This article was downloaded by: [University of Connecticut] On: 16 October 2013, At: 14:56 Publisher: Taylor &

Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Occupational and Environmental Hygiene

Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uoeh20

Investigation of a Pork Shoulder Deboning Operation

Patrick G. Dempsey & Raymond W. McGorry
a a a

Liberty Mutual Research Institute for Safety , Hopkinton , Massachusetts Published online: 17 Aug 2010.

To cite this article: Patrick G. Dempsey & Raymond W. McGorry (2004) Investigation of a Pork Shoulder Deboning Operation, Journal of Occupational and Environmental Hygiene, 1:3, 167-172 To link to this article: http://dx.doi.org/10.1080/15459620490424465

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the Content) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Journal of Occupational and Environmental Hygiene, 1: 167172 ISSN: 1545-9624 print / 1545-9632 online Copyright c 2004 JOEH, LLC DOI: 10.1080/15459620490424465

Investigation of a Pork Shoulder Deboning Operation

Patrick G. Dempsey and Raymond W. McGorry
Liberty Mutual Research Institute for Safety, Hopkinton, Massachusetts

Downloaded by [University of Connecticut] at 14:56 16 October 2013

An ergonomic investigation of force exertions during pork shoulder deboning was conducted. Nine male workers volunteered to participate in the study. Three cycles of the shoulder deboning task were measured for each worker, using an instrumented knife that provided direct measurements of grip forces and cutting moments applied to the knife. The average grip force values during the task were between 11 and 35% of the maximum voluntary grip force. A key nding was that there is high between-subject variation in exposure to integrated grip forces and cutting moments, which can have implications for exposure to risk for upper extremity musculoskeletal disorders. There was also high between-subject variation in cutting time. Measures of knife sharpness indicated that there is also variation in knife maintenance among workers during a 5-hour period. The implications of the results for training and further study are discussed. Keywords hand tools, knife sharpness, meat processing, musculoskeletal disorders

Address correspondence to: Patrick G. Dempsey, Liberty Mutual Research Institute for Safety, 71 Frankland Road, Hopkinton, MA 01748; e-mail: patrick.dempsey@libertymutual.com

asks involved in processing animal carcasses have long been associated with elevated injury and illness risk. Aside from acute injuries, there is risk of musculoskeletal disorders (MSDs) often attributed, at least partially, to the forces, postures, and repetition of the meat-processing tasks.(13) Due to the technical infeasibility of automating many of the hand-intensive tasks, ergonomic efforts often focus on engineering controls, such as knife and workstation design, as well as administrative controls, particularly job rotation. In spite of the research conducted previously, there are still voids in the ability to quantify the exposure to MSD risk factors such as force, repetition, and posture. Unlike chemical exposures, there are no widely accepted limits for exposure. Some of the lack of specicity may be due to the difculty encountered when attempting to quantify exposure. There have been considerable advances in the measurement of repetition and wrist posture, but advances in quantifying exposure to muscular force have not been as signicant. This is evidenced by the use of estimates of exposure to force exertion in several

epidemiologic studies.(2,4,5) Previous approaches to measuring forces exerted while using hand tools include using electromyography to estimate forces.(6,7) The study reported here is an observational study; the study was conducted at the workplace and there were no independent variables manipulated. The primary goal was to perform an initial investigation into the extent of within- and betweensubject variation in grip forces and cutting moments exerted on a knife handle, through direct measurement techniques. The rationale for this is that it is largely unknown how much exposure variation exists between workers and how much force is exerted while processing meat. Large between- or withinworker variation would have implications for the ergonomics practitioner, as well as for epidemiologic studies. In the latter case, the sampling strategy would need to consider this variation to properly assess exposure, as large between-subject variation would argue for exposure assessment at the individual level. There may also be individual factors such as experience that affect force exposure, and this could be used to develop training programs or administrative controls. The secondary goal of the study was to perform an observational study of knife sharpness. Although sharpness is regarded as a factor that can contribute to exposure,(810) it is unknown how sharp different workers are able to maintain their knives. For this study, a hog shoulder deboning task was studied. The task involves making multiple cuts on a shoulder to remove the blade and shank bones. METHODS Subjects Nine of nine male workers performing the deboning task during the data collection period volunteered to participate in the study. The anthropometric characteristics of the sample are given in Table I. The volunteers gave informed consent as approved by the Liberty Mutual Research Centers Institutional Review Committee. Subjects were not paid for participation, but were given a t-shirt for volunteering. Apparatus The basic apparatus used included a tape measure and a video camera. The video camera used the MiniDV format, and the video was transferred to a personal computer for analysis.
March 2004 167

