This article was downloaded by:[Bibliotheek Vrije Universiteit Amsterdam]On: 25 June 2008Access Details: [subscription number 789286654]Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

ErgonomicsPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t713701117

Effect of lifting height and load mass on low backloadingMarco J. M. Hoozemans a; Idsart Kingma a; Wiebe H. K. de Vries b; Jaap H. vanDieën aa Research Institute MOVE, Faculty of Human Movement Sciences, VU UniversityAmsterdam, Amsterdam, The Netherlandsb Roessingh Research & Development, Enschede, The Netherlands

Online Publication Date: 01 July 2008

To cite this Article: Hoozemans, Marco J. M., Kingma, Idsart, de Vries, Wiebe H.K. and van Dieën, Jaap H. (2008) 'Effect of lifting height and load mass on lowback loading', Ergonomics, 51:7, 1053 — 1063

To link to this article: DOI: 10.1080/00140130801958642URL: http://dx.doi.org/10.1080/00140130801958642

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction,re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expresslyforbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will becomplete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should beindependently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings,demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with orarising out of the use of this material.

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

E!ect of lifting height and load mass on low back loading

Marco J.M. Hoozemansa*, Idsart Kingmaa, Wiebe H.K. de Vriesb andJaap H. van Dieena

aResearch Institute MOVE, Faculty of Human Movement Sciences, VU University Amsterdam, Vander Boechorststraat 9, 1081 BT, Amsterdam, The Netherlands; bRoessingh Research & Development,

Roessinghsbleekweg 33b, 7522 AH, Enschede, The Netherlands

The objective of this study was to quantify the e!ect of lifting height and mass lifted onthe peak low back load in terms of net moments, compression forces and anterior–posterior shear forces. Ten participants had to lift a box using four handle heights. Lowback loading was quantified using a dynamic 3-D linked segment model and a detailedelectromyographic driven model of the trunk musculature. The e!ects of lifting heightand lifting mass were quantified using a regression technique (GEE) for correlated data.Results indicate that an increase in lifting height and a decrease in lifting mass wererelated to a decrease in low back load. It is argued that trunk flexion is a majorcontributor to low back load. For ergonomic interventions it can be advised to prioritiseoptimisation of the vertical location of the load to be lifted rather than decreasing themass of the load for handle heights between 32 cm and 155 cm, and for load massesbetween 7.5 and 15 kg. Lifting height and load mass are important determinants of lowback load during manual materials handling. This paper provides the quantitative e!ectof lifting height and mass lifted, the results of which can be used by ergonomists at theworkplace to evaluate interventions regarding lifting height and load mass.

Keywords: lifting; lifting height; low back; spine; ergonomics

1. Introduction

Work-related physical and psychosocial factors are generally acknowledged to be relatedto low back complaints (Bongers et al. 1993, Bernard 1997, Burdorf and Sorock 1997,Hoogendoorn et al. 1999, 2000, Kuiper et al. 1999, Kerr et al. 2001). Among the physicalrisk factors, scientific evidence indicates manual lifting as a strong predictor of thedevelopment of low back complaints at work. Despite intensive automation andmechanisation, heavy lifting is still frequently performed in, for instance, industry andhealth care. To develop e!ective ergonomic interventions to prevent low back complaintsassociated with lifting loads, scientific research has been aimed at identifying aspects oflifting that a!ect low back loading.

Psychophysical, physiological and biomechanical criteria have been used to developthe National Institute for Occupational Safety and Health (NIOSH) work practices guidefor manual lifting (National Institute for Occupational Safety and Health 1981, Waterset al. 1993). The NIOSH lifting equation takes into account several aspects of a liftingtask, among which the initial vertical location of the load is one of the more important

*Corresponding author. Email: m.hoozemans@fbw.vu.nl

ErgonomicsVol. 51, No. 7, July 2008, 1053–1063

ISSN 0014-0139 print/ISSN 1366-5847 online! 2008 Taylor & FrancisDOI: 10.1080/00140130801958642http://www.informaworld.com

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

determinants of the recommended weight limit. However, surprisingly little scientificresearch has been aimed at exploring the quantitative relationship between lifting heightand low back loading.

