Are hard hats a risk factor for WRMSD in the cervical-thoracic region?

2.1 Subjects

Twenty college students were recruited without regard to age or gender using an Institutional Review Board approved script. Before testing, all 20 participants were confirmed to be in good health with no prior neck injury or back pain present. Six subjects were later excluded due to issues with their EMG data. Subjects provided informed consent prior to the start of the study. Each subject was given a small, nominal payment for participation.

2.2 Experimental design

A repeated measures design was used to assess the effect of hard hat weight and neck posture on muscle activity in the CT trapezius. Muscle loading was assessed using sEMG while wearing three different head protection weights (none, head protection only, and head protection with accessories) at two different neck postures (neutral and 20 degree flexion). Each subject was assessed under each of the six head protection weight and neck posture conditions.

2.3 Measurement procedures

Participants were weighed to estimate head and neck weight as 7.9% of their total body weight, as originally described by Dempster in 1955 [8]. A Delsys Trigno© Wireless sEMG system (Delsys, USA) was used to measure muscle loading in the upper trapezius muscles of the CT region. One Trigno Mini sensor was applied to the surface of the skin above the upper trapezius muscle on the dominant hand side of the participant. The sensor placement was oriented parallel to the direction of the upper muscle fibers of the trapezius and placed approximately three to five centimeters (depending on the size of the subject) distal and superior from the C7 spinous process. A reference electrode was placed over the clavicle on the same side of the body. The sEMG electrodes were worn throughout the duration of the experiment. Subjects were prepared for the sensor attachment using an alcohol pad to clean the site and a safety razor was used, when necessary, to remove excessive hair that could prevent the sensor electrode from contacting the skin.

A hook providing a point of resistance was fastened to a stationary setup on a table. The hook setup was height adjustable +/–15 cm from the table top. A sling was attached to the end of the hook via a steel cable. The sling was fashioned from a hard hat suspension, which was fully adjustable to fit tightly and provide support below the occipital bone during neck extension. Participants were seated in front of the stationary hook set up in a height adjustable chair. The heights were adjusted for each participant so that they would have a flexed, chin-to-chest posture while wearing the sling that was attached to the hook. The hook setup, sling, and chair system provided resistance and isolation for isometric neck extension.

Subjects were asked to extend their neck as hard as possible against the resistance of the hook system for five seconds, three times with a five second rest period in between extensions. Subjects were instructed to relax in between each extension. The sEMG signal recorded during this process provided the subjects’ maximum voluntary contraction (MVC) reference value for sEMG signal normalization.

Participants were then placed in each of the six head protection weight and neck posture conditions. Twenty-degree neck flexion was determined via a goniometer. Subjects were instructed to sit naturally in an adjustable chair. The center of the goniometer was aligned over the external auditory meatus. The proximal arm of the goniometer was perpendicular to the ground while the distal arm was aligned with the alar rim of the nose. Subjects were instructed to lower their chin until the distal arm reached the appropriate 20 degree flexion position on the goniometer. This positioning occurred once for each 20 degree of flexion position for all subjects. Once a subject was situated in each condition, a five second sEMG measure was taken, until the subject had been measured in each of the six head protection weight and neck posture condition. All subjects received all conditions in a random order to minimize ordering effects, such as learning or fatigue.

The head protection used in this study were four-point ratchet suspension hard hats (3M, USA), which weighed 340 grams. The weighted head protection was equipped with hearing protection in the form of earmuffs, a face shield bracket, a personal cooling unit, and a LED headlamp. The muffs used were Peltor X-Series Cap-Mounted earmuffs (3 M, USA) with a total weighing 226 grams. The face shield was a Safety Shield (MCR, USA) with an aluminum bracket. The total weight of this shield bracket was 158 grams. The headlamp was a Headlamp model XML L2 (Boruit, China) which weighed 430 grams. The personal air-cooling unit was Workflow Hardhat Cooling System (Workflow Industrial Products, USA) which weighed 907 grams. The weight of the entire weighted head protection was 2.06 kilograms. The weighted head protection provided a worst-case scenario in terms of head protection weight for comparison with the much lighter head protection without accessories. Typical head protection accessory usage varies by industry and situation, but the total hard hat weight would not likely exceed that of our weighted head protection.

2.4 Data analysis

Data was processed and stored on a laboratory computer using Delsys EMG Works version 4.3 and Delsys EMG Analysis version 4.2.0.0, copyright Delsys Inc., Natick, MA, USA. Two members of the research team independently reviewed the raw sEMG signal from each subject’s MVC effort to determine whether the three maximal extensions were clearly distinguishable from the signal background noise. Six of the 20 subjects without three discernable and distinct efforts in the sEMG signal were excluded from further analysis in the study. This signal quality check ensured that each subject included in the analysis had a valid reference contraction for sEMG comparison of the different head protection weight and neck posture conditions.

