One reviewer (TR) initially screened all retrieved records (n=2700) to exclude studies that did not meet the inclusion criteria. The remaining studies were then compiled into a structured database, documenting key bibliographic details for systematic evaluation.

Following title and abstract screening, 2 independent reviewers (TR and MZ) conducted a full-text assessment of the remaining studies to determine their eligibility. In cases of disagreement, a third reviewer provided arbitration to reach a consensus. To enhance efficiency and ensure consistency in decision-making, Rayyan software [36] was used for study selection.

All eligible studies underwent a final review before inclusion in the analysis. The overall selection process is presented as a flow diagram in the Results section.

Among the design methods identified in the reviewed studies, collaborative approaches are predominant, with multiple stakeholders participating in various stages of game development. The collaborative design approach is prominently characterized by the involvement of various stakeholders, including end users, ergonomists, and game developers. This approach aligns with the principles of design-based research, as outlined by Sanchez and Monod-Ansaldi [53]. Design-based research emphasizes a systematic yet flexible methodology that seeks to enhance educational practices through iterative analysis, design, development, and implementation, driven by collaboration among researchers and practitioners within real-world contexts.

Three articles provided very limited details regarding their design methodology, concentrating almost exclusively on the technical characteristics of the game [39,41,51]. In these cases, the perspective appears to be more oriented toward demonstrating technological feasibility than toward adopting a participatory approach that incorporates the needs of end users. While such approaches may serve as exploratory or experimental objectives, they do not provide clear insight into how learning objectives, user needs, engagement strategies, and evaluation criteria were defined and integrated into the game, thereby limiting the understanding of whether the intended educational or behavioral outcomes were achieved.

Although not all studies explicitly detail their design methodologies, a common practice observed is the involvement of users in varying degrees and stages of the design process with an iterative approach. Several studies [39,43,45,49] have documented user involvement in the early stages of the development process. For example, Lanzotti et al [43] engaged users in shaping various design aspects of the game, such as character perspective, avatar visualization, and graphical detail. This user feedback, gathered from a sample of approximately 10 participants, guided design decisions. The optimal configuration was determined through a 10-point evaluation process. On the other hand, Rodrigues et al [48] and Idriss et al [39] included users primarily during the testing phase to evaluate whether the serious game met its basic functional requirements. However, their focus did not extend to aspects such as user experience or broader usability considerations.

The methods used to engage users in the design phase also exhibit notable diversity, with some studies using virtual simulations and structured feedback mechanisms. For example, Sisto et al [49] used a virtual simulation of decontextualized activities within the game that are conceptually similar to real-world work tasks. Ergonomics experts and game developers collaborated closely to define these game scenarios, using insights drawn from field data related to actual operational activities to ensure that the simulated activities effectively represent the types of tasks encountered in real work environments. Rapp et al [45] similarly integrated users into their game development process through a combination of semistructured interviews and web-based questionnaires. The data obtained helped specify the educational objectives of the game and refine the initial design elements.

While most included studies incorporate user input at specific stages, 3 studies integrate stakeholders throughout the entire development process, offering a more comprehensive co-design approach. Pietrafesa et al [44] used an advanced collaborative design approach that used both qualitative and quantitative methods. Users were engaged actively across various phases of the game’s development, including through surveys, focus groups, questionnaires, and discussions. User testing was conducted toward the end of the development cycle. Particularly, the authors sought input from both primary users and developers involved in the design process. Similarly, Kuipers et al [47] and Rebelo and Filgueiras [46] adopted an iterative design method grounded in collaboration between prospective users and experts. This method involved distinct stages: defining requirements, setting objectives and selecting technology, constructing prototypes, deploying the prototypes, evaluating them with target users, and conducting a comprehensive review by an expert panel.

Overall, the sophistication of the design methodologies appears to correlate with the effectiveness of the developed games. Games developed through advanced co-design methodologies often exhibit greater sophistication and functionality. Many researchers emphasize the importance of multidisciplinary and participatory collaboration in the development of SGs [53-55]. Hulin [56] underscores the specifics of the co-design approach within a research-action framework, introducing a novel method for analyzing design processes through stratigraphic modeling of experiences. Additionally, involving end users as consultants and testers—focusing their feedback on game dynamics—has been shown to enhance the game impact [57,58].