Psychophysical experiments were performed to determine the e!ects of lifting heighton the maximum acceptable mass to be lifted by a predefined population of workers(Snook 1978, Snook and Ciriello 1991, Mital et al. 1997). In these experiments that formeda basis for the NIOSH lifting equation, indeed the lifting height was found to be one of themajor factors a!ecting the workload during lifting. Psychophysical workload can beexpected to be closely related to energetic workload, which may manifest itself in terms of(local or whole body) fatigue symptoms. When loads are lifted to upright standing, anincrease in lifting height (starting position) will result in lower energetic costs because oflower potential energy di!erences between starting and ending positions of (upper) bodyand load.

Other studies focused on biomechanical e!ects of lifting height. Kee and Chung (1996),comparing several models for the prediction of low back load, did not observe a significante!ect of lifting height on low back load in two models, while the third electromyography(EMG)-assisted model showed that the low back load significantly increased with anincrease in lifting height (vertical distances of 72, 97 and 122 cm resulted in compressiveforces of 3050, 3600 and 3950 N respectively). Using the surface EMG of the erectorspinae Nielsen et al. (1998) studied the muscular load during lifting with variouscombinations of lifting height and lifting frequency. They concluded that the maximumload on the low back occurred at the low lifting heights. In a study on the evaluation of thee!ect of transferring boxes from di!erent positions on a pallet, Marras et al. (1999)observed an increase in maximum compression force and maximum anterior–posteriorshear force when lifting boxes closer to the ground. In addition, Lavender et al. (2003)quantified the e!ect of lifting height on low back loading. They found that the peak netmoment significantly increased when lifting from lower lifting heights and concluded fromthe regression model developed that the vertical location from which lifts are initiated isone of the principle factors determining the magnitude of the peak dynamic L5-S1 sagittalplane moment.

Building upon the study of Lavender et al. (2003), the relationship between liftingheight and low back loading could be further explored with regard to the quantification ofspine compression and shear forces. While net moments can be estimated in a valid waywith minimal (anthropometrical) assumptions, estimates of spinal forces require more andonly partially validated assumptions (van Dieen and Kingma 2005). However, comparedto net moments, these forces can be interpreted more directly in terms of injury risk. Thisstudy focuses on peak low back load as the literature indicates that peak values of lowback load are more powerful predictors of injury risk (van Dieen and Toussaint 1997) andof low back pain (Norman et al. 1998) than cumulative or average loads. Furthermore,usability of the results could be improved by including lifting height as a continuousvariable in a regression model. The advantage of quantifying lifting height as a continuousvariable over a qualitative approach using a categorical variable (i.e. knee height, hipheight or shoulder height) is that values not incorporated in the experimental design, butwithin the range of conditions tested, can be interpolated. Eventually, the method mightbe applicable at the workplace and be used by ergonomists in the evaluation of liftingtasks.

Although the present study primarily focuses on the e!ect of lifting height on low backload, the load mass has to be taken into account. When lifting from floor height, low backload is mainly determined by external moments caused by the load and the upper body.

1054 M.J.M. Hoozemans et al.

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

When lifting height increases, the contribution of the load mass to the total externalmoment changes, as also the contribution of the upper body changes due to a decrease intrunk inclination. Therefore, the relationship between lifting height and low back load islikely to be influenced by load mass. Whether load mass should be considered a covariateor an e!ect modifier (interaction) is not clear.

Therefore, the objective of the present study was to quantify the e!ect of lifting heightand lifting mass on the peak low back load in terms of net moments, compression forcesand anterior–posterior shear forces.

