Ergonomic evaluation,with the RULA method,of greenhouse tasks of trellising crops

Abstract

BACKGROUND: The present study uses the RULA method to define the most favourable working conditions to lower climbing greenhouse plants, the most problematic task involved in trellising crops.

OBJECTIVE: The aim is to prevent potential musculoskeletal disorders to the upper limbs in workers performing this task.

METHODS: The variety of scenarios involving this task was simulated in the laboratory to evaluate different working heights, crop weights, and types of guides handled.

RESULTS AND CONCLUSIONS: From the 2,544 tests made, only some ergonomically favourable conditions were found to be met when the work was done at a height of 1.4 m, although heights of between 1.2 and 1.6 m could prove acceptable. Furthermore, the study shows the importance of the type of hanger needed to perform the work correctly, of the group of hangers tested; ergonomically acceptable handling was possible only with the new-generation ones having the lock device. In the case of using other types of hangers, it is recommended not to support weights greater than 2 kg. In addition to the above recommendations, it is considered vital to guarantee the correct organization of the workday to include recovery times from the earliest working hours, given the repetitive nature of the work, as well as to promote training sessions to avoid incorrect postures, especially regarding the limbs, trunk, wrist, and neck.

Keywords

Agricultural ergonomics risk prevention repetitive movements musculoskeletal disorders MSDs

1 Introduction

A technique regularly practiced in climbing or sprawling plants protected under plastic is to maintain the plant upright and to keep the leaves and especially the fruits from touching the soil, thereby improving the general aeration of the plant, favouring sunlight reception, and facilitating cultivation tasks (pruning, harvesting, etc.). Overall, this leads to healthier plants, higher yields, better quality fruit, and an easier harvest.

Plants are usually trained on polypropylene twine guides attached on one end to the basal zone of the stem (tied, with rings, etc.) and on the other end to a wire trellis [1]. As the plant grows, it entwines its branches in the guides or is tied to the guides with rings, plant tape or any other attachment system until it reaches the wire trellis. As it grows, the plant is lowered by loosening the twine guides. In greenhouses of south-eastern Spain, this work is done with an attachment mechanism called “hangers or hooks”, which consist of having twine rolled on crop hangers, this being a simple way to lower the plant as it continues to grow. The lowering of the crop is ergonomically the most problematic task in the training of greenhouse crops, given that it is completely manual work; it requires a certain degree of physical strength, given that the plants in full production can weigh up to 15 kg per square meter of floor surface [2]; it is repetitious, as all the plants in the greenhouse must be attended, normally 1 to 2 plants per square meter [3]; and the process must be repeated up to 8 times over a crop cycle (9 months). So, each worker over a workday may lower hundreds of plants, each one with a suspended weight up to 15 kg (Fig. 1). All of this submits the worker’s upper limbs (wrist, arm, shoulder) to awkward postures. The repeated adoption of uncomfortable postures generates fatigue and over the long term can provoke musculoskeletal disorders (MSDs), such as wrist and shoulder tendonitis, carpal-tunnel syndrome, lateral epicondylitis. MSDs are considered one of the greatest occupational health topics today [4] and, in particular, upper-limb MSDs are currently the commonest form of occupational disorder in Europe [5]. Physical ergonomic factors such as the combination of strained working postures [6 –11], heavy weight lifting [6 , 11–17], the handling of manual materials [18], prolonged periods of standing [19 –21], long working hours per shift [22, 23], repetitive movements [6 , 25] and poor working conditions [10 , 26–29] are associated with MSDs. All these factors are characteristic of the task of lowering crops in greenhouses.

In agriculture there have been a lot of improvements over the years in technology, working methods, etc., however few studies are available on improving the working conditions for workers of the horticultural sector. Notable studies on working conditions, workers’ physical and mental health, and skill-learning potential, have been conducted in Sweden [30, 31], as well as some in south-eastern Spain. In this sense, using a general ergonomic-assessment model, LEST method [32], potential problems in greenhouses have been detected in relation to heat stress, psychosocial risks, and physical load [26, 33]. The typical south-eastern Spanish greenhouses were found not to guarantee workers a comfortable working environment. The type of greenhouse and crop proved to directly affect the ergonomic-psychosocial conditions of the workers [33], and the authors concluded that work needs to be appropriately organized and that further training of workers and employers is needed to improve risk problems. Research has also been undertaken on the working conditions outside greenhouses to help prevent risks to construction workers building Almeria-type greenhouses [34, 35] and also on industrial greenhouses [36, 37]. In addition, the conditions of heat stress have been studied to improve the organizational ergonomics of agricultural companies [27, 38]. Regarding physical load in greenhouse, the risk of MSDs due to repetitive work has been assessed for workers employed in the sorting phase in greenhouses, using the ergonomic “OCRA index” [39], revealing that there is a high risk for workers and recommending that the work be executed with automatically calibrated machinery in order to avoid potential MSDs. Also, this method has been used to assess the risk of MSDs for workers employed in manual grafting in greenhouses [40], specifying that the main risk factors are: The frequency of movement, posture, and repetition of cycles. In general, all the above authors concur that greenhouse workers are surrounded by potential health risks. Furthermore, it should be taken into account that roughly three million people work in greenhouses worldwide, of which some 45,000 work in Almeria (Spain) [26].

