ARTICLE IN PRESS
`
`Applied Ergonomics 36 (2005) 199–206
`
`www.elsevier.com/locate/apergo
`
`The effect of simulated school load carriage configurations on
`shoulder strap tension forces and shoulder interface pressure
`Hamish W. Mackiea, , Joan M. Stevensonb, Susan A. Reidb, Stephen J. Leggc
`
`aSchool of Sport, UNITEC, New Zealand, Private Bag 92025, Auckland, New Zealand
`bErgonomics Research Group, Queen’s University, Kingston, Ont./ Canada
`cDepartment of Human Resource Management, Centre for Ergonomics, Occupational Safety and Health, College of Business,
`Massey University, Palmerston North, New Zealand
`
`Received 11 March 2004; accepted 25 October 2004
`
`Abstract
`
`Recently, several studies have addressed the physical demands of school student’s load carriage, in particular the load weight
`carried, using physical demands indicators such as oxygen consumption, gait, and posture. The objective of this study was to
`determine the effects of different load carriage configurations on shoulder strap tension forces and shoulder interface pressure
`during simulated school student’s load carriage. A load carriage simulator was used to compare shoulder strap forces and shoulder
`pressure for 32 combinations of gait speed, backpack weight, load distribution, shoulder strap length and use of a hip-belt. The
`results showed that the manipulation of backpack weight, hip-belt use and shoulder strap length had a strong effect on shoulder
`strap tension and shoulder pressure. Backpack weight had the greatest influence on shoulder strap tension and shoulder pressure,
`whereas hip-belt use and then shoulder strap adjustment had the next greatest effects, respectively. While it is clear that researchers
`and practitioners are justified in focusing on load magnitude in backpack studies as it has the greatest effect on shoulder forces, hip-
`belt use and shoulder strap adjustment should also be examined further as they too may have significant effects on the demands
`placed on backpack users. Based on the present findings, school students should wear their backpacks with the least weight possible,
`use the hip-belt if present, allow a reasonable amount of looseness in the shoulder straps and should position the heaviest items
`closest to their back. However, more detailed work using human participants needs to be undertaken before these recommendations
`can be confirmed.
`r 2004 Elsevier Ltd. All rights reserved.
`
`Keywords: Strap tension; Pressure; Schoolbag; Simulator
`
`1. Introduction
`
`Growing suspicion that the loads school students
`carry to, around and from school are frequently too
`high has prompted research into the physical demands
`of school student’s load carriage (Chansirinukor et al.,
`2001; Cheung and Hong, 2000; Grimmer et al., 2002;
`Grimmer and Williams, 2000; Hong et al., 2000; Mackie
`et al., 2003; Malhoutra and Sen Gupta, 1965; Pascoe
`
` Corresponding author. Tel.: +64 9 815 4321x8012;
`fax: +64 9 815 6796.
`E-mail address: hmackie@unitec.ac.nz (H.W. Mackie).
`
`0003-6870/$ - see front matter r 2004 Elsevier Ltd. All rights reserved.
`doi:10.1016/j.apergo.2004.10.007
`
`et al., 1997; Sander, 1979; Voll and Klimt, 1977;
`Whittfield et al., 2001). However,
`it
`is difficult
`to
`demonstrate that loads carried by school students are
`directly associated with reported musculoskeletal pain
`or discomfort as there are many other factors such as
`physical capability, other physical activities, poor
`seating, growing pains or psychosocial
`factors that
`may contribute to reported pain or discomfort (Trous-
`sier et al., 1994; Watson et al., 2002).
`Researchers have therefore tended to study the effects
`of load carrying on physiological and biomechanical
`measures in children and adolescents such as oxygen
`consumption (Hong et al., 2000; Malhoutra and Sen
`
`Petitioner Ex. 1081 Page 1
`
`

`
`200
`
`H.W. Mackie et al. / Applied Ergonomics 36 (2005) 199–206
`
`ARTICLE IN PRESS
`
`Gupta, 1965), gait (Cheung and Hong, 2000; Pascoe et
`al., 1997; Wang et al., 2001) and posture (Chansirinukor
`et al., 2001; Grimmer et al., 2002; Grimmer and
`Williams, 2000; Malhoutra and Sen Gupta, 1965;
`Pascoe et al., 1997; Wang et al., 2001). Wang et al.