2. Methods

2.1. Participants

A total of 10 healthy young males, age 24.8 (SD 2.9) years, stature 1.80 (SD 0.11) m, bodymass 73.8 (SD 12.8) kg, participated in the experiments. All participants gave informedconsent prior to the experiments and reported no history of low back pain or othermusculoskeletal problems.

2.2. Tasks and procedures

Prior to the actual experiment, participants performed seven attempted maximumisometric contractions of the trunk muscles three times as described by McGill (1991). Toavoid fatigue e!ects the 2–3 s contractions were separated by 1 min rest breaks. For eachmuscle, the maximum EMG amplitude over the 21 measurements was considered to be themaximum voluntary contraction (MVC) value.

During the experiment, participants were standing on a force plate in front of abookcase with four shelves. This standard and commercially available bookcase was usedto mimic everyday lifting environments. Participants had to lift symmetrically with twohands a box (width 6 depth 6 height ! 450 6 360 6 230 mm) from one of the fourshelves to an upright standing posture holding the box at waist height. On top of the box,at both the left and right side, holes were present so that the top of the box had twohandles. Participants were specifically instructed to use the handles of the box. No specificinstructions regarding foot placement, lifting technique and speed of the lift were given.Participants had to perform lifts with the box weighing 7.5 kg and starting at handleheights of 0.32 m, followed by 0.73, 1.14 and 1.55 m respectively. Subsequently, theparticipants had to lift a box of 15 kg following the same sequence of handle heights. Amaximum of 15 kg was chosen to allow the participants to repetitively lift the loadsthroughout the experiment without risking fatigue e!ects and to avoid the risk of low backproblems afterwards. Adding a lower load mass of 7.5 kg to the experiment was expectedto be su"cient to gain insight into the relationship between lifting height, load mass andlow back load. Load masses of 7.5 and 15 kg are just about the average load mass of 11 kgfound in industry as reported by Dempsey et al. (2000).

2.3. Three-dimensional linked segment model

A dynamic 3-D linked segment model, described in detail by Kingma et al. (1996), wasused to estimate net moments at the level of the L5-S1 intervertebral disc. The current,slightly modified, model uses anthropometrical data according to McConville andChurchill (1980), combined with force plate data and kinematics from LED markers oncu!s on the feet, lower legs, upper legs, pelvis and thorax (the latter segment for

Ergonomics 1055

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

kinematics only). As anthropometrical data were based on male subjects, only malesparticipated in the present study. Marker positions were recorded using Optotrak (System3020; Northern Digital, Waterloo ON, Canada) and stored at 50 Hz. Kinematic data werelow-pass filtered at 10 Hz. A global equation of motion was used, as described by Hof(1992), except that the moment about the vertical measured by the force plate was added(Kingma and van Dieen 2004). The use of this global equation of motion allows the use ofone instead of two force plates. A large, custom-made 1 6 1 m strain gauge forceplatform was used. Force plate data were synchronised to the Optotrak system and storedat 50 Hz.

2.4. Three-dimensional electromyography-driven trunk model

A total of 12 pairs of surface EMG electrodes (Blue Sensor; lead-o! area 1.0 cm2, inter-electrode distance 2.5 cm) were attached to the skin after abrasion and cleaning withalcohol. Electrodes were bilaterally attached over the internal oblique, external oblique,rectus abdominis, iliocostalis lumborum, and over the longissimus thoracis parslumborum and pars thoracis (for exact location see van Dieen and Kingma 2005).Signals were amplified 20 times (Porti-17TM; TMS, Enschede, The Netherlands; inputimpedance 41012 O, CMRR 4 90 dB), band-pass filtered (10–400 Hz), A-D converted(22bits at 1000 Hz) and synchronised with force plate and Optotrak recordings.