Journal of Occupational and Environmental Hygiene

TABLE I. Anthropometric and Demographic Characteristics of Nine Male Subjects Parameter Age (yrs) Height (cm) Weight (kg) Elbow height (cm) Grip force MVC (N) Experience on job (yrs) Experience cutting meat (yrs) Mean Standard Deviation 32.8 169.1 84.7 113.0 205.6 0.9 2.8 8.2 9.3 9.7 2.8 59.1 0.9 2.8

Downloaded by [University of Connecticut] at 14:56 16 October 2013

The cameras clock function was activated so that a time record for cycle time measurements was captured on the video. An instrumented knife was used to measure grip forces and cutting moments. The design and implementation is described briey, and more details are provided by McGorry.(11) The handle core was fabricated from 15.6-mm-diameter round stock (14-4 stainless steel), that was milled to create three beams, supported at both ends, and suspended about a hollow center. Thin foil strain gauges were xed to each end of the three beams. The tool end of the grip core has a threaded hole to which the knife blade was attached. A round cross-section was milled into the core, between the tool end tting and the three suspended beams of the handle. Two pairs of strain gauges were applied to this end that accepted the knife blade (one pair for moments in the horizontal plane, one pair for moments in

the vertical plane). The wires from the applied moment gauges and from the grip force gauges on the suspended beams were routed through the hollow center, and exit through a hole at the butt of the grip core. The handle of the knife used in the deboning operation (Victorinox 40515, Ibach, Switzerland) was used to make a mold. The mold was then used to make a casting of the handle. After fabrication, the handle was split into three sections. Each handle section was attached to the grip core beams with screws. For each worker, a fresh blade was used that had been sharpened on a mechanical sharpener (Tru Hone, Ocala, Fla., adjustable bevel and speed knife sharpener). The instrumented knife is shown in the bottom of Figure 1, with the regular knife used in the operation shown for comparison purposes. Knife sharpness was measured with the device described by McGorry et al.(8) The operation of the device will be briey summarized. The basic principle of the measurement approach is that a sharp blade requires less force to cut than does a dull blade. The handle of the knife blade is placed in a clamping xture. A linear actuator is used to drive the xture containing the knife at approximately 40 mm/sec downward and at a 45 angle through a piece of plastic-coated berglass mesh with a 1.5 2.5 mm weave. This produces a dynamic, yet controlled and repeatable motion with a reasonable approximation of a cutting motion, that allowed measurement of the cutting force along the entire blade edge. Since the instrumentation is in the sharpness tester, any knife can be tested. A mesh material was selected for the test material because the process of cutting through a series of independent strands

FIGURE 1.