`(2001) also studied ground reaction forces in order to
`determine the effects of carrying school-related loads.
`Physiological and biomechanical measures such as
`oxygen consumption and gait are undoubtedly altered
`as a result of load carriage (Goldman and Iampietro,
`1962; Kinoshita, 1985; Knapik et al., 1996; Legg and
`Mahanty, 1985,1986) but whether these changes are
`indicative of eventual injury is unknown. Increases in
`oxygen consumption or increases in support phase time
`during gait may be the body’s natural way of safely
`accommodating the extra load placed on it.
`A more direct method of determining the physical
`demands of load carriage in school students would be to
`measure the external
`forces that directly relate to
`carrying a backpack, such as the pressure on the
`shoulders that occur as a result of the tension in the
`shoulder straps of a backpack. Bryant and Reid (1996)
`described a biomechanical model for the forces that act
`within the person/backpack system when load carrying.
`In this model the weight force of the backpack is resisted
`mostly by the resistive forces of the shoulders, hips and
`lower back via the shoulder straps and hip-belt. Given
`that using the hip-belt to increase the load on the hips is
`seen as positive during load carriage, measuring the
`forces at
`the shoulder during load carriage would
`provide a relevant indicator of the demands placed on
`the backpack user.
`The magnitude of the loads that school students carry
`has also been the focus of
`school
`load carriage
`researchers (Cheung and Hong, 2000; Hong et al., 2000;
`Malhoutra and Sen Gupta, 1965; Pascoe et al., 1997; Voll
`and Klimt, 1977; Whittfield et al., 2001), and 10% of
`body weight (BW) is generally accepted as a recom-
`mended maximum load for school students (Sander,
`1979; Voll and Klimt, 1977). Recently studies have shown
`that no significant changes in oxygen consumption or gait
`occur until school students are carrying 15–20% of BW
`(Cheung and Hong, 2000; Hong et al., 2000; Pascoe et al.,
`1997), which may support a school load carriage limit of
`10% BW. What seems more certain is that 20% BW as a
`load for school students is excessive (Cheung and Hong,
`2000; Hong et al., 2000).
`The variations reported in school student’s responses
`to carrying loads may be because a person’s carrying
`capacity is affected not only by the magnitude of the
`load they carry but also by the way the load is carried,
`the duration of carriage, the frequency of carriage and
`the physical capabilities of the person. These other
`factors must also be considered when attempting to
`determine the overall physical demands placed on
`the user.
`
`Bygrave et al. (2004) appear to be the only authors
`to have studied the adjustment of a single backpack
`in adults. They found that the tightness of fit of a
`backpack (adjustment in the shoulder straps, chest
`strap and hip-belt of 3 cm) had an effect on lung
`function in 12 healthy males wearing a 15 kg back-
`pack. Using different backpack designs Lloyd and
`Cooke (2000) and Kinoshita (1985) both found that
`distributing the weight of
`the backpack between
`the front and the back of the body lead to improve-
`ments in gait measures. In children, Grimmer et al.
`(2002)
`found that more
`loose
`shoulder
`straps
`allowed a more upright, natural posture than tighter
`shoulder straps where the backpack is carried higher
`on the back.
`Although these studies have addressed backpack
`configuration, no studies to date have attempted to
`study the effects of many different backpack adjust-
`ments on the backpack forces that directly affect school
`students. However, in order to carry out such a study, a
`large number of trials would need to be performed in
`order
`to test different combinations of backpack
`adjustments for each individual from a sample group
`large enough to account for the variation of results
`expected from human participants.