One of the universally recognized assessment methods [41 –43], the RULA (Rapid Upper Limb Assessment method), was developed for use in ergonomics investigations in workplaces where work-related upper-limb disorders are reported [44]. The method provides an overall score that takes into account postural load on the entire body with particular attention to the neck, trunk, shoulders, arms, and wrists. The overall score also takes into account the force used and the repetitiveness of the movement. Several studies have validated the evaluation provided by ergonomics methods. [45], studying musculoskeletal disorders in different factories with six different biomechanical methods, concluded that the real risk conditions were consistently confirmed by the RULA method. Similarly, an examination has been made for the agreement between 5 ergonomic risk-assessment methods calculated on the basis of quantitative exposure measures (surface electromyography) concerning the exposures of sawmill workers performing 4 repetitive jobs [46, 47], stating that the RULA was one of the best biomechanical methods. The RULA was also used to study ergonomic risk factors among female sewing-machine operators in Turkey [48], noting that it provided a quick and effective evaluation. Other authors [42], however, hold that it provides a quite rigid evaluation. The method has also been used in the agricultural sector to assess the pricking task [41], revealed that the pricking task with existing methods (fork and hand tool) requires immediate investigation and changes in working posture. Moreover, the method has been used to evaluate upper-limb stress of workers operating a prototype of a tubing-type grafting robot for fruits and vegetables [49], offering success in validating the prototype and making improvement proposals. Similarly, [4] used the RULA to evaluate the prevalence of ergonomic risk levels and to identify ergonomic factors related to lower-back pain in rubber tappers, recommending the development and implementation of a programme using ergonomic concepts to reduce the pain. Recently [28], showed the need to characterize the ergonomic conditions of greenhouse tasks in order to relate psychosocial risks with MSDs.

The present work uses the RULA method to evaluate the ergonomic conditions of a simulation of lowering crops to identify the most appropriate working conditions and thereby prevent potential MSDs.

2 Material and methods

2.1 Participants

The test involved 30 volunteers who gave their written consent before participating. None of the participants had a history of physical disorders, musculoskeletal ailments, or neuromuscular problems. The experiment was approved by the Department of Engineering of the University of Almeria (Spain).

2.2 Crop hangers tested

The crop hangers regularly used in Spanish Mediterranean coastal agriculture can be classified into two broad groups:

Traditional hangers, consisting of a wire or plastic hanger around which the training twine is wound.

New-generation hangers, consisting of a spool with the twine attached to the body of the hanger. This group of hangers can in turn be classified into two groups, according to whether they have a lock system to stop the training twine.

In consideration of the variety of the hangers sold on the Spanish Mediterranean coast, the present study used six different models of hangers (Fig. 2), two being traditional (hanger 1 and 2) and the other four being considered new-generation models. One hanger (3) lacks a lock device, while the others have a stop (4 and 6) or a drum lock (5).

2.3 Equipment used

The experimental tests required the design of a support structure for the hangers. The structure was designed to work at different heights. For this, the structure was composed of two metal posts 2.25 m apart, connected with a horizontal bar, of adjustable height, onto which the different crop hangers were attached. Sony model Carl Zeiss Vario-Tessar video cameras provided continuous filming to be used afterwards in measuring the posture angles of the worker.

2.4 The RULA method

The method known as the RULA [44] divides the human body into two groups: Group A, which includes the upper limbs, i.e. arms, forearms, and wrists; and Group B, composed of the legs, trunk, and neck. By tables associated with the method, each body zone (legs, wrists, arms, trunk, etc.) is scored and, according to the scores, overall values are assigned to each group (A and B). Then the overall scores of each group are weighted as a function of the type of muscular activity undertaken (static, repetitive, occasional), and of the force applied during the execution of the task. Several groups were differentiated according to whether the load was less than 2 kg, between 2 and 10 kg, or higher than this latter value. The present study was made with loads of 1, 5, and 10 kg, allowing analyses of the different portions of load differentiated by the method.

Based on the overall weighted values (C scoring for the upper limbs, and D for the other limbs), a final score was established. The final value reached with the RULA method is proportional to the level of action involved in the task, so that the high values indicate greater risk that MSDs will appear. The method organizes the final scores into action levels that orient the evaluator towards decisions to make after the analysis. The levels of action fluctuate from level 1, which evaluates the posture as acceptable; level 4, which indicates the urgent need to change the activity. An action level of 2 could indicate the need for changes in the job, whereas an activity level of 3 would require such changes.

2.5 Experimental design

For the aim of the study, the following variables were analysed:

Crop-hanger type: 6 models;

Working height: 1.2, 1.4, 1.6, 1.8, and 2.0 m;

Weight handled: 1, 5, and 10 kg;

Worker height: 30 persons;

Corresponding variables characterizing the workers: gender, age, height, and body-mass index (hereafter BMI).

2.6 Work methodology

For a definition of the most favourable ergonomic conditions for lowering crops, the field working conditions were simulated in the laboratory, specifying the work cycles comprising the task and studying the most extreme work postures adopted by each of the workers, for each of the work cycles.

Prior to the tests, all the participants (Table 1) were trained in the use of each hanger, each taking the time necessary to learn the system (c. 15 min on average) until succeeding in adopting a working posture considered by each participant to be the most appropriate one. Next, the tests were administered individually. The tests began at a height of 1.2 m with a weight of 1 kg, using each of the different hangers. The task consisted of releasing the training twine until a counterweight reached the soil, simulating the lowering of a crop plant. After finishing the lowering at 1.2 m with the 6 hangers tested at 1 kg, the procedure was repeated for weights of 5 kg and afterwards with 10 kg. Then, the complete test cycle was repeated for the other working heights (1.4, 1.6, 1.8, and 2.0 m) (Fig. 3).

