Summary The incidence of musculoskeletal disorders (MSD) among both male and female workers in Québec was recently estimated at 732,000 cases (Stock et al., 2014). These MSDs are associated with occupational activities involving prolonged, intense and/or repetitive physical effort, such as assembly, handling and personal assistance tasks, or remaining in the same position for prolonged periods at a workstation. After the back, the upper limbs, which represented 30.1% of MSD cases reported and accepted in Québec from 1998 to 2007 (Michel et al., 2010), are the most affected. The shoulder joint alone accounts for 46.5% of these injuries, 80% of which are tendinitis and sprains (Duguay et al., 2012). In the United States, these MSDs cause the highest number of lost workdays per year (Statistics 2015, s.d.). Women and new or younger employees are also at greater risk of developing an MSD in the shoulder (Breslin and Smith, 2005; Nordander et al., 2008; Treaster and Burr, 2004). In a context where MSDs of the shoulder have become a real public and occupational health issue, a better understanding of shoulder biomechanics during the performance manual tasks appears necessary and will lead to occupational health and safety recommendations to reduce exposure to the risks of shoulder MSDs. The overall objective was to analyze and compare handling techniques involving the shoulder between men and women (study 1) and between expert and novice handlers (study 2) in the distribution sector, using synthetic risk indicators. Several manual handling tasks were simulated in the laboratory with the aim of finding less stressful strategies for the shoulder and to assess the original synthetic indicators for their potential to estimate musculoskeletal stresses in order to recommend exposure indicators. The participants recruited had to move instrumentation-equipped boxes of various weights (6 kg, 8 kg, 12 kg), from a low point at hip level to a high point above shoulder level. A kinematic analysis was conducted to describe the articular strategies used by the participants. The results support the viewpoint of a gender-specific strategy for upper-limb joint contribution during a manual handling task. With a 6 kg weight, women used their glenohumeral joints more than men. However, with the 12 kg weight, the opposite was observed, with men’s glenohumeral joints contributing more than women’s. For the heavier weight, women compensated for the lower contribution of their glenohumeral joints by using their wrist and elbow joints more. Most of the differences reported were found during the deposit phase, when the arms were at or above shoulder level. The results also suggest that expertise plays a role in upper-limb joint contribution during the box-lifting task. During the pulling phase, novices used their wrist, elbow, shoulder and torso joints to bring their hands closer to the box to hold it. Experts, on the other hand, brought their entire bodies close to the box, involving their lower limbs to limit the contribution of their arms and to maintain their torsos in a neutral position. Between the pulling and lifting phases, both groups mainly used their wrists and elbows, while their shoulders contributed about 30% of the box height. There was greater involvement of the sternoclavicular and acromioclavicular groups by the experts, which means that they stabilize the joint more effectively during this transition. Between the lifting and deposit phases, both groups used a similar technique, in which shoulder flexion contributed from 55% to 60% of the deposit height of the box. An electromyographic analysis (EMG) of the ten muscles of the shoulder girdle and upper arms was conducted. As expected, this experimental EMG measurement established that the participants expended more muscular activity to lift the 12 kg box than to lift the 6 or 8 kg boxes. The increase in muscle activity was greater in the motor muscles (anterior deltoid, lateral deltoid, pectoral, biceps brachialis) than in the stabilizer muscles (supraspinatus, infraspinatus, subscapularis) and the antagonist muscles (posterior deltoid, latissimus dorsi, triceps brachialis). A similar tendency was observed with respect to the role of gender on muscular activity. Women’s motor and antagonist muscles generated greater muscular activity than men’s, for a similar absolute load. In fact, women’s muscular activation when lifting a 6 kg box was similar to that of men lifting a 12 kg box, reaching up to 48% of maximal activation in the anterior deltoid. This result is consistent with several studies that show that women’s maximal strength is 30 to 60% less than men’s for the various muscle groups of the upper limbs (Douma et al., 2014; Faber et al., 2006; Harbo et al, 2012). The results do not seem to show a difference in muscle activation between experts and novices. While novices appear to exhibit greater muscle activation than experts in the lifting phase, the distribution of muscle activation over the course of the task is similar between the groups. A difference in strategy involving a longer pulling phase among experts proportionate to the time the task takes could be the cause for a shift in muscle activation between the lifting and deposit phases of both groups, implying a statistical difference that may not necessarily be physically present. Finally, a musculoskeletal model was designed to estimate the internal forces applied to the upper limb joints when a box is lifted. This model is sensitive to variations in weight, gender and expertise, as well as to the various movement phases. As anticipated, the sum of muscular activation and the sum of muscular force are higher with a 12 kg weight than with 6 kg and 8 kg weights in the various movement phases. Women and experts exhibited greater muscular activation and strength than men and novices, respectively. The relative time spent above a dislocation compression-shear ratio of the glenohumeral joint is higher with a 12 kg weight, but the same is true for women and experts. One recommendation that could be made from this study is that it is probably safer to bring the box closer to the body during a handling task. This technique reduces moment of the arm and may reduce stress at the shoulder and elbow. This kinematic change could also influence the direction of muscular force, thus enabling less activation for the same stability. In addition to lessening loading on the spinal column (Marras et al., 2006) this technique could also be a significant factor in reducing shoulder injuries by limiting forcing during extreme ranges of motion.