`Bryant et al. (2001) recommend that a load carriage
`simulator is useful
`in screening a large number of
`backpack designs or configurations prior to more
`detailed analyses using human participants. A load
`carriage simulator might, therefore, be an efficient way
`of evaluating a large number of school load carriage
`configurations, prior to a more detailed evaluation of
`potentially beneficial configurations using school stu-
`dents in the future. The objective of this study, there-
`fore, was to determine the effects of
`load weight,
`shoulder strap length,
`load distribution, gait speed,
`and the use of a hip-belt on shoulder strap tension forces
`and shoulder interface pressure during simulated school
`student’s load carriage.
`
`2. Methods
`
`All trials were conducted on a load carriage simulator
`that was designed and built by the Ergonomics Research
`Group at Queens University, Ontario, Canada and is
`the property of Defence Research and Development
`Canada (Stevenson et al., 2004). The load carriage
`simulator (Fig. 1) consists of a programmable three
`degree of freedom pneumatically driven platform, which
`supports interchangeable rigid mannequins. Vertical
`displacement, rotation about the anterior/posterior axis
`(side lean), and rotation about the medial/lateral axis
`(forward lean) are user programmable from a menu. A
`skin analogue (Bocklites) covers the surface of the
`mannequin.
`
`Petitioner Ex. 1081 Page 2
`
`

`
`ARTICLE IN PRESS
`
`H.W. Mackie et al. / Applied Ergonomics 36 (2005) 199–206
`
`201
`
`Load cell
`
`Load cell
`
`Fig. 1. Load carriage simulator used for data collection (tight shoulder
`straps configuration shown).
`
`Anterior/posterior lean of the mannequin is typically
`set by balancing the anterior–posterior moment due to
`backpack loads. In previous studies (Cheung and Hong,
`2000; Malhoutra and Sen Gupta, 1965; Pascoe et al.,
`1997) the change in anterior lean of the trunk in school
`children when carrying different loads has been shown
`to be very small or negligible until a load change of
`17–20% BW was administered. Therefore, in this study,
`the mannequin was fixed to the motor of the simulator
`with an anterior tilt of 51 (balanced in the anteroposter-
`ior plane) to maintain consistency between trials.
`A mannequin representing a 5th percentile Canadian
`armed forces female (weight 52.8 kg and height 1.55 m)
`was used (Fig. 1) as it most closely resembled the
`anthropometric characteristics of 13 year old school
`students, which have been reported as carrying the
`greatest loads across all school students (Grimmer and
`Williams, 2000, Pascoe et al., 1997; Whittfield et al.,
`2001) and therefore may be at the greatest risk of injury.
`A commercially available school backpack (Fig. 1), with
`no internal or external
`frame, but with adjustable
`shoulder straps and waist belt was used for the study.
`The backpack was modified to accommodate custom
`built load cells at the top and at the bottom of the
`
`shoulder straps so that tension could be measured at
`these points, giving an indication of
`the shoulder
`reaction force. The linearity of the load cells’ response
`to loading was tested up to 50 N. Correlation coeffi-
`cients of r ¼ 0:999 and 0:998 were determined for the
`bottom and top shoulder strap load cells, respectively.
`Forces were measured on the right side of the backpack
`while dummy load cells of identical dimensions were
`used on the left side to ensure the symmetry of the
`school backpack. The load cells were hardwired to an
`amplifier and personal computer and force data were
`collected at 20 Hz, which was the limit of the capability
`of the system.
`Shoulder pressure during load carriage has previously
`been measured using Tekscan pressure sensors (Martin
`and Hooper, 2000). A pressure sensor (Fscan 9811,
`Tekscan) was placed over the most superior aspect of
`the right shoulder of the mannequin so that changes in
`pressure due to forces from the shoulder straps could be
`measured. Gathering absolute quantitative data using
`this sensor when placed on a curved surface proved
`ineffective as the bending of the sensor created an offset,
`so only changes in raw pressure (the sum of
`the
`pressures measured in each of 96 pressure sensitive
`cells) was used. Raw pressure measurements were
`collected at 50 Hz using the same data acquisition
`software as the load cells. Extra precautions were taken
`by collecting unloaded baseline data from the Tekscan
`system before and after each trial, to account for any
`drift in the signal from the sensor.
