1. Introduction
Musculoskeletal disorders (MSDs) are a set of injuries and symptoms affecting the osteomuscular system and associated structures, such as the bones, muscles, joints, tendons, ligaments, nerves, and the circulatory system [
1]. Work-related MSDs (WMSDs) are the leading cause of occupational disabilities in industrialized countries [
2,
3]. Assessment of exposure to WMSD risk factors is regarded as a critical step in protecting workers in industries from developing WMSDs [
4].
Observational techniques that aim to evaluate risk factors are more widespread in industries [
5] because they (1) do not interfere with job processes; (2) do not require the use of expensive equipment for measuring the angular deviation of a body segment from the neutral position; (3) are user-friendly, applicable, and repeatable in various conditions; and (4) have higher validity and lower subjectivity than that had by self-reports, such as worker diaries, interviews, and questionnaires [
6,
7,
8].
While many observational methods have been developed and applied for assessing risk factors for WMSDs, the Rapid Upper Limb Assessment (RULA) [
9] has been the most frequently applied method in industries [
10,
11]. The RULA has been cited approximately 3500 times in relevant literature, which exceeds the citations of other observational methods [
12]. Some comparative studies [
13,
14,
15,
16,
17,
18,
19] claimed that the RULA might be the best system among three representative observational techniques, including the Ovako Working Posture Analysis System (OWAS) [
20], RULA, and Rapid Entire Body Assessment (REBA) [
21] for evaluating postural loads and the association with MSDs.
Many existing observational methods were mostly developed based on rankings provided by ergonomists and occupational physiotherapists or on subjective opinions by experienced workers rather than objective and consistent experimental results [
9,
16,
20,
21]. Recently, Kee [
22] developed a novel observational technique, the “Loading on the Entire Body Assessment” (LEBA), mainly based on the experimental results obtained from the studies performed by the author and other researchers. In the study, the LEBA was validated using postural load criteria, including discomfort, compressive force at L5/S1, maximum holding time and % capable at the shoulder and trunk, and epidemiological data on MSDs [
22].
A study that compares various categories of observational methods would be helpful in selecting an observational method that could most accurately quantify postural stress and estimate the association with MSDs. Contrary to the RULA, which was frequently compared with various observational methods in previous studies [
5,
13,
14,
15,
16,
17,
18,
19,
23,
24], few studies have compared the LEBA with other observational methods because of its recent development. Kee [
22] asserted in a study that the LEBA, compared to other representative observational methods, may be a more useful tool for precisely quantifying postural loads, as well as determining the corrective actions required and the association with MSDs. Therefore, a comparative study is needed to prove the superiority and validity of the LEBA. This study aimed to compare and evaluate the LEBA and RULA based on various postural load criteria and epidemiological data on MSDs. Among several observational methods, the RULA and LEBA were chosen for this study because (1) the RULA has been widely applied in industries and is known to be one of the best tools for estimating postural loads [
13,
14,
15,
16,
17,
18,
19], and (2) the LEBA has advantages, such as a high correlation with various postural load criteria and a strong relationship between the LEBA risk levels and WMSDs [
22].
RULA and LEBA
The RULA was developed by McAtamney and Corlett [
9]. It provides a quick assessment of the loading on the musculoskeletal system due to postures of the neck, trunk, and upper limbs, muscle function, and external loads exerted. Based on the grand score of its coding system, four action levels, which indicate the level of intervention required to reduce the risks of injury due to physical loading on the worker, were suggested [
9,
14]:
Action level 1: posture is acceptable if it is not maintained or repeated for long periods;
Action level 2: further investigation is needed and changes may be needed;
Action level 3: investigation and changes are required soon;
Action level 4: investigation and changes are required immediately.