EMG signals were full-wave rectified and low-pass filtered at 2.5 Hz. Subsequently, thesignals were normalised to the highest values recorded in the MVC tests and non-linearlytransformed into an estimate of muscle activation as described by Potvin et al. (1996). Toestimate forces acting on L5-S1, a model of the trunk muscles was used, which has beendescribed in more detail by van Dieen (1997) and van Dieen and Kingma (2005). Toobtain the compression and shear forces at the L5-S1 intervertebral joint, forces due to themuscle force (calculated by the EMG-driven model) were added to the net reaction forcesthat had been calculated by the linked segment model.

2.5. Statistical analyses

The 3-D components of the net moment and the net reaction force at the L5-S1intervertebral disc were used to calculate the resultant net moment, the compression forceand the (anterior–posterior) shear force at L5-S1. Shear forces were considered positivewhen they would move the fifth lumbar vertebra anterior relative to the sacrum. For eachtrial, the peak values of these variables were determined.

The e!ects of lifting height and lifting mass, both defined as continuous variables, onthe maximum net moment, compression force and shear force at L5-S1 were quantifiedusing generalised estimating equations (GEE) (Liang and Zeger 1993). This analysisconsiders the measurements within participants as repeated measurements and account forthis dependency. In the GEE analysis the regression equation is defined as:

Outcome ! constant " B1 # load mass $kg% " B2 # handle height $m%" B3 # load mass $kg% # handle height $m%

where B1, B2 and B3 are regression coe"cients and the constant represents the value of theoutcome measure (maximum net moment, maximum compression force, maximum shearforce) when lifting a load mass of zero kg from the ground (handle height is zero). The

1056 M.J.M. Hoozemans et al.

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

interaction between load mass and lifting height was screened separately for significanceand only included in the model when significant. A significance level of 5% was used.

3. Results

With an increase in lifting height, the maximum low back moment decreased significantly(Table 1, Figure 1). The regression analysis predicted that per 10 cm increase in handleheight the low back moment would decrease by 7.0 Nm. The e!ect of the mass of the loadon the low back moment was predicted to be 2.8 Nm per kg. Corrected for lifting height,the increase in low back moment when lifting 15 kg instead of 7.5 kg was 21.0 Nm.Furthermore, the regression equations suggest that the low back moment is largelydetermined by body mass, as the constant of the regression was estimated to be 166.4 Nm,which represents the low back load when lifting zero kg with a handle height of zerometres.

Results of the maximum compression force at the L5-S1 intervertebral disc were highlycomparable to those of the maximum low back moment (Table 1, Figure 2). Again, liftingheight as well as the mass of the load significantly a!ected the compression force. Theregression analysis indicated that the compression force would decrease by 146.9 N per10 cm increase in handle height and that the compression force would increase by 54.5 Nfor every kg that is added to the mass to be lifted. For lifting zero kg from the ground, theregression analysis predicted a compression force of 3817.0 N.

The maximum shear force at L5-S1 was significantly a!ected by lifting height but notby the mass of the load (Table 1, Figure 3). The results of the regression analysis showed aconstant of 869.9 N (lifting zero kg from the ground) and a decrease in shear force of15.7 N with an increase of 10 cm in handle height.

4. Discussion

The present study aimed at quantifying the e!ect of lifting height and lifting mass on thepeak low back load. Results generally indicated a decrease in low back load with anincrease in lifting height, and an increase in low back load with an increase in lifting mass.These conclusions are comparable to those stated by most others (Nielsen et al. 1998,Marras et al. 1999, Lavender et al. 2003). The additional value of the present study is thequantification of the relationships using a regression technique for correlated data and by

Table 1. Results of the generalised estimating equation regression analyses quantifying the e!ect ofload mass (kg) and handle height (m) on the net moment (Nm), compression force (N) and shearforce (N) at the L5-S1 intervertebral disc.

