools, such as the NIOSH lifting equation (Revised NIOSH Lifting Equation, Waters et al., 1993), and the Snook lifting tables (Snook and Ciriello, 1991, Liberty Mutual). However, they do not directly calculate lumbar loading, but instead provide an estimate of the maximum weight that a population of workers can lift safely. In the past, Arjmand et al. (2011, 2012) developed regression equations based on the detailed kinematic model developed by Shirazi-Adl’s team. They establish a simple relationship between lumbar loading (compression and shear forces at L4-L5 and L5-S1) and independent input variables easily measurable by the professional. However, these equations are limited to symmetrical postures and do not take into account variations in workers’ weight and height. The objective was to (1) develop equations to predict spinal loads, taking into account, for the first time, personal factors such as the subject’s weight and height; (2) improve the equations used to predict spinal loads during asymmetrical handling activities, which would be used by ergonomists to assess which situations could cause injuries.

Method: Our kinematic finite element model of the torso was modified to incorporate personalized variables, including the subjects’ weight, height, age and gender, and then used to assess the effect of these individualized variables (weight, height, gender and age) on compression and shear forces at L4-L5 and L5-S1. Direct measurements were taken of 19 asymptomatic subjects during symmetrical and asymmetrical manual handling tasks, in order to record 3D torso and pelvic rotations and the EMG activity of superficial muscles. The improved kinematic model made it possible to estimate the compression and shear forces during symmetrical/asymmetrical tasks and simulations extended the input and output data useful for the development of new regression equations. These equations can predict compression/shear forces using simple input variables such as the subject’s weight and height, load size, distance (lever arm) of the load in the hands, slope and angle of asymmetry of the torso.

Results: Prediction equations for compression and shear forces on the spinal column at L4-L5 and L5-S1 were developed for manual handling: (1) in symmetrical postures, taking into account variations in loads, posture, weight, height, gender and age; and (2) in asymmetrical postures, taking into account load size, load position in the hands, asymmetrical torso rotation and individualized parameters of the subjects, such as input variables. These regression equations had the best match (followed by OpenSim, AnyBody, the McGill polynomial and 3DSSPP) with the in vivo intradiscal measurements. In addition, in people with high body weight, the estimates of recommended weights using the NIOSH lifting equation generated lumbar compression forces higher than the limit recommended by the NIOSH. In contrast, in lighter individuals, the weights recommended by the NIOSH remained conservative (compression < 3 400 N and shear < 1 250 N).

Benefits: This study now provides OHS professionals with simple predictive equations of compression and shear forces on the spine to help in making decisions to minimize manual handling risks involving symmetrical and asymmetrical postures.

Additional Information

Category: Research Report
  • Aboulfazl Shirazi-Adl
  • Navid Arjmand
  • André Plamondon
Research Project: 2014-0009
Online since: February 05, 2021
Format: Text