To understand the technological landscape of SGs for MSD prevention, we examined the platforms and control mechanisms used in the reviewed studies. Among the reviewed studies, a variety of platforms and control mechanisms for SGs are evident. Four games were designed for personal computers (PCs) and use keyboard and mouse controls [42-45], while others apply haptic controllers [48] and cameras [41]. Three games were developed for smartphones and tablets [42,44,50]. Several games used motion-capture technologies such as Microsoft Kinect [39,47,48,52], inertial measurement unit (IMU) [51], Leap Motion [49], or similar tools like Vive Tracker [40]. One study used a more advanced motion capture system combining a reflective-marker suit, infrared cameras, and surface electromyography electrodes [38]. In some cases, the games used virtual reality (VR) devices [40,49] alongside motion-capture technologies, with the latter often serving as measurement tools. It is noteworthy that all the reviewed SGs were digital in format [59].

Sisto et al [49] used a combination of motion-capture technologies (Microsoft Kinect and Leap Motion) within a VR game. Their approach aimed to capture subtle hand movements with Leap Motion while using Microsoft Kinect to assess the positions of various body parts, including the back, arms, and neck. Similarly, Stranick and Lopez [52] used a VR headset in order to immerse the user in the game environment, leveraging Microsoft Kinect to adopt the expected posture for stretching exercises. Rodrigues et al [48] used multiple gaming peripherals, incorporating gestural control via Microsoft Kinect, and haptic devices. Their game architecture featured segments that are played on a PC with haptic controllers, while other segments used Kinect to detect when players needed breaks and to suggest exercises.

Some games, such as those developed by Greuter and Tepe [42] and Pietrafesa et al [44], privileged minimal hardware, designed specifically for tablets, smartphones, and PCs. The motivation for choosing minimal hardware requirements in game development often revolved around increasing accessibility, particularly for classroom settings or enabling students to use the game on their smartphones beyond class hours. However, the rationale behind specific sophisticated technology choices, such as VR, or haptic controllers, is rarely detailed in the reviewed studies. For example, Rodrigues et al [48] used haptic devices in their game, but the study lacks a clear justification for this choice, missing a clear explanation of the benefits and limitations of using haptic technology.

Games designed for rehabilitation, or the acquisition of physical skills, predominantly used motion capture technologies [38-40,47-49]. The choice of technology in these cases is closely tied to its functional utility within the game's context. However, the implementation of such technologies can introduce specific constraints. For example, in Valdivia et al [51], although the use of an IMU allows for faster sampling and more accurate capture, it is invasive and requires the expertise of specialists. Similarly, Stranick and Lopez [52] and Sisto et al [49] reported notable technical limitations associated with the use of VR technology. In Stranick and Lopez [52], the Microsoft Kinect sometimes failed to capture player movements in time. In Sisto et al [49], players had to keep their hands within their field of vision to avoid losing in-game items, and the VR system occasionally demonstrated suboptimal fluidity, requiring players to repeat certain actions. Additionally, the setup of the VR system was time-consuming and required the expertise of a VR specialist to address potential system bugs and ensure the game's proper functioning.

To clearly define the scope of gameplay in the reviewed studies, it is necessary to distinguish between its core components and related elements such as narrative scenarios. Here, “gameplay” is defined as the interactive and functional aspects of a game, encompassing its rules, mechanics, objectives, and the overall player experience. In contrast, a “scenario” refers to the narrative or storyline that provides the contextual framework for the game and structures the sequence of events during gameplay [59]. A single gameplay framework can support multiple scenarios.

The reviewed studies reveal a wide range of formats, from very simple structures to more elaborate ones. A first category of gameplay includes short and simple mini-games, based on basic interactions and often limited to a single task or objective [41,51,52]. A second category consists of relatively simple games that incorporate a progression or performance tracking system, allowing for a sense of continuity in the player experience [43,49,50]. A third configuration involves games composed of multiple successive mini-games, designed to diversify situations and gaming experiences [38-40,48]. The last category of gameplay consists of a more complex structure, combining various mechanics, levels of games, and integrating more advanced progression systems [42,44,45,47]. This diversity in gameplay reflects a variety of pedagogical intentions, as well as technical and contextual constraints specific to each project.