`Both the load cells and the pressure sensors proved to
`be highly reliable. Correlation coefficients for test/re-test
`mean and peak forces were r ¼ 0:986 and 0:979;
`respectively. Correlation coefficients
`for
`test/re-test
`mean and peak pressures were r ¼ 0:945 and 0:956;
`respectively.
`The validity of the load carriage simulator’s ability to
`predict musculoskeletal discomfort in soldiers has been
`established by Bryant et al. (2001). Significant positive
`correlations were shown between shoulder pressure and
`forces on the simulator and soldier’s reported muscu-
`loskeletal discomfort. In the present study, statistically
`significant (Po0:01) correlation coefficients of r ¼ 0:556
`and 0:635 for mean and peak load cell/pressure sensor
`comparisons, respectively, demonstrated the validity of
`the overall measurement system. There appear to be no
`studies that demonstrate the validity of the simulator’s
`ability to reproduce human movement.
`Before each trial the backpack was placed on the
`mannequin in a standardised manner. Measurements
`between markers on the side and back of the neck of the
`mannequin and the shoulder strap and the top of the
`backpack were used to ensure consistent backpack
`placement.
`Five
`load carriage adjustment parameters were
`determined based on the variations of load carriage
`
`Petitioner Ex. 1081 Page 3
`
`

`
`202
`
`H.W. Mackie et al. / Applied Ergonomics 36 (2005) 199–206
`
`ARTICLE IN PRESS
`
`that school students were considered to most commonly
`experience. Gait speed (‘Walking’ and ‘running’), back-
`pack weight, load distribution, shoulder strap length
`and use of a hip-belt were manipulated so that 32
`possible combinations of load carriage configuration
`were evaluated.
`Simulator walking and running step rates (1.3 and 1.5
`steps per second, respectively) and centre of mass
`vertical displacements (4.5 and 6.0 cm, respectively)
`were used based on gait kinematics information from
`Unnithan and Eston (1990) and Rose and Gamble
`(1994). Step rate and centre of mass vertical displace-
`ment were the only programmable components of the
`simulator’s gait speed. It is acknowledged that only
`manipulating these two variables is not sufficient to
`realistically differentiate between real walking and
`running, however they are likely to have the greatest
`effect on the forces that effect the shoulder during load
`carriage.
`Backpack weights used were 10% (5.3 kg) and 15%
`(7.9 kg) of the representative BW of the mannequin.
`These weights were chosen as they represented the
`current recommended load carriage limit for school
`students (10% BW) and 5% greater than the recom-
`mended limit, so that the effects of heavier, yet realistic
`loads could be examined. Load distribution was termed
`as ‘close’ and ‘distant’. Five text books were used to
`pack the school backpack with the heaviest books
`closest to the back of the mannequin for the ‘close’ load
`distribution condition (centre of mass 5.5 cm from inner
`backpack wall) and the heaviest books farthermost from
`the back of
`the mannequin for the ‘distant’
`load
`distribution condition (centre of mass 11 cm from inner
`backpack wall). The shoulder straps were adjusted and
`checked using a tape measure before each trial, with the
`‘tight’ straps condition defined as a distance of 7 cm
`from the tip of the shoulder strap adjustment buckle to
`the lower connection of
`the shoulder strap to the
`backpack. This adjustment represented the backpack
`fitting close to the upper back (Fig. 1). The ‘loose’ straps
`condition, representing the backpack sitting lower on
`the back of the mannequin, was defined as a distance of
`24 cm from the tip of the shoulder strap adjustment
`buckle to the lower connection of the shoulder strap to
`the backpack. The hip-belt was either used or not used.
`When it was used the hip-belt tension was standardised
`to 13.6 kg using a Shimpo tensiometer before each trial.
`For each trial, the simulator was allowed to run for 10
`gait cycles, prior to data collection. Two 10 s trials were
`collected for each backpack configuration so that the
`reliability of the system could be evaluated. Between
`each trial, the backpack position on the mannequin was
`checked and adjusted if necessary.