The LEBA was proposed by Kee [
22]. It was based on discomfort and epidemiological data from previous research, from which posture classification and scoring systems of representative observational methods were adopted and modified. The LEBA reflects the effects of posture, external load, motion repetition, static loading, and coupling. The LEBA has more detailed posture classifications compared to representative observational techniques, such as the OWAS, RULA, and REBA. The LEBA classifies leg postures into 13 categories and subcategories, and classifies motion repetitions into five categories. In addition, three equations were provided according to the hand position for rating the effects of the external load or exertion. The degrees of the assessed harmfulness are grouped into four action categories, according to the urgency for the required workplace interventions:
Action category 1: normal postures that do not need any corrective actions;
Action category 2: postures that require further investigation and corrective changes during a subsequent regular check, but immediate intervention is unnecessary;
Action category 3: postures that require corrective actions, including redesigning workplaces or working methods, within a short time;
Action category 4: postures that require immediate consideration and corrective actions.
4. Discussion
This study compared the efficiency of the RULA, one of the best observational methods [
13,
14,
15,
16,
17], with that of the LEBA, a method that was recently developed by the author [
22]. Of the two types of validation studies for observational methods, this study assesses concurrent validity, which evaluates the extent of correlation of a method (LEBA in this study) with more valid and established ones (RULA in this study), instead of predictive validity, which evaluates the extent to which scores accurately predict an item (in this study, the association of the risks estimated by the LEBA with MSDs) [
5]. The comparisons were mainly based on various previous studies, including both types of validity studies (i.e., Kee [
13] and Kee et al. [
16,
17]: concurrent validity; Kee [
14]: predictive validity).
The results from the comparison suggest that the LEBA is a better observational method than the RULA in all categories compared in this study, including the general characteristics, risk levels, postural load criteria, association with MSDs, and intra- and inter-rater reliabilities. Although the author could not compare the LEBA with findings from a study that assessed the usability of the RULA, the LEBA exhibited high usability with “agree” or “strongly agree” (first and second highest level of five verbal anchors employed in the usability test) in the four perspectives evaluated in the usability test, such as “easy and time-effective”, “effective”, ”helpful to decide acceptance”, and “useful in establishing interventions” [
22].
The agreement rates between the LEBA and RULA were low (<60.0%) in three studies [
13,
17,
22]. The Spearman correlation coefficient between the LEBA action category and the RULA action level for the 148 WMSD cases from industries was also low (0.571) (
Table 4). These results imply that the risk assessment results by the two methods do not agree or correlate well. However, it is inferred that the LEBA may evaluate postural loads more precisely than the RULA. This is because the correlation coefficients between the postural load criteria, which included discomfort, MHTs, compressive force at L5/S1, and % capable at the shoulder and trunk, and the LEBA score were generally higher than those between the postural load criteria and the RULA grand score (
Table 3).
The logistic regression analysis revealed that the LEBA predicted the association with MSDs more accurately than the RULA, but the percentage concordant for the LEBA action category was <60% (the value for the RULA action level was 44.8%) (
Table 5). This implies that when a practitioner decides whether or not an MSD is work-related, mainly based on the LEBA action category, almost half of those decisions may not be correct. This may be attributed to several reasons [
22]. First, the RULA and LEBA do not appropriately assess work-related musculoskeletal loads, which are known to be the main risk factor for the development of MSDs [
33,
34,
35]. Second, it is difficult to estimate the association with different MSDs in various body parts that were caused by several factors using a single ergonomic method. This suggests that it may be better to use several methods to rigorously evaluate musculoskeletal loadings and accurately predict WMSDs rather than using a single best technique. Thus, unlike most existing methods that are designed for the whole body and all MSDs, new techniques should be designed for specific body parts and/or MSDs [
22]. This would enable health and safety practitioners in industries or ergonomists to assess musculoskeletal loadings more precisely, which would, in turn, increase the accuracy of predicting the association with MSDs.