Several of the developed games show similarities in their underlying concepts and rules, while their distinctions appear at the level of narrative elements, artistic attributes, and control mechanisms. For instance, France and Thomas [40] and Jansen-Kosterink et al [38] both crafted memory-based games characterized by identical rule sets but differentiated by graphical elements and control mechanisms. Moreover, we observed games of similar genres, especially within the risk management category, within those designed for PCs [42,44,45].

In some PC-based games, player interaction extends beyond individual actions to managing environments and nonplayer characters (NPCs), adding a strategic component to gameplay. For example, in Pietrafesa et al [44], players are responsible for maintaining a safe work environment by overseeing NPCs' tasks and protecting them from workplace hazards. The player's success is tied to the productivity of the NPCs, which is affected by the implemented safety measures. Greuter and Tepe [42] require players to identify and address hazardous conditions affecting NPCs' construction work. Players mitigate these risks by providing necessary equipment and adjusting the work environment, enabling NPCs to continue their tasks. In Rapp et al [45], a multiplayer game simulates restaurant management, where players oversee various aspects such as personnel recruitment, budget management, and employee welfare. The game includes a debriefing phase aimed at enhancing learning outcomes, although details on the gameplay mechanics were not extensively provided in the study.

Some studies incorporated features designed to enhance player motivation and engagement, such as competition and collaboration [42-44,47]. Lanzotti et al [43] considered incorporating a shared scoreboard to introduce competition as a motivational element. Pietrafesa et al [44] added a competitive component by having 2 groups of students compete against each other, though the study does not provide detailed information on the conditions of this competition. Greuter and Tepe [42] and Kuiper et al [47] organized weekly group gaming sessions to facilitate testing, but this collaborative aspect was not further examined in the study.

Another critical aspect of gameplay complexity is the extent to which games integrate dynamic feedback and progression systems to reinforce learning objectives. Stranick and Lopez [52] incorporate elements such as achievements and ranking in their game, which are clear feedback and progression mechanisms for the player. Sisto et al [49] integrate task completion time and a posture risk assessment score as core gameplay elements. These scores are calculated using an algorithm that factors in exposure time and joint angles within the simulated environment. At the end of each game level, players can view their risk scores for awkward postures alongside their time performance scores, enabling them to refine their strategies for task execution within the game. In contrast, the game by Kuipers et al [47] is notably more sophisticated, incorporating elements of customization, scoring, and progression. Players are encouraged to explore the in-game world by visiting specific locations that trigger mini-games, which allow them to personalize the virtual environment. The primary objective of this game is to induce behavioral changes, motivating players to experiment with new behaviors while confronting challenging tasks and objectives. Throughout the gameplay, players receive precise and tailored feedback to ensure compliance with safety and in-game rules. Both the games developed by Kuipers et al [47] and Sisto et al [49] incorporate in-game experiences and dashboards with data visualization to support players in understanding and interpreting the information presented during gameplay.

Interestingly, despite the emphasis on interactivity, none of the studies, except for Greuter and Tepe [42], report considering audio elements and music in the game design which could influence player immersion and experience. In their study, participants provided negative feedback about the sound environment, which the authors hypothesized was due to all 20 games being played simultaneously in the same room, creating an overwhelming auditory experience. The authors suggest that the use of headphones might have mitigated this issue, although it could have also impeded player collaboration. Table 1 summarizes the technologies, gameplay, and expected outcomes used in the reviewed studies.

Defining the expected outcomes and learning objectives of SGs is crucial, as they specify the skills, knowledge, and competencies that developers aim for users to acquire. For SGs designed for the rehabilitation of patients with work-related MSDs [38-41], the scenarios are structured around movements and postures critical for recovery or training. The expected outcome is to engage patients in performing rehabilitation exercises through play scenarios and to help them overcome fear of injury through interactive participation. An example of a rehabilitation-focused SG is the VR-based intervention developed by France and Thomas [40], designed to gradually improve lumbar flexion. While the game may not directly influence psychological factors for behavior change, it serves as a distraction from pain, reinforcing movement by increasing expectations of lumbar flexion. This approach helps individuals cope with movement-related fear, promoting the understanding that lumbar flexion is not inherently harmful, particularly for those with chronic low back pain and pain-related fear.