`Pressure and force data were analysed using SPSS
`statistical analysis software. Data from the two trials for
`each load carriage configuration were combined and
`
`means and standard deviations were calculated both for
`the overall data and for the peaks in each cycle for each
`trial. Separate, single factor, within groups, analyses of
`variance (ANOVA) with an alpha level of 0.05 were
`used to compare the data between each variation of
`walking/running, backpack weight,
`load distribution,
`strap length and use of a hip-belt. Between groups
`ANOVA were used to test for interactions between
`backpack configurations.
`
`3. Results
`
`Tables 1 and 2 show the mean and standard deviation
`(SD) overall and peak shoulder
`strap forces and
`shoulder pressures for each variation of backpack
`weight, use of hip-belt, strap length, load distribution
`and walking/running. The percentage difference be-
`tween the means of each variation of overall and peak
`force and pressure is also shown along with the p-value,
`demonstrating the level of statistical significance of the
`differences between the means of each variation.
`Load weight had the greatest influence on shoulder
`strap forces with a load of 15% BW producing 50%
`greater overall force (po0:001) and 36% greater peak
`force (po0:001) than a load of 10% BW. This was
`followed by hip-belt use where the non-use of the hip-
`belt produced 40% greater overall forces (po0:001) and
`41% greater peak forces (po0:001) than when a hip-belt
`was used, and shoulder strap length where tight straps
`produced 37% greater overall forces (po0:001) but only
`10% greater peak forces (p ¼ 0:151) than loose shoulder
`straps.
`Variations in load placement and walking/running
`had much less effect on shoulder strap forces than load
`weight, hip-belt use and shoulder strap adjustment. For
`load placement, having the weight distributed farther-
`most away from the back only increased overall
`shoulder strap forces by 6% (p ¼ 0:494) and peak
`shoulder strap forces by 10% (p ¼ 0:143). For walking
`and running, running only increased overall shoulder
`strap forces by 1% (p ¼ 0:914) and peak shoulder strap
`forces by 8% (p ¼ 0:286).
`The pattern of results for shoulder pressure was
`similar to those shown for shoulder strap forces. Load
`weight had the greatest influence on shoulder pressure
`with a load of 15% BW producing 70% greater overall
`shoulder pressure (po0:001) and 65% greater peak
`shoulder pressure (po0:001) than 10% BW. This was
`followed by hip-belt use where the non-use of the hip-
`belt produced 44% greater overall shoulder pressure
`(p ¼ 0:001) and 47% greater peak shoulder pressure
`(po0:001) than when the hip-belt was used. For strap
`length,
`tight
`straps produced 40% greater overall
`shoulder pressure (po0:001) and 28% greater peak
`shoulder pressure (p ¼ 0:020) than loose straps.
`
`Petitioner Ex. 1081 Page 4
`
`

`
`ARTICLE IN PRESS
`
`H.W. Mackie et al. / Applied Ergonomics 36 (2005) 199–206
`
`203
`
`Table 1
`Mean and standard deviation (SD) overall and peak shoulder strap forces (Newtons) for different load carriage configurations
`
`Load carriage variable
`
`Adjustment 1
`
`Adjustment 2
`
`Load weight
`
`Hip belt
`
`Straps
`
`Load placement
`
`Gait speed
`
`10% of body weight
`Overall
`Peak
`22.5 (7.0)
`38.0 (10.1)
`
`15% of body weight
`Overall
`Peak
`33.8 (8.4)
`51.7 (9.6)
`
`% Diff overall
`
`% Diff peak
`
`50***
`
`36***
`
`Used
`Overall
`23.5 (10.1)
`
`Loose
`Overall
`23.8 (9.4)
`
`Close to back
`Overall
`27.4 (10.8)
`
`Walking
`Overall
`28.1 (9.8)
`
`Peak
`37.2 (10.6)
`
`Peak
`42.7 (14.1)
`
`Peak
`42.7 (13.1)
`
`Peak
`43.2 (11.4)
`
`not used
`Overall
`32.9 (6.2)
`
`Tight
`Overall
`32.5 (7.7)
`
`Peak
`52.5 (7.7)
`
`Peak
`47.0 (9.1)
`
`Distant from back
`Overall
`29.0 (8.3)
`
`Peak
`47.1 (10.5)
`
`Running
`Overall
`28.3 (9.5)
`
`Peak
`46.5 (12.5)
`
`40***
`
`41***
`
`37***
`
`6
`
`1
`
`10
`
`10
`
`8
`
`*Difference statistically significant (po0:05). **Difference statistically significant (po0:01). ***Difference statistically significant (po0:001).