Constant Load mass (kg) Handle height (m)

Low back moment (Nm) Coe"cient 166.4 2.8 770.4SE 17.2 0.7 8.4

p-value 50.001 50.001 50.001

Compression force at L5-S1 (N) Coe"cient 3817.0 54.5 71468.7SE 375.2 15.7 137.4

p-value 50.001 0.001 50.001

Shear force at L5-S1 (N) Coe"cient 869.9 4.0 7156.8SE 105.5 4.3 64.3

p-value 50.001 0.352 0.015

Ergonomics 1057

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

including the compression and shear forces as dependent variables, determined using adynamical 3D EMG-assisted biomechanical model.

Lavender et al. (2003) used a univariate regression model to predict the net moment atL5-S1 using lifting height, lifting speed and lifting mass:

L5& S1 moment ! 143:6 " 0:21 # mass $N% & 35:8 # height " 17:1 # speed;

Figure 1. Mean (n ! 10) and standard deviation (indicated by error bars) of the maximum netmoment at the low back (L5-S1) determined using a dynamic 3D linked segment model for lifting 2load masses (7.5 kg, 15 kg) at 4 handle heights (0.32 m, 0.73 m, 0.14 m, 1.55 m).

Figure 2. Mean (n ! 10) and standard deviation (indicated by error bars) of the maximumcompression force at the low back (L5-S1) determined using a 3D EMG assisted trunk model forlifting 2 load masses (7.5 kg, 15 kg) at 4 handle heights (0.32 m, 0.73 m, 1.14 m, 1.55 m).

1058 M.J.M. Hoozemans et al.

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

with mass (N), height defined as 1 (floor level), 2 (knee level) or 3 (knuckle level) and speeddefined as 0 (normal) and 1 (fast). Although it was not stated whether a regressiontechnique for correlated data was used and although defining lifting height as a three levelcategorical variable (coded as 1, 2 and 3) limits the validity of the model, the regressionequation, i.e. the regression coe"cients, seems to be comparable to the equation observedin the present study, except for the regression coe"cient of height. For lifting a 7.5 kgmass from the floor, knee and knuckle level, the Lavender et al. (2003) regression equationpredicts net moments of 123 Nm, 87 Nm and 53 Nm respectively, while in the presentstudy 187 Nm, 152 Nm and 125 Nm are predicted. Di!erences are mainly caused bydi!erences in the regression coe"cient of height, which may be the result of defining heightas a categorical variable with three levels in the study of Lavender et al. Although theregression considers di!erences 1–2 and 2–3 to be equal, the actual di!erences floor–kneeheight and knee–knuckle height are not, which will a!ect the regression analysis. Otherdi!erences may be caused by di!erences in the performance of the participants.Experimental setup and the participants themselves may have caused di!erences in liftingtechnique and thus low back loading. However, biomechanical studies indicate that liftingtechnique a!ects low back load only to a minor extent (van Dieen et al. 1999, Kingmaet al. 2004). Despite these di!erences and despite the application of di!erentbiomechanical models, these reasonably comparable equations strengthen the conclusionconsidering the e!ect of lifting mass and lifting height on low back loading.

In the present study, lifting height was defined as the height of the handles of the box.It was observed that participants always used both handles during lifting. However, theactual location of the hands on the handles of the box might have varied somewhatbetween lifting heights. It is expected that defining lifting height in terms of hand heightwould not have a!ected the regression coe"cients considerably, since the di!erence invertical position of handles and hands can at most be several cm. To allow use of theregression equations as a tool to quantify e!ects of lifting height and load mass at the

Figure 3. Mean (n ! 10) and standard deviation (indicated by error bars) of the maximum shearforce at the low back (L5-S1) determined using a 3D EMG assisted trunk model for lifing 2 loadmasses (7.5 kg, 15 kg) at 4 handle heights (0.32 m, 0.73 m, 1.14 m, 1.55 m).

Ergonomics 1059

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

workplace, it was decided to define the independent variables in terms of workplacevariables, i.e. lifting mass and handle height.