The entire test was filmed with continuous videos from two opposite recording planes, both being normal to the worker. Afterwards, the images were processed in the laboratory for slow motion in order to measure the most extreme posture angles, thereby evaluating the working posture.

2.7 Statistical analysis

2.7.1 Characterization of the workers

The group of workers were categorized on the basis of the descriptive statistical analysis of the variables gender, age, height, and BMI, studying their distribution by frequency. The variables worker height and BMI were defined by sections, as indicated in Table 1.

2.7.2 Analysis of the evaluations made with the RULA method

First, the entry data were analysed to identify absent and lost data. Next, the frequency distribution of the evaluations made by applying the method was analysed and it was confirmed that the data did not verify the conditions of independence, homoscedasticity, or normality of the variables, and therefore, Pearson’s chi-square test was applied to analyse the independence of the variables, and a contingency analysis was used to study the dependency relations detected. The study was completed with a simple correspondence analysis to determine to what degree the different categories of variables contributed to this relation. Finally, the work was completed with multiple-correspondence analyses to determine what categories of each variable correspond to each other in order to detect significant correspondences between the categories differentiated in the variable “level of action” with respect to the other differentiated categories for the remaining variables analysed.

The statistical analyses were conducted using the SPSS version 20.0 with a significance level set at p < 0.05 for all tests.

3 Results

3.1 Characterization of the workers

Qualitative variables collected for each worker: Gender, age, height and BMI, are summarized in Table 1.

No statistically significant dependency relation was detected between the variables characterizing the workers (gender, age, height or BMI), with respect to the rest of the variables analysed.

3.2 Characterization of the evaluations made

Table 2 shows the frequency distribution of the results found for the application of the method. In addition, the overall mean values registered for each parameter analysed is indicated, as well as the standard deviation (SD).

The position of the arm reached a mean value of 3.22±1.12, indicating that on average it bent more than 90°, and the frequency analysis indicated that in 11.3% of the tests the scores increased, given that the arms were in a rotated or abducted position. The mean score of the forearm was1.36±0.48, indicating that in some workers, their position was inappropriate, given the bending angles of less than 60° or greater than 100°. The wrist position adopted on average was quite neutral (1.12±0.32), so that in only 11.8% of the cases were there slight flexions or extensions, in all cases with an angle of less than 15°. For the wrist, pronation and supination were detected in an extreme range in 15.6% of the tests.

Given that, to complete the task of lowering a single crop, the cycle was repeated more than 4 times per min, both the scoring of Group A as well as that of Group B, was lumped for its repetitive nature, an aspect reflected in the columns corresponding to muscle activity. In reference to the force applied, the mean overall values were very low (0.82±1.2), because in 67.2% of the tests the load was less than 2 kg (test weight 1 kg) or else the new-generation hangers were used with a lock mechanism (hangers 4, 5, and 6), which can be handled without the need to bear the weight of the plant. In general, the weighted mean scoring of Group A (scoring C) was 5.18±1.83, higher than the mean value established by method (4), indicating that on average the position adopted by the upper limbs was quite strained and in some cases extremely so.

In reference to the position adopted by the neck, it was found that in 54.4% of the tests the position was neutral, with slight bending, between 10 and 20°, in 23.5% of the cases, as well as the presence of extensions in 22.2% of the cases. In the trunk, the posture was correct in 86.8% of the tests, with only slight bending (<20°) in the remaining cases (13.2%), in any case, without turning or lateral leaning of the trunk. Finally, the posture of the legs was appropriate in all cases, given that the weight was symmetrically distributed. In general, the weighted mean scoring of Group B (scoring D) was 3.98±1.99, reaching a medium position, given that the evaluations fluctuated between 1 and 7+, the latter value being more the limit.

Finally, the levels of activity fluctuated from 1 to 4. The mean values recorded were 2.85±0.93. In 0.3% of the tests, the activity level was 1, i.e. 51.4% the activity level of 2, 11.6% the activity level of 3, and 36.8% the activity level of 4.

3.3 Contingency tables

Below, the contingency tables are used to analyse the relation between the variable “level of activity” with respect to the variables “working height”, “hanger type”, and “crop-plant weight”. Thus, Table 3 shows the relation between the variable “working height” and “activity level”.

The results for Pearson’s chi-square test indicate that the categorical variable “working height” is related to “activity level” at a significance level of 0.05.

The analysis of the cases showed that the frequency registered for the working heights of 1.2, 1.4, and 1.6 m are practically identical. The working heights of 1.8 m implied a decrease in the frequency of the cases with an activity level of 2, so that the number of cases with the activity level of 3 and 4 (55.7% of the tests) increased, with the great majority of the cases concentrated at an activity level of 4 (93.3%) for heights of 2 m.