`
`Table 2
`Mean and standard deviation (SD) overall and peak shoulder pressure (Raw pressure) for different load carriage configurations
`
`Load carriage variable
`
`Adjustment 1
`
`Adjustment 2
`
`Load weight
`
`Hip belt
`
`Straps
`
`Load placement
`
`Gait speed
`
`10% of body weight
`Overall
`222 (95)
`
`Peak
`271 (112)
`
`15% of body weight
`Overall
`378 (128)
`
`Peak
`446 (136)
`
`% Diff Overall
`
`% Diff Peak
`
`70***
`
`65***
`
`Used
`Overall
`246 (138)
`
`Loose
`Overall
`250 (138)
`
`Close to back
`Overall
`295 (151)
`
`Walking
`Overall
`336 (141)
`
`Peak
`290 (143)
`
`Peak
`315 (165)
`
`Peak
`352 (164)
`
`Peak
`390 (147)
`
`Not used
`Overall
`355 (114)
`
`Tight
`Overall
`350 (117)
`
`Distant from back
`Overall
`305 (122)
`
`Running
`Overall
`264 (124)
`
`Peak
`427 (129)
`
`Peak
`402 (125)
`
`Peak
`365 (140)
`
`Peak
`326 (152)
`
`44**
`
`47***
`
`40***
`
`3
`
`21*
`
`28*
`
`4
`
`16
`
`*Difference statistically significant (po0:05).**Difference statistically significant (po0:01).***Difference statistically significant (po0:001).
`
`For shoulder pressure, variations in load distribution
`again had much less effect on shoulder pressure than
`load weight, hip-belt use and shoulder strap adjustment.
`Having the weight distributed farthermost away from
`the back only increased overall shoulder pressure by 3%
`(p ¼ 0:772) and peak shoulder pressure by 4%
`(p ¼ 0:720). Walking and running had the opposite
`effect on shoulder pressure than it did on shoulder strap
`forces. Walking produced 21% more overall shoulder
`pressure (p ¼ 0:031) and 16% more peak shoulder
`pressure (p ¼ 0:096) than running.
`One interaction between load carriage adjustments
`was statistically significant. The interaction between the
`
`shoulder strap adjustment hip-belt use was statistically
`significant (po0:001) for overall and peak shoulder
`strap forces and shoulder pressure. The interaction
`meant that the loose shoulder strap adjustment was
`more effective in reducing shoulder forces when the hip-
`belt was worn.
`
`4. Discussion
`
`Load weight was clearly the most influential of the
`load carriage variables that were studied. This seems
`reasonable as the gravitational pull on the contents of
`
`Petitioner Ex. 1081 Page 5
`
`

`
`204
`
`H.W. Mackie et al. / Applied Ergonomics 36 (2005) 199–206
`
`ARTICLE IN PRESS
`
`the backpack due to the added load would have the
`greatest effect on the forces at the shoulder straps. More
`surprising was that the magnitude of the pressure on the
`shoulder increased disproportionately to the increase in
`load added to the backpack. The variation in backpack
`loads was 50% (10–15% BW), therefore forces and
`pressure at the shoulder would have been expected to
`increase by 50% in accordance with Newton’s law of
`reaction forces. The differences in overall and peak
`shoulder strap force between 10% and 15% BW were
`50% and 38%, respectively, which seems approximately
`proportional to the load increase, whereas the differ-
`ences in overall and peak shoulder pressure between
`10% and 15% BW were 70% and 65%, respectively,
`suggesting that the load might be increasingly demand-
`ing disproportionately to the weight carried. This might
`be explained by the frictional forces at the shoulder and
`back that partially support the load of the backpack
`having less of an effect at higher loads. This phenom-
`enon was reflected by Bryant and Reid (1996) who
`found that the proportion of the load weight being
`supported by the shoulders compared with other contact
`points on the body increased as the load weight
`increased. However,
`it must be remembered that
`currently,
`the
`relationship between frictional and
`pressure forces on the simulator compared with school
`students is not known. This is further complicated by
`the fact that for much of the shoulder area, the shoulder
`straps were in direct contact with the clothing rather
`than the Bocklites skin analogue, although Hooper
`and Jones (2002) suggested that clothing layers have no
`effect on the transmitted pressure to the skin.