The sEMG signal for each subject under each condition was normalized using the subject’s root-mean-squared MVC. This amplitude analysis provided the relative muscle loading of subjects under each condition in terms of a percent MVC. The mean percent MVC, measured in the five-second-measurement period for each subject, was recorded for each head protection weight and neck posture condition.

Statistical analyses of these data were performed using SAS software version 9.2, copyright, SAS Institute Inc., USA. Differences in mean muscle loading between the six head protection weight and neck posture conditions were analyzed using a repeated measures analysis of variance (ANOVA) model. Both head protection weight and neck posture were coded as within-subject variables, with no between subject variables included in the model. The null hypotheses for this analysis were:

There was no significant effect of head protection weight on muscle loading in the upper trapezius.

The assumption of sphericity was assessed using Mauchly’s Sphericity test. Because the assumption of sphericity was not fulfilled by the data nor following generally accepted transformations, the multivariate alternative Wilk’s lambda test was used instead. A post-hoc comparison of significant variables was accomplished using a Wilcoxon Signed-Rank test. The level of significance for each pair-wise comparison was adjusted to maintain a family-wise Type I error rate of 5% by using a Šidák correction according to the following formula: α _(P–W)  = 1–(1–0.05)^1/m, where ‘α _(P–W) ’ is the pair-wise level of significance and ‘m’ is the number of comparisons [9].

Because the assumption of sphereicity was not met for these data nor was sphereicity adequately improved by transformation, the non-parametric alternative Wilk’s Lambda multivariate test was performed. The effect of head protection weight on muscle loading in the upper trapezius was small but significant (Λ= 0.58, p = 0.038). The effect of neck posture on muscle loading was not statistically significant (Λ= 0.89, p = 0.167). The nonparametric Wilcoxon Signed-Rank test was used to compare the muscle loading associated with individual head protection weights and neck postures. Six conditions were compared, resulting in a pair-wise level of significance of α= 0.009. The only significant post-hoc difference at this level was the head protection only when compared the head weight only without head protection in the flexed posture (p = 0.006). The post-hoc comparison results are provided in Table 3. There was no significant interaction between the head protection weight and neck posture (p = 0.983).

4.1 Study limitations

The findings of this study should be considered in light of several methodological and sample limitations. Subjects were limited to a small, convenience sample of college students ranging from 21 to 33 years of age. Additional demographic factors such as fitness level and experience wearing head protection were not considered. This study was a pilot study designed to estimate the effect size of head protection on muscle loading in the neck relevant to industries where hard hat usage is common. Given that the results were largely not statistically significant, this may in part be due to underpowered statistics and the effect sizes reported here can be used to inform future investigations. However, the absolute differences between hard hat weight and neck postures were quite modest and generally within 1% MVC.

In addition, only one muscle and one type of posture were assessed. Workers who wear hard hats must do so while in a wide variation of body and neck postures for short to extended periods of time. Industries such as construction and mining require workers to use multiple dynamic head positions and extreme neck movements while wearing hard hats. Furthermore, the only muscle examined in this study was the upper trapezius due to equipment and logistic limitations. Evaluation of additional para spinal CT muscles in a variety of neck postures would allow for a further understanding of the muscle loading associated with wearing head protection typically used in industry.

Subjects in the present study were not exposed to any additional ergonomic risk factors including whole body vibration (WBV), which can be commonly encountered in industries that use head protection. However, the presence of WBV may increase the risk of neck pain and development of WMSDs, although the association is less clear than between low back pain and WBV [12]. Further, working in a “twisted” position while being exposed to WBV has been associated with increased muscle load in the neck and shoulders [12]. The present study was limited to two neck postures, not including any sort of rotation or WBV. Postural deviation in flexion was modest at best. Ranges of motion greater than 20 degrees from neutral are likely to result in increased muscle fiber recruiting and overall activity. Thus, it would be inappropriate to discount the potential for head protection to increase WMSD risk in the presence of these additional risk factors.

4.2 Study strengths

A main strength of this study was the focus on hard hats typically worn by industrial workers. Further, employing the use of typical (head protection only) and worst-case (head protection loaded with a number of accessories) weights provides a range of results that should encompass the large variety actual head protection weights seen in industry, especially given the variety of hard hat and accessory configurations that are available to workers.

Another benefit of this study was the inclusion of both male and female participants. Prior studies involving military helmets were limited only to male participants. However, in industry today you will find both male and female workers. This is important because of the difference between male and females as it relates to ergonomic risk factors. Anthropometric and physiological factors may make women more prone to WSMDs in the work place [13]. When exposed to identical industrial work tasks, women as compared to men had higher muscle activation and WMSD risk levels [13].

Finally, the use of a repeated measures design was advantageous because it allowed for each subject to act as their own control. By designing the study to obviate the between-subjects variability inherent in ergonomic exposure measurements, the effect of different head protection and neck posture conditions was isolated. A typical ANOVA, for instance, would have required a much larger sample due to the inclusion of both between- and within-subject variability.