The marginal-frequency analysis showed that the number of tests made for each of the heights studied in general was 540, representing a total of 2,544 tests made. If all the heights presented the same distribution with respect to the different activity levels, in each column all the cells should present the same frequency. For example, in the second column, all the frequencies should be 261 (1,308/5). However, the frequencies found in the second column for working heights of 1.2 (360), 1.4 (359), and 1.6 m (344) were higher than 261, while the frequencies found for 1.8 (221) and 2.0 m (24) were lower than this value, showing that an activity level of 2 is more probable to be reached with working heights of 1.2, 1.4, and 1.6 m than with the other working heights studied. With similar reasoning, it was more probable to reach activity levels of 3 with working heights of 1.2, 1.4, and 1.6 m, and activity levels of 4 with working heights of 1.8 and 2.0 m.

Similar results were found with the corrected residuals analysis (indicated in parenthesis), reflecting that the working heights of 1.2, 1.4, and 1.6 m were related fundamentally to the activity levels of 2. Activity levels of 1 were associated with heights of 1.4 m, and to a lesser extent with 1.6 m. Finally, it was found that working heights of 1.8 m and especially heights of 2 m were related to activity levels of 4.

In addition, the marginal frequency analysis by heights indicated that for working heights of 1.8 m, some 3.3% of the participants (1 worker) could not do the tests, and for heights of 2 m, 13.3% of the participants (4 operators). However, no statistically significant relation was detected between the variable “worker height” with respect to the categorical variable “test height”.

Also, Table 4 shows the dependent relation between the variable “weight handled” with respect to the variable “activity level”.

As shown in Table 4, there is a direct relation between the two variables in such a way that as the weight handled increase, the activity level also increased. The analysis of cases together with the corrected residual analysis showed the relation between the factor weight of 1 kg with respect to the activity level of 2, a weight of 5 kg with respect to the activity levels of 3, and a weight of 10 kg with respect to the action levels of 4. In addition, the analysis of marginal frequencies revealed that 3.3% of the participants (1 worker) could work with the weights of 1 kg but not with heavier weights (5 or10 kg).

Finally, Table 5 analyses the relation of dependence between the variable “type of hanger” with respect to the variable “activity level”.

The frequency analysis for the hanger type differentiated two subsets: The first composed of type 1, 2, and 3 hangers; and the second comprising hangers 4, 5, and 6. For the first subset, the frequencies were distributed fundamentally by activity levels of 2 (25.2%), 3 (22.6%) and 4 (52.0%), the behaviour of each hanger being very similar. The analysis of the corrected residuals showed the association of these hangers with the activity levels of 3 and 4, with hanger 2 being the most likely to reach activity levels of 4, followed by hanger 1, and finally hanger 3. On the other hand, for the subset made up of hangers 4, 5, and 6, it was found that the frequencies of each hanger were practically similar, but that the values of the previous group differed enormously, as the greatest number of tests (77.6%) reached an activity level of 2. In addition, the corrected residuals analysis showed that hanger type 6 was also related to activity levels of 1, this being the only hanger for which this association was found.

3.4 Correspondence analysis

Multiple-correspondence analysis has two significant dimensions, the first explaining 46.21% of the variance with a Cronbach α coefficient of 0.932 and an autovalue of 8.318, and the second explaining 37.42% with a Cronbach α coefficient of 0.902 and an autovalue of 6.735. For the model as a whole, the total variance explained was 41.81%, the mean Cronbach α coefficient was 0.918, and the mean autovalue 7.526. The model can therefore be regarded as reliable.

Table 6 shows the discrimination measures of each variable with respect to each of the two dimensions of the model, and the mean. As can be appreciated, the leading variable in the ranking of the explanatory variables of the variance of the homogenizing model was the variable “activity level” (0.702), as it presents the highest discrimination, followed by the variables “type of hanger” (0.435), “working height” (0.372) and “weight handled” (0.082), which proved to be less explanatory.

In addition to identifying the variables that most discriminate the objects, the correspondence model plotted the results on a correspondence map, representing the association between the categories of the multiple variables that made up the test. Thus, the correspondence map of Fig. 4 shows the relation between the categories of the variable “working height” with respect to the categories differentiated in the variable “activity level”. As results, coinciding with those found in the contingency-table analysis, three subsets of heights were differentiated: The first represented by test heights of 1.2 m (1), 1.4 m (2), and 1.6 m (3); the second by heights of 1.8 m (4); and the third by heights of 2.0 m (5). The first subset was associated with activity levels 2, heights of 1.4 m (2) showing a certain relation with action levels of 1. The second subset (1.8 m) was associated with action levels of 3 and 4, and the third subset (2.0 m) with action levels of 4.

Also, Fig. 5 shows the correspondence map of the variable “weight handled” with respect to the variable “activity level”. The results show that the three weight categories tested were well differentiated; weights of 1 kg (1) were associated with activity levels 1 and 2, especially with the latter; weights of 5 kg (2) with activity levels of 3; and weights of 10 kg (3) strongly with activity levels of 4.

On the other hand, Fig. 6 shows the correspondence map of the variable “hanger type” with respect to the variable “activity level”. The analysis differentiated three subsets of hangers: The first included hangers 1, 2, and 3; the second hangers 4 and 5, and the third (not differentiated by the contingency analysis) hanger 6. The first subset was associated with activity levels 3 and 4, the second with activity level 2, and the third with activity level 2 with a weak association with level 1.