Knife used in deboning operation (top) and experimental instrumented knife

168

Journal of Occupational and Environmental Hygiene

March 2004

minimized problems associated with tear propagation, which could occur with a solid material. The mesh strip, approximately 4 cm wide by 40 cm long, is oriented vertically. The lower end of the mesh is clamped to the test frame base. The upper end is attached to a load cell with an 18 kg measurement range. As the blade cuts the strands of the mesh, the reactive force is measured by the load cell, and the amplied output is sampled at 100 Hz by an A/D converter of a laptop computer and stored in memory. The force required to cut the mesh provides a relative measure for comparison of blade sharpness. The load cell was calibrated prior to testing. To assess repeatability of the device, 3 repeated samples from 10 knives were taken. The correlations between the sharpness readings for the 3 trials ranged between 0.98 and 0.99. Procedures The workers were read the informed consent and signed the form if they agreed to participate in the study. All the workers volunteered, eliminating potential selection bias. The worker was then given a brief explanation of the data collection procedures. The maximum voluntary contraction (MVC) for grip force of the hand used for cutting was collected using the instrumented knife handle. The basic protocol used was the Caldwell regimen(12) with a minor modication. A 15% within-trial variation was used instead of the 10% prescribed because subjects sometimes have difculty achieving this criterion for hand grips.(13) Two consecutive trials within 15% of each other were averaged and used as the MVC value. The data collection began with videotaping ve cycles using the regular Victorinox knife. This was performed so that the typical cycle time could be compared to the cycle time using the instrumented knife. A complete hog shoulder deboning was considered one cycle. Next, the battery pack for the instrumented knife was attached to the workers abdomen guard belt, and the cable was secured to the upper arm of the worker to minimize interference. The length of cable was adjusted to a length that the worker preferred. One cycle was completed with the instrumented knife so that the worker could become accustomed to the knife, particularly the cable that extended from the end of the handle. Following the practice cycle, 34 cycles were completed during which data were collected. Due to technical difculties with several trials, three cycles for each worker were used for analysis. The data were collected at 100 Hz and stored in text les for later analysis. The data les were examined and the beginning and end of each le were identied using a cutting moment criterion. The beginning of a trial was dened as the cutting moment exceeding 0.5 N m and the end of a trial was dened as the last sample exceeding 0.5 Nm. The amount of time between these two data was dened as the cycle time. This value was chosen by examining several les. Similarly, cutting time was dened as the amount of time, within each cycle, during which the cutting moment exceeded 0.5 Nm. The number of samples

that the criterion was exceeded were counted and multiplied by the inverse of the sampling frequency to arrive at the cutting time value. This is what has been called duty cycle by some. This approach does not consider low-force exertions, such as holding the knife between cycles. Several force variables were calculated from the data. Peak and average grip force and cutting moment were calculated using the samples within cutting time. Grip force was calculated by taking the mean of the output of the six strain gauges mounted on the handle beams. Cutting moment was calculated about the midpoint of the long axis of the knife handle.(11) In addition to peak and average forces and moments, the grip force and cutting moments were integrated across cycles to yield a measure of the amount of force exerted per cycle (i.e., force exertion per shoulder processed). A similar measure was used in the authors investigation of ice cream scooping,(14) and represents the force exposure per cycle. The grip force data were also normalized to each workers grip MVC to assess the percentage of peak and average MVC the workers exert. As with the ice cream scooping task assessment,(14) the goal was to examine the jobs relative to MVC guidelines found in the literature. The videotapes of the trials were analyzed so that the number of individual cuts could be counted. A cut was operationally dened as starting when the knife blade entered the meat and ended when the knife blade was removed from the meat. For some cuts, sawing motions were noticed where the worker would make a cut using sawing motions. The number of sawing motions per cycle were counted, with forward and backward motions each counted as one motion. The same analyst analyzed all video for consistency. For the sharpness study, the actual knives normally used during the deboning operation were used. The actual sharpness tester was instrumented. Knives sharpened by the plant sharpener were measured at the beginning of the day. Since the sharpening process is a metal-removal process, the length of the knife blades decrease over time. The researchers asked that knives be selected that were at approximately the midpoint of their life, since the properties may change as the blade changes size. It was assumed that selecting the knives in this manner would yield a representative average sample. The knives were tested with the sharpness tester two hours after the shift started, and again after ve hours. At each test period, two trials were taken for each knife and the results were average. The data from the sharpness test were reduced by extracting the peak value (at the tip of the knife typically), as well as averaging the data from within the range from the tip of the blade to 3 cm down the blade. All samples within this range were averaged to produce a single datum for the range. The data from two replications for each knife were then averaged. Because the authors used the sharpness test in a relative measure, the sharpness readings at the beginning of the shift were subtracted from the readings taken at the 2-hour point and 5-hour point. A positive value would indicate higher forces were required, and that the knife was not as sharp as at the initial reading and vice versa.
March 2004 169

Downloaded by [University of Connecticut] at 14:56 16 October 2013

Journal of Occupational and Environmental Hygiene

TABLE II. Variable

Summary Statistics for Job Analysis Parameters Mean 49.9 21.6 16.2 130.9 4.0 39.9 84.9 864.2 24.1 20.1 68 21 Standard Dev. 11.4 7.5 3.1 26.5 0.6 4.4 25.8 339.0 4.2 13.6 22 8 Minimum 30.8 13.9 10.0 94.2 3.2 28.5 49.0 458.0 19.0 4.0 44 11 Maximum 80.0 42.9 22.8 181.4 5.3 46.3 138.9 1859.6 34.0 44.0 120 35