The values of the constants in the regression equations, i.e. 166.4 Nm, 3817.0 N and869.9 N, are the estimated values representing the low back load when getting up from aposition with the hands at floor level (zero handle height and zero load mass). Theseconstants are, however, only validated for the (lower) range of handle height studies, i.e.0.32–1.55 m. Assuming that this relationship continues linearly outside of the rangestudied, it indicates that just reaching down to near the ground would result in relativelyhigh low back loads, which exceed some suggested thresholds such as the NIOSH limit of3400 N. Examining the quantified e!ect of the mass of the load – for instance, when lifting10 kg more, the net moment, compression forces and shear force would increase by28 Nm, 545 N and 40 N respectively – suggests that, within the ranges of lifting height andload mass investigated in the present study, the low back load is mainly determined by themoment due to the upper body mass and not by lifting mass. This conclusion is supportedby the upper body kinematics. Figure 4 presents the maximum flexion of the trunk foreach lift at the four handle heights. Trunk flexion is hereby defined as the angle (8) betweenthe trunk and the vertical minus the angle between the pelvis and the vertical. An angle of08 would be a straight back. The figure shows that trunk flexion increases with lower liftingheights. Furthermore, maximum trunk flexion is hardly a!ected by load mass. Althoughstandard deviations indicate that di!erences in technique between participants areapparent, these data confirm that the contribution of the upper body to the moment at thelow back increases with a decrease in lifting height and that this contribution will besubstantial at low lifting heights. For ergonomic interventions, it can, therefore, be advisedto prioritise optimisation of the lifting height, e.g. the vertical factor in the NIOSH liftingequation, rather than decreasing the mass to be lifted, again within the ranges for whichthe regression equations are validated. When actually comparing the NIOSH lifting

Figure 4. Mean (n ! 10) and standard deviation (indicated by error bars) of the maximum flexionof the trunk for lifting 2 load masses (7.5 kg, 15 kg) at 4 handle heights (0.32 m, 0.73 m, 1.14 m,1.55 m). Trunk flexion is defined at the angle between the pelvis and the vertical. An angle of zerodegrees would be a straight back.

1060 M.J.M. Hoozemans et al.

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

equation to the results of the present study, an optimum in the e!ect of lifting height, assuggested by the vertical factor in the NIOSH equation (Vf ! 1 7 (0.003 6 j V 775 j )), could not be observed. For handle heights between 32 cm and 155 cm and for loadmasses between 7.5 and 15 kg, a linear relationship was found in which the low back loaddecreases with an increase in lifting height. The data did not indicate that a non-linearrelationship should have been applied (Figures 1–3). The non-linearity in the NIOSHequation is based on psychophysical experiments and likely reflects increased shoulderloads at high lifting heights.

In addition to the net moment, the compression and shear forces were also estimatedusing an EMG-assisted biomechanical model. As suggested by Lavender et al. (2003),including the compression and shear forces as dependent variables, when examining thee!ect of lifting height and mass lifted, is warranted to account for the possible presence ofco-contraction of the trunk muscles. However, a considerable level of co-contraction wasnot observed. This is in line with previous studies that concluded that, for lifting, results ofEMG-assisted biomechanical models are comparable to results of optimisation models(van Dieen and Kingma 2005). Data presented by Potvin (1997) on regression-basedmodels that use the NIOSH lifting equation horizontal and vertical factors to calculate L5-S1 compression forces were also used to estimate the relationship between peakcompression on the one hand and lifting height and mass on the other. With horizontalfactors of 0.5 m and a body mass of 75 kg this relationship was estimated as:

L5& S1 compression ! 3133 " 103 # mass $kg% & 1252 # height $m; hand height%:

This is reasonably comparable to what was observed in the present study. For lifting a10 kg mass at 0.30 m, the model of Potvin predicts a compression force of 3787 N whilethe model of the present study predicts a compression force of 3921 N. Di!erences may becaused by di!erences between the groups of participants and by the fact that theparticipants of Potvin had to lift only 5 kg. It is likely that lifting movement patternschange when heavier masses are lifted. Furthermore, in the study of Potvin, participantswere instructed to lift with either a stoop technique or a squat technique. In the presentstudy, instructions as to lifting technique and location of the feet were not given to be ableto observe motor behaviour likely to be displayed at the workplace. Also, in the presentstudy participants had to lift from the shelves of a bookcase. They might have adapted thelifting technique to, for instance, prevent knocking their heads.