Finally, the multiple correspondence analysis draws a relation, in a two-dimensional space, between the categories of the different nominal variables studied (Fig. 7), showing that the levels of action 1 and 2 were associated with weights handled of 1 kg (1), working heights of 1.2 (1), 1.4 (2) and 1.6 m (3), and hanger types 4, 5 and 6. On the other hand, the activity levels of 3 were associated with weights of 5 kg (2), working heights of 1.8 m (4), and with hanger types 1, 2, and 3, and activity levels of 4, were associated with weights handled of 10 kg (3), working heights of 1.8 (4) and 2.0 m (5), and with hangers 1, 2, and 3.

4 Discussion

The present study identifies the most favourable conditions to undertake the most problematic task of training greenhouse crops (lowering the plants on their guides), making recommendations to prevent potential MSDs in upper limbs due to this work.

The main variables which characterized the task of lowering plants (working height, weights handled, hanger types) was simulated in the laboratory, avoiding the possible posture alterations associated with the extreme thermal conditions that, as mentioned by [27] and [38], occasionally occur inside greenhouses.

As in the study by [48] concerning female sewing-machine operators in Turkey, it was not possible to establish any statistically significant dependent relationship between the variables characterizing the workers (gender, age, height and BMI) and the probability of suffering MSDs, perhaps because of the small sample size analysed. In other studies, intense physical activity, high BMI, and advanced age were found to increase the prevalence of MSDs [19, 22].

Of the 2,544 tests analysed, according to the RULA method the working posture could be considered acceptable only in 51.7% of the cases. The rest (49.3%) indicated that 11.6% of the tests required the task to be redesigned soon (activity level 3) and in the others (36.8%), urgent changes were needed in the job (activity level 4). The weighted mean scoring of the group constituted by the arm, forearm, and wrist (scoring C) was 5.18±1.83, indicating that the position adopted by these limbs was quite strained and in some cases extremely so; in fact, some workers did not complete some of the tests. However, the weighted mean scoring of the group that included the neck, trunk, and legs (scoring D) reached a value of 3.98±1.99, revealing that the position was situated at a middle action level, the scoring not being as extreme as that for the upper members. It was observed that the high levels of risk were occasioned fundamentally by a forced position of the arm, with bending more than 90° (43.6% of the cases), of the forearm, normally by bending less than 60° (36.1% of the cases), and of the neck, caused by extensions (22.2% of the cases). In general, these inappropriate postures in the arm, forearm, and neck, were associated with working heights of more than 1.6 m, and thus the task of lowering plants at working heights of between 1.2 and 1.6 m avoids awkward postures in these limbs and thus for these heights the activity levels recorded are lower. Furthermore, it was noted that on occasions the activity levels were aggravated by inappropriate positions of the wrist, whether for deviations with respect to the neutral position (11.8% of the cases) or for turning (15.6% of the cases); of the trunk, for slight bending (13.2% of the cases); and of the neck, as well as slight bending and straightening (23.5% of the cases). In any event, these inappropriate postures were considered avoidable by worker training programmes.

In addition to the inappropriate posture, the application of the method has revealed that this is a highly repetitive activity, since the task repeats more than 4 actions per min to complete the lowering of a plant, on the average 8 actions. In addition, sometimes, the ergonomic conditions were aggravated by the need to bear an excessive work load. Therefore, it can be stated that, according to the RULA method, the high levels of action recorded were caused by: An improper posture, the repetitive nature of the task, and by excessive load to handle. It bears mentioning that the posture adopted and the repetition of the work cycles has been detected as especially problematic by [39] in the task of sorting tomato fruits, as well as by [40] in greenhouse grafting. Nevertheless, in both studies, the authors highlighted the frequency factor of movement as the most problematic one, but this factor is not considered when applying the RULA method and thus no conclusions can be drawn in this sense. Nevertheless, it should be commented that on average, the group of workers spent roughly one min to complete an entire cycle of lowering plants (to become positioned next to a hypothetical plant and execute the repeated cycles of lowering the plant). Therefore, given the duration of the normal workday in Spain (8 h), and the period in which the workers execute their tasks inside the greenhouses, which according to [50] can be considered 80% of the work day, it can be stated that each worker over the workday may lower more than 380 plants. If it is considered that to lower a plant 1.4 m in height, the cycle must be repeated 8 times, each worker over the workday repeats the lowering cycle some 3,040 times. [4] showed the repetitive nature of the rubber-tapping process by the repetition of the work cycles hundreds of times per day, and thus the repetitive nature of training plants in this way explains the danger of suffering MSDs, despite that these have never been evaluated, since, as stated by other authors [6 , 25], one of the factors strongly associated with the probability of suffering MSDs is the repetitive nature of a task.

With the aim of avoiding these ailments, some authors recommend the use of automatic equipment. Thus, [39] recommend the use of automatic machines to sort tomato fruits, and [41] recommend machines to can and package fruit. Despite that this could be a feasible alternative, it should be commented that for the plant-training tasks, there is currently no mechanization of the tasks, nor any procedure of alternative work, making research necessary along these lines. Nevertheless, with the aim of avoiding potential MSDs, it is considered indispensable to guarantee the correct organization of the workday, including recovery times starting from the earliest working hours, given that the work is highly repetitive. This measure to improve the general working conditions in greenhouses of south-eastern Spain has been proposed [28 , 33–35], while [40] have demonstrated that the proper organization of the workday improves the ergonomic conditions of grafting tasks in horticulture.