Cycle time (sec) Cutting time (sec) Max. cutting moment (Nm) Max. grip force (N) Average cutting moment (Nm) Average grip force (N) Integrated cutting moment (Nm*sec) Integrated grip force (N*sec) Number of cuts Number of sawing motions Percent grip MVCpeak Percent grip MVCaverage

Downloaded by [University of Connecticut] at 14:56 16 October 2013

RESULTS

able II provides summary statistics for the job analysis parameters. Aside from these descriptive statistics, a variance components analysis (using PROC VARCOMP of the SAS system)(15) was performed to estimate the contribution of replication (within-subject) and worker (between-subject) to variation in cutting time, integrated grip force, and integrated cutting force. Table III provides the minimum variance quadratic unbiased estimators (often denoted MIVQUE(0)) of the variance components. The percentage of variance due to replication (within-worker) and worker (between-worker) are indicated in the table. Overall, the results suggest strong between-subject variation, and little variation within-subject. To investigate the potential relationships between the subject demographics and job analysis parameters, linear Pearson correlation coefcients were estimated. One of the notable trends in the data was the correlations between years on the job and cutting time (r = 0.66, p < 0.0002), integrated grip force (r = 0.65, p < 0.0002), and integrated cutting moment (r = 0.72, p < 0.0001). These relationships are shown in Figure 2ac, respectively. Cutting time and the latter two measures are interrelated, and cutting time was more strongly correlated with the integrated measures than the force or moment components. For example, the correlation between

TABLE III. Analysis Variance Component

Results

of

Variance

Components

Cutting Time A

Integrated Cutting Moment A

Integrated Grip Force A

FIGURE 2. The relationships between years on the job and cutting time, integrated grip force, and integrated cutting moment

Worker 58.7 (95.4%) 689.9 (95.6%) 112807.1 (91.2%) Replication 2.8 (4.6%) 32.1 (4.4%) 10900.8 (8.8%)
A MIVQUE(0)

estimates (percentage of variance).

170

Journal of Occupational and Environmental Hygiene

March 2004

deemed unusable. In the other case, the worker exchanged his knife for a freshly sharpened knife before the 5-hour reading because his knife had become too dull. DISCUSSION

O
FIGURE 3. Relationship between integrated grip force and cutting time

Downloaded by [University of Connecticut] at 14:56 16 October 2013

cutting time and integrated grip force (0.96, p < 0.0001) (see Figure 3) was much higher than the correlation between cutting time and average grip force (0.11, p > 0.55). Correlations involving moments were generally of low magnitude or not signicant. Two parameters positively correlated with cutting time were the number of cuts (r = 0.71, p < 0.0002) and number of sawing motions (r = 0.82, p < 0.0001). There were negative correlations between years on the job and the number of cuts (r = 0.49, p < 0.02) and number of sawing motions (r = 0.61, p < 0.002). The differences in sharpness readings between the 2- and 5-hour measurements and the initial readings are given in Table IV for each of the workers. In two cases, the knives were dulled to the point that the readings were very erratic and TABLE IV. Results of Sharpness Tests Knife Location Time 2 hours 2 hours 2 hours 2 hours 2 hours 2 hours 2 hours 2 hours 2 hours 5 hours 5 hours 5 hours 5 hours 5 hours 5 hours 5 hours 5 hours 5 hours Subject Number 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 Knife Tip 61.6 11.0 21.9 3.9 0.3 21.7 30.4 17.2 71.7 61.5 64.4 29.9 17.9 1.1 44.3 60.6 58.7 03 cm 0.4 0.1 4.6 0.8 1.7 6.9 3.3 4.8 0.9 3.0 5.8 0.7 1.1 1.8 3.5

Note: Values expressed in Newtons.