Although it is still di"cult to interpret moments, compression forces and shear forcesas used in the regression equations in terms of risk of low back pain as epidemiologicalevidence is still insu"cient, the results of the study of Norman et al. (1998) might be usedto get an indication of an increase in risk in relation to changes in the variables of theregression equations (peak moment, compression force and shear force). Norman et al.(1998) found odds ratios of around 2 for di!erences in peak moment, peak compressionand peak shear of 80 Nm, 1348 N and 203 N respectively. For peak compression forcesthis means that a decrease in lifting height of 92 cm or an increase in load mass of 25 kgcould result in a two-fold higher risk of low back pain.

In conclusion, it can be stated that an increase in lifting height is associated with adecrease in low back load, independent of the mass lifted, and that heavier lifting massesare associated with higher low back loading. For handle heights between 32 cm and155 cm, and for load masses between 7.5 and 15 kg, the findings of this study imply thatfor the ergonomic optimisation of lifting activities at the workplace the vertical location of

Ergonomics 1061

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

the load is more important to consider than the mass lifted. Obviously, other factorsrelated to lifting, such as lifting frequency and the capacity of the workers, but notaddressed in this study, may be additional determinants of low back load and as suchdeserve attention in prevention of low back pain.

References

Bernard, B.P., 1997. Musculoskeletal disorders and workplace factors: a critical review ofepidemiological evidence for work-related musculoskeletal disorders of the neck, upper extremity,and low back. Washington, DC: National Institute for Occupational Safety and Health, USDepartment of Health and Human Services (DHHS), 97–141.

Bongers, P.M., et al., 1993. Psychosocial factors at work and musculoskeletal disease. ScandinavianJournal of Work, Environment and Health, 19, 297–312.

Burdorf, A. and Sorock, G.S., 1997. Positive and negative evidence of risk factors for back disorders.Scandinavian Journal of Work, Environment and Health, 23, 243–256.

Dempsey, P.G., et al., 2000. Field evaluation of the revised NIOSH lifting equation. In: Proceedingsof the IEA 2000/HFES 2000 congress, San Diego, USA (Santa Monica, CA: Human Factors andErgonomics Society).

Hof, A.L., 1992. An explicit expression for the moment in multibody segments. Journal ofBiomechanics, 25, 1209–1211.

Hoogendoorn, W.E., et al., 1999. Physical load during work and leisure time as risk factors for backpain. Scandinavian Journal of Work, Environment and Health, 25, 387–403.

Hoogendoorn, W.E., et al., 2000. Systematic review of psychosocial factors at work and private lifeas risk factors for back pain. Spine, 15, 2114–2125.

Kee, D. and Chung, M.K., 1996. Comparison of prediction models for the compression force on thelumbosacral disc. Ergonomics, 39, 1419–1429.

Kerr, M.S., et al., 2001. Biomechanical and psychosocial risk factors for low back pain at work.American Journal of Public Health, 91, 1069–1075.

Kingma, I. and van Dieen, J.H., 2004. Lifting over an obstacle: e!ects of one-handed lifting andhand support on trunk kinematics and low back loading. Journal of Biomechanics, 37, 249–255.

Kingma, I., et al., 1996. Validation of a full body 3-D dynamic linked segment model. HumanMovement Science, 15, 833–860.