The learning outcomes of the games addressing occupational health and safety [42-45] revolve around raising awareness and facilitating learning. These studies typically focus on the risks of work-related accidents, particularly safety procedures and equipment. For instance, the game developed by Greuter and Tepe [42] aligns with occupational health and safety training objectives, aiming to educate users on hazard identification, control measures, communication processes, and reporting procedures. However, the study did not assess whether playing the game improves students' ability to identify occupational hazards. Moreover, Rapp et al [45] expanded this approach by integrating broader workplace learning objectives, aiming to enhance leadership, communication, and safety culture. For example, they aim to enhance knowledge by involving employees in operational decision-making, develop skills by allowing players to explore various actions with feedback, and promote awareness and attitude shifts toward workplace health and safety by emphasizing its importance for both employee well-being and organizational success. However, the authors did not provide any results demonstrating the achievement of these objectives through their game.

Finally, a distinct category of SGs focuses on MSD prevention by training professionals to adopt proper postures and ergonomic behaviors [46-52]. Rodrigues et al [48] developed 3 complementary mini-games: one is designed to raise awareness about the risks associated with static desk postures, while the other two encourage players to engage in regular stretching exercises. Kuipers et al [47] and Sisto et al [49] designed games aimed at promoting behavioral changes, providing tailored feedback to support the effective adoption of new ergonomic practices. Within this category, three studies [50-52] specifically used exergames to promote stretching and movement exercises. Intipanya et al [50] created a smartphone-based exergame where head movements control an aircraft, targeting cervical pain prevention. Valdivia et al [51] used Kinect and IMU sensors to prompt office workers to perform bending and extension movements through a room painting task. Stranick and Lopez [52] designed a VR-based exergame where players reproduce postures to overcome obstacles, promoting stretching exercises among industrial workers. The exergames reviewed primarily focus on the correct execution of physical exercises, with short-term learning objectives centered on performance. Players’ cognitive involvement remains limited: few games encourage understanding the purpose of movements, reflecting on posture, or transferring learning beyond the game context. Moreover, game mechanics often emphasize repetition or movement accuracy without providing elements that help players understand their relevance or their connection to MSD prevention. The game included in Kuipers et al [47], however, goes further by promoting behavioral change among health care workers, particularly in patient handling practices.

According to Gris and Bengston [60], the assessment of game-based learning effectiveness revolves around the key components including engagement, user experience, usability, and educational impact. The evaluation uses a combination of quantitative and qualitative measures for assessing educational impact, with qualitative methods being more commonly used for evaluating engagement, user experience, and usability [60]. In medical literature, another prevalent approach is to assess a game's ability to facilitate skill transfer [61]. The following section examines how different included studies evaluated various components of game effectiveness.

The assessment of engagement, usability, and user experience is evident in 10 of the studies reviewed [38,39,41-43,46-48,51,52], although these evaluations exhibit variations in the questionnaires, variables, and methodologies used. Among these studies, some focused more specifically on user experience, using distinct evaluation methods and tools. Greuter and Tepe [42] evaluated participants' user experience by a questionnaire across various dimensions, including fun, engagement, success, control, motivation, feedback, usability, and difficulty [62]. Additionally, semistructured interviews were conducted following the test sessions to gather further insights. While the authors reported satisfactory levels of motivation, engagement, and learning among participants, with most students indicating increased motivation as they progressed through the game, several limitations in the study's design should be acknowledged. These include a small sample size, the subjective nature of the evaluation (which relied solely on self-reported questionnaires from a single group), and the absence of a longitudinal follow-up period.