`Although load weight had the greatest effect on
`shoulder forces and pressure, the use of a hip-belt and
`looser
`shoulder
`straps also significantly reduced
`shoulder forces. The effect of the hip-belt is under-
`standable as its use means that more of the weight is
`borne by the hips, lower back and abdominal region,
`therefore reducing the demands on the shoulders.
`The effect of looser shoulder straps is not as obvious.
`Perhaps the looser shoulder straps meant that there was
`more of each strap in contact with the body, which
`would lead to greater frictional forces and therefore less
`pressure on the shoulder. Alternatively, this phenomen-
`on might be explained by the fact that in the loose
`position the straps are pulling more vertically, which is
`in the direction required to counter the effects of gravity
`on the backpack, and would therefore require less
`overall force in the shoulder straps. A complication to
`this trend is that when the loose and tight shoulder
`straps data is further categorised by hip-belt use, the
`positive effects of looser shoulder straps is much greater
`when the hip-belt is worn. Likewise, the effect of the hip-
`belt appears to be more effective when loose shoulder
`straps are used. Simultaneous measurement of shoulder
`strap and hip-belt pressures in future research would
`
`more accurately describe how contact pressures are
`shared when the hip-belt is worn during schoolbag
`carriage. Although Jones and Hooper (2003) have
`measured pressure in shoulders and hips in response to
`military load carriage, this has never been carried out
`for school students.
`Grimmer et al. (2002), found that more loose straps
`allow school students to stand in a more up-right
`posture. Based on the findings to date, there may be
`some benefit in school students adjusting shoulder straps
`to a more loose position, especially if the hip-belt is
`used. Conversely, walking with a lower backpack centre
`of mass has been shown to cause greater forward lean in
`adults
`(Bloom and Woodhull-McNeal, 1987), and
`therefore further clarification is required.
`Load distribution had much less of an effect on the
`shoulder strap forces and pressure at the shoulder than
`load, hip-belt use and shoulder strap adjustment. The
`greater torque that is generated by distributing the
`heavy contents of the backpack further away from the
`back, should increase shoulder strap forces due to the
`increased resistance torque that the wearer must exert.
`However, the difference in weight distribution in this
`study was clearly not enough to invoke significant
`differences in overall or peak shoulder strap forces and
`shoulder pressure. By more greatly changing the weight
`distribution of a backpack via the use of balance pockets
`on the front of a backpack, both Kinoshita (1985) and
`Lloyd and Cooke (2000) found that positive differences
`in load carrying ability were obtained, which is more
`conclusive than the findings of the present study. The
`effects of increasing the distance of the centre of mass
`position from the back of the mannequin might have
`been better detected by measuring the lumbar force
`applied by the backpack.
`The effect of gait speed on shoulder strap forces and
`shoulder pressure was unexpected. It was expected that
`both shoulder strap forces and shoulder pressure would
`increase as gait speed increased in accordance with the
`Newton’s second law in which force is a function of
`mass and acceleration. The increased vertical accelera-
`tion from running should have produced greater
`shoulder forces. However, running produced signifi-
`cantly less overall pressure on the shoulders and there
`was no effect on peak shoulder strap forces and peak
`shoulder pressure. One possible explanation for this is
`that the different gait speeds produced different relative
`simulator and backpack movements due to different
`timing of the phases of the simulator and the backpack
`movement. If this is the case then the effects of phase
`differences in person-backpack movement on forces on
`the shoulder should be examined more thoroughly as it
`might justify the use of devices such as springy shoulder
`straps, which may promote such interactions.