While the odds ratios for the LEBA action category (2.42 and 7.00 for the LEBA action category 3 and 4, respectively) were excessively higher than those for the RULA action level (0.88 and 2.56 for the RULA action level 3 and 4, respectively), and the percentage concordant for the LEBA score (69.6%) was also excessively higher than that for the RULA grand score (52.4%), the odds ratio for the LEBA score (1.05 per 1 point) was lower than that for the RULA grand score (1.36 per 1 point) (
Table 5). This may be attributed to the wide range of the LEBA scores (range: 2–63) compared to that of the RULA grand scores (range: 1–7) used for the logistic regression analysis performed in this study.
The influencing factors of the LEBA score were slightly different from those of the RULA grand score. If the factors such as the external load, motion repetition, static loading, and coupling, which were not considered in the study by Joshi and Deshpande [
27], were excluded, the influence of the upper arm (or shoulder) on the LEBA and RULA grand scores was ranked first, and that of the trunk was ranked second. While the effects of the other factors were significant in the order of the neck, wrist, lower arm, and leg in the RULA grand score, the effects were significant in the order of the leg, elbow, neck, and wrist in the LEBA score. This difference may be attributed to the following facts: (1) the RULA was developed based on results from previous relevant studies, which can result in incorrect posture classifications without appropriately considering real postural loads, rather than the consistent experimental data used to develop the LEBA; (2) while the RULA has just two posture classification codes for the leg, the LEBA is equipped with four categories and thirteen subcategories for leg postures. The differences between the two methods may explain the inability of the RULA to properly reflect postural loads compared to the LEBA.
This study compared the RULA, which focuses on upper limb postures, and the LEBA, which was developed for evaluating whole-body postures. This might be justified based on the following studies. First, although the RULA has a significant limitation of comprising only two classifications for leg postures, many previous studies have assessed postural loads using the RULA, including even unstable lower limb postures, such as squatting and kneeling. Gómez-Galán et al. [
12], after performing a bibliometric review of 226 RULA-relevant publications between 1993 and 2019, reported that the RULA was applied to manufacturing (74 studies), human health and social work activities (38), agricultural activities, forestry and fishing (18), construction (4), and mining and quarrying (2). Kee and Karwowski [
16] also applied the RULA to iron and steel industries, and general hospitals. The above are representative industries where various unstable lower limb postures occur frequently. In principle, the RULA, with two classifications for leg postures, including balanced or unbalanced, could not be applied to the above industries. However, the aforementioned showed that the RULA has been applied to whole-body postures irrespective of leg postures. Second, while the RULA was developed for assessing upper limb postures [
9], the OWAS and the REBA were developed for evaluating whole-body postures [
20,
21]. Kee [
15] showed, based on a literature survey, that 44 journal papers dealt with the assessments of postural loads using 2 or more of the OWAS, RULA, and REBA, and that of the 44 studies, 39 adopted the RULA. Nine of the ten studies dealing with the OWAS and RULA applications revealed that the postural loads shown by the RULA were higher than those shown by the OWAS. Of the 36 studies that adopted the RULA and REBA as ergonomic risk assessment tools, 30 demonstrated that the RULA showed higher postural loads for the selected postures than the REBA. Third, several studies compared various observational techniques, including the three methods of the OWAS, RULA, and REBA, based on scales for posture classification, main functions, correspondence with valid reference, association with WMSDs, repeatability between observers, potential users, ergonomic experts’ evaluation, exposure factors assessed, postural loads, discomfort, maximum holding time, ergonomic experts’ evaluation results, etc. [
5,
7,
11,
13,
14,
15,
16,
17,
18,
19,
23,
24,
32]. The LEBA was developed for estimating whole-body postures like the OWAS and REBA. These imply that the LEBA can be compared to the RULA, a representative observational method.
The results presented in this study should be interpreted with caution, because there are so many RULA-based studies or cases [
12], but very few studies or cases are LEBA-based.
The comparisons in this study were based on general characteristics, risk levels, postural load criteria, association with MSDs, and intra- and inter-rater reliabilities. The data were obtained from laboratory experiments, as well as automotive manufacturing, automotive parts manufacturing, and construction industries. Further comparative studies using postures and MSD cases from various industries may be required to obtain more reliable comparison results.