ne of the most important ndings of this study was the high between-subject exposure to force. The variance components analysis results presented in Table II indicates that within-subject variation was relatively small. It should be noted that the three replications were taken consecutively, and higher within-subject variation may exist. Further study with sampling at different times within and across days will be required to examine additional potential sources of variation. The results of this study do suggest that exposure assessments for epidemiologic studies of meat processing tasks should be conducted at the individual level. Similarly, ergonomic assessments for the purpose of loss prevention should consider the variation between workers, since measurement of one worker may be misleading. When expressed as a percentage of MVC, the average and peak grip forces also indicate considerable variation in force exertion. The mean grip forces ranged between 11 and 35% of MVC, and clearly some of these values exceed most ergonomic guidelines (for example, Putz-Anderson).(14) The peak grip force values all exceeded 40% of MVC, with some reaching over 100% of MVC. The reason these values can exceed 100% is because the MVC is taken with the knife mechanically isolated and the only way to exert force on the handle is through grip force. During the task, the worker can exert additional forces onto the knife handle when the blade is in contact with the meat or table. Additionally, the MVC is an average of a 3-sec exertion, whereas the peaks observed were in the order of milliseconds. Finally, the MVC procedure is the maximal voluntary force that the worker exerts, and is not truly a maximal measure of strength. Although the process investigated was self-paced, a machine-paced setting would introduce additional signicance to the results. The total exposure to force exertion per day would be equal to the average integrated force per shoulder multiplied by the number of shoulders deboned. If all workers were required to perform an equal number of cycles, substantial variation in total exposure would result. The self-paced setting studied could result in less substantial differences since the workers with higher integrated grip forces and cutting moments also had longer average cycle times. By processing fewer shoulders, exposure would be reduced. New workers in a machine-paced environment could not only be exerting more force per cycle, but their rest or micro-breaks would be limited due to their longer cycle times. There is evidence that newer workers experience a higher rate of upper limb musculoskeletal disorders than more experienced workers,(17) and the results of this study provide a plausible reason why new workers in a machine-paced environment would experience higher risk. The nding that integrated grip force was strongly related to cutting time rather than average force was similar to a previous
March 2004 171

Journal of Occupational and Environmental Hygiene

nding(14) that integrated grip force per unit weight of ice cream scooped was more strongly related to time than average force. In the previous study,(14) it was suggested that because force levels were already relatively high, subjects chose to increase time to accomplish the task. However, in the current study, experience working at the deboning job appeared to be a primary determinant of cutting time, which in turn was strongly related to the amount of integrated grip force and cutting moments exerted during a cycle. This would imply that forces would be fairly similar, but that the time spent exerting the force varied. If additional study conrms this nding, assuring proper training would be an administrative means of minimizing exposure. As mentioned earlier, the number of cuts and number of sawing motions were positively correlated with cutting time, thus it appears that these measures can be used as a gross indication of training level. They can also be used to train workers, as minimizing the number of cuts required and eliminating the need to use sawing motions should lead to reduced force exposure. However, this hypothesis needs to be tested as it remains uncertain if the slope of the learning curve with respect to these two specic issues can be increased, or if they require time on the job regardless of training. Because this was an observational study, the sawing motions cannot be assume to cause longer cycle times and higher integrated forces, although it is viable that these work methods would lead to the results noted. The sharpness data provided some interesting insights. In particular, one worker actually had sharpness readings after 2 and 5 hours of working, which indicated that the knife was as sharp, or slightly sharper, than at the beginning of the shift. This worker was the only one who used a steel to maintain his knife, whereas the other eight used what are colloquially called mousetraps. These devices are intended to maintain knife sharpness while requiring little or no training. Because knife sharpness affects force exertion, further study is warranted to investigate if other steel users are able to maintain their knives as well as the worker mentioned. Clearly, a number of knives tested after being used were dulled considerably, indicating that more frequent use of the mousetrap is warranted, or possibly the use of a different method of sharpening. Another option would be more frequent substitution of knives for freshly-sharpened ones. There was considerable variation among workers in sharpness differences between the morning reading and 2- and 5-hour readings, and these could lead to different levels of exposure. One potential administrative control would be to maintain knife sharpness so that exposure is minimized. Further research will be required to determine what devices best maintain knives, as well as how often knives should be machine sharpened. There are several limitations to the study reported. As was mentioned, data were not collected across days or weeks, so conclusions regarding the nature of within- and betweensubject variation are limited to the conditions studied. Similarly, since nine volunteers were available, the conclusions