Similarly, Jansen-Kosenterink et al [38] adopted a multifaceted approach to evaluate user experience, incorporating standardized and customized assessment tools. In their study, user experience was evaluated using multiple metrics: usability was measured with the System Usability Scale (SUS) [63], satisfaction was gauged through 2 specific questions, motivation was assessed using 2 additional questions, and gaming experience was evaluated with the Core Elements of Gaming Experience Questionnaire [64]. All 10 participants in the study reported a positive user experience, indicating that the game effectively distracted them from pain and enhanced their focus on gameplay. This study addressed the longitudinal aspects by tracking participants over a 4-week period of gameplay. However, the study's limitations include the absence of a control group and a relatively small sample size. Similarly, Husna et al [41] used the SUS and the User Experience Questionnaire to assess their game, although the choice of these tools is questionable given the simplicity of the prototype, in which a character is guided through a maze using hand gestures captured via video. The minimal graphics and interactive complexity of the system raise concerns about the actual relevance of evaluating usability and user experience in this context.

Some studies focused more on usability assessment, with Lanzotti et al [43] evaluating their SG based on effectiveness, efficiency, and satisfaction. Effectiveness and efficiency were measured using quantitative metrics, including interaction errors, incorrect movements, gameplay duration, and task completion rates, while satisfaction was assessed through a subjective questionnaire. The usability index was calculated as the average of these 3 factors. The study involved 2 groups of 6 operators: the case group, which used an SG for training on mould press safety procedures, and the control group, which completed a written test with a similar structure. Although the test group reported higher satisfaction, their overall usability index was lower, particularly in correctly sequencing commands for safe machine operation. Interpretation of the results was challenging due to the small sample size and the lack of information on participants' baseline knowledge levels. The authors suggested that the written test may have assessed theoretical knowledge, whereas the SG required practical application, potentially leading to discrepancies. Additionally, no weighting was applied to usability factors in calculating the index, and the questionnaire lacked details on reliability and validity. Furthermore, the study did not specify how variables such as error frequency were quantified, particularly for the control group, limiting the clarity of the findings.

In addition to evaluating usability, some studies also explored user preferences and subjective satisfaction. Rodrigues et al [48] assessed user preferences among 20 participants, divided into 2 age groups, focusing on 5 key aspects: game relevance, ease of use, effectiveness of stretching exercises, perceived health (before and after the session), and overall satisfaction. Participants rated each aspect on a 3-point scale (low, medium, and very good). Most participants rated the game as “very good” or “medium” across the 5 categories. However, fewer than 40% rated the effectiveness of stretching exercises as “very good,” and only about half gave the highest rating for overall satisfaction. This suggests that while participants found the game engaging, its perceived effectiveness in promoting stretching exercises was more variable. A key limitation of this study is the lack of details on the validity, reliability, and development of the assessment questionnaire. The absence of such information raises concerns about the robustness of the evaluation method, as it is unclear whether the tool effectively captured user perceptions and game impact.

Other studies further investigated usability and acceptability through comparative analysis between healthy participants and those with MSD symptoms. In that respect, Idriss et al [39] conducted an evaluation of the usability and acceptability of a serious game that featured 2 distinct scenarios: football and object manipulation. Feedback was collected through a questionnaire. The study also involved a comparison of game scores between 2 groups: healthy participants and those with MSD symptoms. The results indicated that healthy participants achieved significantly higher scores than those with MSD symptoms. Participants were asked to rate various parameters, including the clarity of objectives, level of difficulty, and the attractiveness of the game environment, using a 5-point Likert scale. Both groups found the games to be motivational, attractive, and challenging. However, the study did not include a comparative analysis of the usability results between the 2 groups. Additionally, the questionnaire used for evaluation had not been validated for reliability or validity, raising concerns about whether it adequately captured all relevant aspects of usability, engagement, and playability in the games.

Finally, two studies focused on the assessment of engagement. Valdivia et al [51] mainly assessed exergame’s engagement skills through the Game Engagement Questionnaire, which measures dimensions such as immersion, focus, and sense of control. This evaluation followed two brief 1-minute exercise sessions in which participants performed trunk flexion and extension movements using 2 versions of the exergame—one based on the Kinect V2 sensor and the other using an IMU. Despite the use of a standardized questionnaire, such short-duration exposure with a small subject raises questions about the ability of the sessions to foster meaningful engagement or allow users to fully grasp the potential of the system. While most studies relied on self-reported measures, Kuipers et al [47] introduced a different approach by quantifying engagement through voluntary re-engagement metrics. They used a quantitative measure for evaluating player engagement, where the number of game sessions voluntarily conducted by players over a given period served as an indicator of their engagement. This approach is analogous to the voluntary reengagement observed in Sanchez et al [65], where players' self-initiated participation in the task was used as a measure of motivation. During the experimentation phase, participants were given free time in which they could choose to replay the game at their discretion. The decision of a participant to replay the game independently was interpreted as a sign of motivation. However, Sanchez et al [65] emphasized that this measure is more closely associated with intrinsic motivation rather than engagement. This distinction between the two concepts should be considered in the interpretation of results.