Kingma, I., et al., 2004. Foot positioning instruction, initial vertical load position and liftingtechnique: e!ects on low back loading. Ergonomics, 47, 1365–1385.

Kuiper, J.I., et al., 1999. Epidemiological evidence on manual materials handling as a risk factor forback disorders: a systematic review. International Journal of Industrial Ergonomics, 24, 389–404.

Lavender, S.A., et al., 2003. The e!ects of initial lifting height, load magnitude, and lifting speed onthe peak dynamic L5/S1 moments. International Journal of Industrial Ergonomics, 31, 51–59.

Liang, K.-Y. and Zeger, S.L., 1993. Regression analysis for correlated data. Annual Review of PublicHealth, 14, 43–68.

McConville, J.T. and Churchill, T.D., 1980. Anthropometric relationships of body and bodysegment moments of inertia. AFAMRL-TR-80–119. Air Force Aerospace Medical ResearchLaboratory (Dayton, OH: Aerospace Medica Divison, Wright-Patterson Air Force Base).

McGill, S.M., 1991. Electromyographic activity of the abdominal and low back musculature duringthe generation of isometric and dynamic axial trunk torque: implications for lumbar mechanics.Journal of Orthopaedic Research, 9, 91–103.

Marras, W.S., et al., 1999. E!ects of box features on spine loading during warehouse order selecting.Ergonomics, 42, 980–996.

Mital, A., Nicholson, A.S., and Ayoub, M.M., 1997. A guide to manual materials handling. London:Taylor & Francis.

National Institute for Occupational Safety and Health, 1981. Work practice guide for manual loadlifting. Washington, DC: NIOSH.

Nielsen, P.K., Andersen, L. and Jorgensen, K., 1998. The muscular load on the lower back andshoulders due to lifting at di!erent lifting heights and frequencies. Applied Ergonomics, 29, 445–450.

Norman, R., et al., 1998. A comparison of peak vs cumulative physical work exposure risk factorsfor the reporting of low back pain in the automotive industry. Clinical Biomechanics, 13, 561–573.

1062 M.J.M. Hoozemans et al.

Dow

nloa

ded

By:

[Bib

lioth

eek

Vrij

e U

nive

rsite

it A

mst

erda

m] A

t: 12

:03

25 J

une

2008

Potvin, J.R., 1997. Use of NIOSH equation inputs to calculate lumbosacral compression forces.Ergonomics, 40, 691–707.

Potvin, J.R., Norman, R.W., and McGill, S.M., 1996. Mechanically corrected EMG for thecontinuous estimation of erector spinae muscle loading during repetitive lifting, European Journal ofApplied Physiology, 74, 119–132.

Snook, S.H., 1978. The design of manual handling tasks. Ergonomics, 21, 963–985.Snook, S.H. and Ciriello, V.M., 1991. The design of manual handling tasks: revised tables of

maximum acceptable weights and forces. Ergonomics, 34, 1197–1213.van Dieen, J.H., 1997. Are recruitment patterns of the trunk musculature compatible with a synergy

based on the maximization of endurance? Journal of Biomechanics, 30, 1095–1100.van Dieen, J.H., Hoozemans, M.J.M., and Toussaint, H.M., 1999. Stoop or squat: a review of

biomechanical studies on lifting technique. Clinical Biomechanics (Bristol, Avon), 14, 685–696.van Dieen, J.H. and Kingma, I., 2005. E!ects of antagonistic co-contraction on di!erences between

electromyography based and optimization based estimates of spinal forces. Ergonomics, 48, 411–426.

van Dieen, J.H. and Toussaint, H.M., 1997. Evaluation of the probability of spinal damage causedby sustained cyclic compression loading. Human Factors, 39, 469–480.

Waters, T.R., Putz-Anderson, V., and Garg, A., 1993. Revised NIOSH equation for the design andevaluation of manual lifting tasks. Ergonomics, 36, 749–776.

Ergonomics 1063