need to be interpreted within the context of this somewhat limited subject pool. Wrist deviation was not quantied, thus the exposure to deviated wrist postures would need to be considered in a comprehensive ergonomic analysis. With respect to knife sharpness, an absolute measure of initial sharpness was not obtained, thus differences in initial sharpness may have affected the results. Finally, only one task was studied, and further studies of a wider variety of meat processing tasks is warranted. The task was fairly complex, and simpler jobs may not have similar ndings. REFERENCES
1. Armstrong, T.J., J.A. Foulke, B.S. Joseph, and S.A. Goldstein: Investigation of cumulative trauma disorders in a poultry plant. Am. Ind. Hyg. Assoc. J. 43:103116 (1982). 2. Moore, J.S., and A. Garg: Upper extremity disorders in a pork processing plant: Relationships between job risk factors and morbidity. Am. Ind. Hyg. Assoc. J. 55:703715 (1994). 3. Occupational Safety and Health Administration (OSHA): Ergonomics Program Guidelines for Meatpacking Plants (Publication No. OSHA 3123). Washington, D.C.: OSHA, 1991. 4. Punnett, L.: Ergonomic stressors and upper extremity disorders in vehicle manufacturing: Cross sectional exposure-response trends. Occup. Environ. Med. 55:414420 (1998). 5. Silverstein, B.A., and R.E. Hughes: Upper extremity musculoskeletal disorders at a paper mill. Appl. Ergonom. 27 :189194 (1996). 6. Armstrong, T.J., D.B. Chafn, and J.A. Foulke: A methodology for documenting hand positions and forces during manual work. Am. Ind. Hyg. Assoc. J. 12:131133 (1979). 7. Fogleman, M.T., A. Frievalds, and J.H. Goldberg: An ergonomic evaluation of knives for two poultry cutting tasks. Int. J. Ind. Ergonom. 11:257265 (1993). 8. McGorry, R.W., P.C. Dowd, and P.G. Dempsey: Cutting moments and grip forces in meat cutting operations and the effect of knife sharpness. Accepted for publication in Applied Ergonomics 34:375382 (2003). 9. Szabo, R.L., R.G. Radwin, and C.J. Henderson: The inuence of knife sharpness on poultry processing operator exertions and the effectiveness of re-sharpening. In Proceedings of the Human Factors and Ergonomics Society 42nd Annual Meeting, pp. 921925. Santa Monica, Calif: The Human Factors and Ergonomics Society, 1998. 10. Szabo, R.L., R.G. Radwin, and C.J. Henderson: The inuence of knife dullness on poultry processing operator exertions and the effectiveness of periodic knife steeling. Am. Ind. Hyg. Assoc. J. 62:428433 (2001). 11. McGorry, R.W.: A system for the measurement of grip forces and applied moments during hand tool use. Appl. Ergonom. 32:271279 (2001). 12. Caldwell, L.S., D.B. Chafn, F.N. Dukes-Dobos, et al.: A proposed standard procedure for static muscle strength testing. Am. Ind. Hyg. Assoc. J. 34:201206 (1974). 13. Berg, V.J., D.J. Clay, F.A. Fathallah, and V.L. Higginbotham: The effects of instruction on nger strength measurements: Applicability of the Caldwell regimen. In F. Aghazadeh, editor, Trends in Ergonomics/ Human Factors V, pp. 191198. Amsterdam: Elsevier, 1988. 14. Dempsey, P.G., R.W. McGorry, J. Cotnam, and T.W. Braun: Ergonomics investigation of retail ice cream operations. Appl. Ergonom. 31:121130 (2000). R 15. SAS Institute Inc.: SAS/STAT Users Guide, 4th ed. Cary, N.C.: SAS Institute, 1989. 16. Putz-Anderson, V. (ed.): Cumulative trauma disorders. London: Taylor & Francis, 1988. 17. Hakkanen, M., E. Viikari-Juntura, and R. Martikainen: Job experience, work load, and risk of musculoskeletal disorders. Occup. Environ. Med. 58:129135 (2001).

Downloaded by [University of Connecticut] at 14:56 16 October 2013

172

Journal of Occupational and Environmental Hygiene

March 2004