`Currently, there are no normative data with which the
`results of the present study can be compared, apart from
`
`Petitioner Ex. 1081 Page 6
`
`

`
`ARTICLE IN PRESS
`
`H.W. Mackie et al. / Applied Ergonomics 36 (2005) 199–206
`
`205
`
`the increasing evidence that a load of 17% of BW may
`be excessive and a load of 20% of bodyweight almost
`definitely is excessive for school students (Cheung and
`Hong, 2000; Hong et al., 2000; Pascoe et al., 1997). If
`shoulder strap forces were found to increase by at least
`50% when the load was increased from a currently
`‘acceptable’ load of 10% BW to a ‘possibly unaccep-
`table’ load of 15% BW in the present study, then the
`increases in shoulder forces and pressure of approxi-
`mately 40% as a result of tight shoulder straps or not
`wearing a hip-belt, must also be significant enough to
`affect the wearer.
`A statistically significant interaction was observed
`where the benefit of looser shoulder straps was greatly
`improved when the hip-belt was used. This phenomenon
`may be explained by the hip-belt controlling the load
`and preventing relative movement between the back-
`pack. There may be other, more subtle interactions that
`also exist, which should be further studied in more
`detail, using human participants.
`Limitations to the present study include measuring
`force at 20 Hz and the validity of using Tekscan pressure
`sensitive pads on a curved surface. However, the effects
`of these limitations on the findings have been minimised
`by reproducing identical trials, only using changes in
`shoulder pressure and measuring baseline values for
`shoulder pressure prior to data collection.
`Another limitation of this study includes the unknown
`ability of the load carriage simulator to accurately
`reproduce human movement and posture, and respond
`to contact pressures. However, the main purpose of the
`load carriage simulator is not to perfectly reproduce
`human movement, but to allow highly reproducible
`comparisons of load carriage systems. In addition, the
`ability of the simulator to predict soldier’s musculoske-
`letal discomfort as demonstrated by Bryant et al. (2001)
`indicates a positive relationship between simulated and
`human backpacking.
`It
`is unknown whether
`this
`relationship is also true for school students.
`A similar study to the present one, using human
`participants to examine these findings in a more realistic
`setting would be useful, however, the logistical implica-
`tions of conducting such a study with remotely near the
`same reliability as the simulator used in the present
`study are enormous. A combination of the two methods,
`as suggested by Bryant et al. (2001), where the simulator
`is used to screen large numbers of different load carriage
`adjustments prior to more specific human-based inves-
`tigations might be appropriate.
`
`5. Conclusion
`
`Load weight, hip-belt use, and shoulder strap length
`had the greatest effects on shoulder strap tension forces
`and shoulder interface pressure. Load distribution had
`
`much less of an effect on shoulder forces, however,
`keeping the load close to the back may still assist in
`reducing discomfort and perhaps injury. It is unclear
`that what effect gait speed had on shoulder forces. Based
`on the demands placed on the shoulder as a result of
`simulated load carriage, school students should limit the
`amount of weight carried, use a hip-belt, adjust the
`shoulders straps to a fairly loose position and perhaps
`position the heaviest items closest to the back. However,
`more detailed work with human participants needs to
`be conducted before these recommendations can be
`confirmed.
`
`Acknowledgement
`
`Part funding for this study was provided by the Royal
`Society of New Zealand.
`
`References
`
`Bloom, D., Woodhull-McNeal, A.P., 1987. Postural adjustments while
`standing with two types of loaded backpack. Ergonomics 30 (10),
`1425–1430.
`Bryant, J.T., Reid, J.G., 1996. A biomechanical model of load
`carriage. Paper presented at
`the Ninth Biennial Conference,
`Canadian Society of biomechanics, Vancouver.
`Bryant, J.T., Doan, J.B., Stevenson, J.M., Pelot, R.P., 2001.
`Validation of objective based measures and development of a
`performance-based ranking method for load carriage systems. In:
`Proceedings of the Soldier mobility: Innovations in Load Carriage
`System Design and Evaluation Conference, K