The statistical analysis performed explains that the activity level is directly related to the variables “working height” and “weight handled”. Thus, it was specified that the minimum activity levels were registered for working heights of 1.4 m, although heights of between 1.2 and 1.6 m could prove acceptable. On the other hand, working heights exceeding 1.6 m imply inacceptable activity levels, these requiring urgent changes in the tasks. In addition, in relation to the variable “weight handled”, it was observed that only weights of less than 2 kg prove ergonomically acceptable, this corresponding to the 1 kg weights tested. The activity level associated with crop weights of between 2 and 10 kg (weight tested of 5 kg), requires the redesigning of the job, making it necessary to make urgent changes in the tasks of manipulating weights greater than 10 kg. The most common horticultural crops in greenhouses in south-eastern Spain that require trellising, are the tomato, cucumber, eggplant, courgette, green bean, and pepper. The regulation on greenhouses for commercial production [2], establishes that the relation of the most restrictive weight is reached in greenhouse tomato and pepper crops (15 kg·m^- 2). It was clarified that the entire crop weight should not be handled during the task of lowering the plants, since the crops are supported by the ground and only a percentage of the weight should be handled. This percentage would vary according to the crop and its phenological state, but in any case should be lower than 10 kg. Therefore, it could be stated that the ergonomic conditions during the training of crops is directly related with the type of crop handled. This aspect could complement the conclusions drawn by [33], who established that the crop directly affects the ergonomic-psychosocial conditions of greenhouse workers. Furthermore, the study shows the importance of the type of hanger needed to perform the work correctly. The present study shows that, of the group of hangers tested, ergonomically acceptable handling was possible only with the new-generation ones having the lock device (in the test hangers 4, 5, and 6), as this type did not require the support of the crop weight. In the case of using other types of hangers, it is recommended not to support weights greater than 2 kg.

Despite that the RULA method differentiated three categories of hangers, it was barely possible to distinguish between the different types of hanger studied in each category. Thus, it would be expected that the evaluations made for hanger type 1 was more limiting than for type 2, as all the workers coincided that its handling was more difficult for being larger. On the other hand, the use of hanger type 4 was far simpler than type 5, which was activated by a reel, and moreover hanger 4 had automatic retraction. Therefore, the results confirm the findings of [42], who indicated that the RULA evaluation method is quite rigid. In this sense, it would useful to make similar analyses of other biomechanical methods, to evaluate the variability between evaluations made by different methods, as in the works of [46, 47], who studied the exposures of sawmill workers. It should not be neglected that no biomechanical method available has been developed to evaluate agricultural workers, and none to assess greenhouse environments. Therefore, as indicated by [33], its application requires an adaptation to the particular conditions of greenhouse work. Nevertheless, previously, it would be necessary to undertake research such as the present study to examine the possible application of existing biomechanical methods. This demand was raised by [28], who called for the development of research on the physical load in the greenhouse sector, in order to relate psychosocial risks detected in the sector with possible MSDs.

In conclusion, the present work indicates that the work of lowering greenhouse plants should be undertaken at a working height of approximately 1.4 m, and preferably using new-generation hangers with a lock device. In the case of using another type of hanger, it is not recommended to handle suspended loads of more than 2 kg. In addition, this reflects the importance of guaranteeing the appropriate organization of the workday, to alleviate the consequences of the repetitive nature of the work as well as to provide training sessions for workers to learn to avoid inappropriate postures, especially in the limbs, trunk, wrist, and neck.

Conflict of interest

The author has no conflict of interest to report.

Footnotes

Acknowledgments

I would like to thank Lecturer Dr. Amelia Victoria Garcia Luengo from the University of Almeria for her assistance with statistical analysis.

References

Bianchi . Tipologie construttive ed impiantistiche nelle colture protette. Colture Protette 1989;7:23–26.

European Committee for standardization CEN. EN-1-1Greenhouses: Design and construction. Part 1: Commercial production greenhouses, Brussels, Belgium; 2002;105.

Castilla

, Hernández

. Protected cultivation in the Mediterranean area. Plasticulture 1995;107(3):13–20.

Meksawi

, Tangtrakulwanich

, Chongsuvivatwong

. Musculoskeletal problems and ergonomic risk assessment in rubber tappers: A community-based study in southern Thailand. Int J Ind Ergonom 2012;42(1):129–135. doi: 10.1016/j.ergon.2011.08.006

Colombini

, Occhipinti

. Preventing upper limb work-related musculoskeletal disorders (UL-WMSDS): New approaches in job (re)design and current trends in standardization. Appl Ergon 2006;37(4):441–450. doi: 10.1016/j.apergo.2006.04.008.

Van Rijn

, Huisstede

BMA

, Koes

, Burdorf

. Associations between work-related factors and specific disorders of the shoulder-a systematic review of the literature. Scand J Work Environ Health 2010;36(3):189–201. doi: 10.5271/sjweh.2895.

Jakob

, Liebers

, Behrendt

. The effects of working height and manipulated weights on subjective strain, body posture and muscular activity of milking parlor operatives-laboratory study. Appl Ergon 2011;43(4):753–761. doi: 10.1016/j.apergo.2011.11.009. Epub 2011 Dec 5.

Herin

, Vezina

, Thaon

, Soulat

, Paris

. Predictors of chronic shoulder pain after 5 years in a working population. Pain 2012;153(11):2253–2259. doi: 10.1016/j.pain.2012.07.024.

Kwon

, Lee

, Jeong

. Work-related hazards among farmers. J Korean Med Assoc 2012;55(11):1046–1053. doi: 10.5124/jkma.2012.55.11.1046.