In summary, in line with Gris and Bengston [60], the assessment of engagement, usability, and user experience in SGs predominantly relies on qualitative measures. Moreover, the evaluation tools used in these studies are generally nonstandardized. Among the studies reviewed that assess these concepts, only three [38,41,51] used a standardized questionnaire. Furthermore, there are notable discrepancies in the concepts and variables measured across different studies. The concept of usability, for instance, is variably defined depending on the study. For example, in the study of Lanzotti et al [43], usability encompasses a range of quantitative and qualitative measures, focusing on the dimensions of effectiveness, efficiency, and satisfaction.

Assessing the educational impact of SGs requires robust evaluation methodologies to determine their effectiveness in training and behavior change. However, many studies fail to implement comprehensive assessment processes, relying instead on limited or inconsistent evaluation methods. Among the studies included in the analysis, many lack a comprehensive process for evaluating the impact of developed SGs on training or achieving targeted outcomes. Three reviewed studies [44,45,49] do not provide any evaluation of the game's impact on learning, while others rely solely on postgame questionnaires without including a control group or had significant limitations. For instance, Lanzotti et al [43] assessed the efficacy of game-based learning by comparing in-game metrics with responses from a control group to a multiple-choice questionnaire addressing the same simulated process. However, the level of training received by the control group was not clearly defined, and the results indicated that the control group outperformed the experimental group, suggesting the serious game's effectiveness was limited. Similarly, Greuter and Tepe [42] evaluated the game's effectiveness through a written questionnaire on covered concepts. Since the game aimed to enhance hazard identification skills, effectiveness was measured through a written test requiring participants to identify risks from photographs. Educational impacts were assessed by comparing pretest and posttest questionnaire results of the experimental group. Although the authors mentioned the presence of a control group, results between the two groups were not provided, and the overall study design lacked robustness due to the use of a validated questionnaire and postsession data only.

However, 3 studies implemented game evaluations through prolonged test sessions involving multiple groups and clear, comparable measurement variables. Kuipers et al [47] measured the evolution of participants' game scores based on the number and duration of sessions over a period of 3 weeks. They observed that the more players played, the higher the scores increased. The increase in scores thus reflects the effectiveness of the game: players autonomously exhibit the desired behavior. The authors posited that performing in-game movements mirrors actual workplace lifting techniques, thereby increasing game scores leading to training outcomes aligning with desired behavioral changes in lifting techniques. Nevertheless, further validation is required through longitudinal studies assessing behavioral changes in real workplace settings.

Similarly, Jansen-Kosterink et al [38] evaluated learning outcomes by analyzing motor skill improvements and clinical indicators before and after gameplay sessions. The game was administered to a cohort of patients afflicted with chronic neck and back pain, who engaged with the game over a 4-week duration. Among the parameters measured, including walking velocity, reach height, cervical range of motion, pain intensity, and pain disability, only cervical range of motion exhibited a significant increase over the testing period. Although the authors asserted that patients generally exhibited faster walking speeds, increased reach, and enhanced neck mobility after 4 weeks of gameplay, these results did not reach statistical significance, likely due to the limited sample size and the short duration of the testing period.

France and Thomas [40] conducted a 9-week clinical trial to evaluate their VR-based rehabilitation game for pain management and functional recovery. They assessed patient-reported disability, pain levels, spinal condition, and pain expectations, with follow-up monitoring over 48 weeks. Results showed that the experimental group experienced greater pain and disability reduction, along with decreased pain expectancy and improved lumbar flexion compared to the control group. These findings suggest that the VR game may help individuals with low back pain and movement-related fear engage in repeated sessions, promoting graded spinal motion.