10.

Nag

, Vyas

, Shah

, Nag

. Risk factors and musculoskeletal disorders among women workers performing fish processing. Am J Ind Med 2012;55(9):833–843. doi: 10.1002/ajim.22075.

11.

, Cheng

ASK

, Li-Tsang

CWP

. Prevalence and risk factors of work-related musculoskeletal disorders in the catering industry: A systematic review. Work: A Journal of Prevention, Assessment and Rehabilitation 2013;44(2):107–116. doi: 10.3233/WOR-2012-1375.

12.

Andersen

, Haahr

, Frost

. Risk factors for more severe regional musculoskeletal symptoms: A two-year prospective study of a general working population. Arthritis Rheum 2007;56(4):1355–1364. doi: 10.1002/art.22513.

13.

Caicoya

, Delclos

. Work demands and musculoskeletal disorders from the Spanish National Survey. Occup Med-Oxford 2010;60(6):447–450. doi: 10.1093/occmed/kqp191.

14.

Barrero

, Pulido

, Berrio

, Monroy

, Quintana

, Ceballos

, Hoehne-Hueckstaedt

, Ellegast

. Physical workloads of the upper-extremity among workers of the Colombian flower industry. Am J Ind Med 2012;55(10):926–939. doi: 10.1002/ajim.22102.

15.

Abdoli-EM , Damecour

, Petersen

, Potvin

. Investigation of pacing as a control measure for an industrial lifting task above waist height. Work: A Journal of Prevention, Assessment and Rehabilitation 2014;47:15–22. doi: 10.3233/WOR-131682.

16.

Dumas

, Preston

, Beaucage-Gauvreau

, Lawani

. Posture analysis of lifting a load for head carriage and comparison between pregnant and non-pregnant women. Work: A Journal of Prevention, Assessment and Rehabilitation 2014;47:63–72. doi: 10.3233/WOR-131686.

17.

MacKenzie

, Grimshaw

, Jones

CDS

, Thoirs

, Petkov

. Muscle activity during lifting: Examining the effect of core conditioning of multifidus and transversus abdominis. Work: A Journal of Prevention, Assessment and Rehabilitation 2014;47:452–462. doi: 10.3233/WOR-131706.

18.

Yeung

, Genaidy

, Deddens

, Alhemood

, Leung

. Prevalence of musculoskeletal symptoms in single and multiple body regions and effects of perceived risk of injury among manual handling workers. Spine 2002;27(19):2166–2172. doi: 10.1097/01.BRS.0000025683.43004.7A.

19.

Andersen

, Gaardboe

. Prevalence of persistent neck and upper limb pain in a historical cohort of sewing machine operators. Am J Ind Med 1993;24(6):677–687. doi: 10.1002/ajim.4700240604.

20.

Balasubramanian

, Adalarasu

, Regulapati

. Comparing dynamic and stationary standing postures in an assembly task. Int J Ind Ergonom 2009;39(5):649–654. doi: 10.1016/j.ergon.2008.10.017.

21.

Memarpour

, Badakhsh

, Khosroshahi

, Vossoughi

. Work-related musculoskeletal disorders among Iranian dentists. Work: A Journal of Prevention, Assessment and Rehabilitation 2013;43:465–474. doi: 10.3233/WOR-2012-1468.

22.

Waersted

, Westgaard

. Working hours as a risk factor in the development of musculoskeletal complaints. Ergonomics 1991;34(3):265–276. doi: 10.1080/00140139108967312.

23.

Osborne

, Blake

, McNamara

, Meredith

, Phelan

, Cunningham

. Musculoskeletal disorders among Irish farmers. Occup Med-Oxford 2010;60(8):598–603. doi: 10.1093/occmed/kqq146.

24.

Asensio-Cuesta

, Diego-Mas

, Cremades-Oliver

, Gonzalez-Cruz

. A method to design job rotation schedules to prevent work-related musculoskeletal disorders in repetitive work. Int J Prod Res 2012;50(24):7467–7468. doi: 10.1080/00207543.2011.653452.

25.

Bodin

, Ha

, Chastang

, Descatha

, Leclerc

, Goldberg

, Imbernon

, Roquelaure

. Comparison of risk factors for shoulder pain and rotator cuff syndrome in the working population. Am J Ind Med 2012;55(7):605–615. doi: 10.1002/ajim.22002.

26.

Callejón-Ferre

, Perez-Alonso

, Sanchez-Hermosilla

, Carreño-Ortega

. Ergonomics and psycho-sociological quality indices in greenhouses Almeria (Spain). Span J of Agric Res 2009;7(1):50–58.

27.

Callejón-Ferre

, Manzano-Agugliaro

, Díaz-Pérez

, Carreño-Sánchez

. Improving the climate safety of workers in Almería-type greenhouses in Spain by predicting the periods when they are most likely to suffer thermal stress. Appl Ergon 2011;42(2):391–396. doi: 10.1016/j.apergo.2010.08.014.

28.

Montoya-Garcia

, Callejon-Ferre

, Perez-Alonso

, Sanchez-Hermosilla

. Assessment of psychosocial risks faced by workers in Almeria-type greenhouses, using the Mini Psychosocial Factor method. Appl Ergon 2013;44(2):303–311. doi: 10.1016/j.apergo.2012.08.005.

29.

Stewart

, Rothmore

, Doda

DVD

, Hiller

, Mahmood

, Pisaniello

. Musculoskeletal pain and discomfort and associated worker and organizational factors: A cross-sectional study. Work: A Journal of Prevention, Assessment and Rehabilitation 2014;48:261–271. doi: 10.3233/WOR-131622.

30.

Gustafsson

, Lundqvist

. Prevention of occupational injuries in greenhouses. Ergonomics 1982;25(6):576.

31.

Lundqvist

. Occupational health and safety of workers in agriculture and horticulture. New Solutions 2000;10(4):351–365. doi: 10.2190/CNC5-ECBE-G7L9-PP7.

32.

NTP 175: Evaluación de las condiciones de trabajo: el método L. E. S.T. Instituto Nacional de Seguridad e Higiene en el Trabajo. Ministerio de Trabajo y Asuntos Sociales. España.

33.

Callejon-Ferre

, Perez-Alonso

, Carreno-Ortega

, Velazquez-Marti

. Indices of ergonomic-psycholsociological workplace quality in the greenhouses of Almeria (Spain): Crops of cucumbers, peppers, aubergines and melons. Safety Sci 2011;49(5):746–750. doi: 10.1016/j.ssci.2010.12.009.

34.

Pérez-Alonso

, Carreño-Ortega

, Callejón-Ferre

, Vázquez-Cabrera

. Preventive activity in the greenhouse-construction industry of south-eastern Spain. Safety Sci. 2011;49(2):345–354. doi: 10.1016/j.ssci.2010.09.013.

35.

Pérez-Alonso

, Carreño-Ortega

, Vázquez-Cabrera

, Callejón-Ferre

. Accidents in the greenhouse-construction industry of SE Spain. Appl Ergon 2012;43(1):69–80. doi: 10.1016/j.apergo.2011.03.007.

36.

Vázquez

, Pérez

, Callejón

, Carreño

. Desing of a new multispan greenhouse capital. Inf Constr 2011;63(521):47–56. doi: 10.3989/ic.09.009.

37.

Carreño

, Pérez

, Vázquez

, Callejón

. A safer method of constructing multi-span greenhouses. Inf Constr 2012;64(525):93–102. doi: 10.3989/ic.09.050.

38.

Pérez-Alonso

, Callejón-Ferre

, Carreño-Ortega

, Sánchez-Hermosilla

. Approach to the evaluation of the thermal work environment in the greenhouse-construction industry of SE Spain. Build. Environ 1725;46(8):1725–1734. doi: 10.1016/j.buildenv.2011.02.014.

39.

Cecchini

, Colantoni

, Massantini

, Monarca

. The risk of musculoskeletal disorders for workers due to repetitive movements during tomato harvesting. J Agric Saf Health 2010;16(2):87–98.

40.

Colantoni

, Marucci

, Monarca

, Pagniello

, Cecchini

, Bedini

. The risk of musculoskeletal disorders due to repetitive movements of upper limbs for workers employed to vegetable grafting. J Food Agric Environ 2012;10(3-4):14–18.

41.

Rai

, Gandhi

, Kumar

, Sharma

, Garg

. Ergonomic intervention in Aonla pricking operation during preserve preparation in food processing industries. Work: A Journal of Prevention, Assessment and Rehabilitation 2012;41(1):401–405. doi: 10.3233/WOR-2012-0190-401.

42.

Chiasson

, Imbeau

, Aubry

, Delisle

. Comparing the results of eight methods used to evaluate risk factors associated with musculoskeletal disorders. Int J Ind Ergonom 2012;42(5):478–488. doi: 10.1016/j.ergon.2012.07.003.

43.

Deepak

, Ajeesh

. Correlation of ergonomic risk factors with RULA in IT professionals from India. Work: A Journal of Prevention, Assessment and Rehabilitation 2012;41:512–515. doi: 10.3233/WOR-2012-0205-512.

44.

McAtamney

, Nigel Corlett

. RULA: A survey method for the investigation of work-related upper limb disorders. Applied Ergonomics 1993;24(2):91–99. doi: 10.1016/0003-6870(93)90080-s.

45.

Sala

, Torri

, Tomasi

, Apostoli

. Risk assessment for upper extremity work related musculoskeletal disorders in different manufactures by applying six methods of ergonomic analysis. Giornale Italiano Di Medicina Del Lavoro Dd Ergonomia 2010;32(2):162–173.

46.

Jones

, Kumar

. Comparison of ergonomic risk assessment output in a repetitive sawmill occupation: Trim-saw operator. Work: A Journal of Prevention, Assessment and Rehabilitation 2008;31(4):367–376.

47.

Jones

, Kumar

. Comparison of Ergonomic Risk Assessment Output in Four Sawmill Jobs. Int J Occup Safety Ergonomics 2010;16(1):105–111.

48.

Öztürk

, Esin

. Investigation of musculoskeletal symptoms and ergonomic risk factors among female sewing machine operators in Turkey. Int J Ind Ergonom 2011;41(6):585–591. doi: 10.1016/j.ergon.2011.07.001.

49.

Chiu

, Chen

, Wu

, Lin

. Three-dimensional computer-aided human factors engineering analysis of a grafting robot. J Agric Saf Health 2012;18(3):181–194.

50.

García

, Padilla

. Estudio de la prevención de riesgos laborales en invernadero. Colegio Oficial de Ingenieros Técnicos Agrícolas de Almería: Spain: 2005.