l : J û l l . \ l l . S t t \ t t t K t i l t l h t t R u u t t h ù J B h h l a t t l ) r t ) D | Q l ù l d L \ t l d t t : t ) t ) l
Using
On-Site
Observations
for Low Back Related
Work Load Assessment
- Methodological
Issues
p i c r r c I t S O N t V I t _ t _ t i . p h i l i p p c M ^ t R I A L T X ('\tit'crsitl.ttl.I'itgc'Octu1ltttionI|tu/|honl]|La/thI|lr'tIlkûUni|.
Liègt:. llcltimt
^bstrxct: on silc crlononric ohscrr'alions havc hccn sclcclcr! rs lhc colc nrcllrrrtl lo irsscss
* o r k l t r l d c x f t r r s l t r c * i l h t n l c o h o r t s l u d y () v c r c l i o l o g i c n n r l p r o g l l ( ) s t i c ( l ! ' t c r n r i n r n l s o l ' N ( ' r k f! l i l l c ( l l r l r c k I l r i r r ' S c l l r e l ) { ) r t s c ( ) l l c c l c ( i l l ) ( ) l r ! l l i r q r r c s l i o r r r r u r r c l r a y 6 l l i r l r e l r e a ' c r
prorv lirr erposure assessnrcnl but lhey ha'e bccn sho*n nol to providc rerinbrc (ratir 1,,
cnsrlre al corrccl classiljcl]tion ol'sul.rjects in clill'crcnl cxposrrrc sroLrps.
The tlcsisn ol'iln observation protocol raiscs several methocloiogical issrrcs. pirsl a choicc
has to bc n.tadc betwecn rlircct or dclaycd observations using I,ideo rccordings. In
conrparison to dirccr obscrv:rrions, clelaycri ones alrorv a nxrrc prccise n,"",rr,r",uiut ul'
d u r a l i o n i n d l i e q u c n c i c s u I e a c h r c t i v i t y b u t a r e ri n r e c o n s u r n i n g . c o s t r y r n r l r a c k t r r e l r )
r ision ol'thc lvorkcr poslLtrc. Whcn opling lôr clircct obscrvationi. a clisconti6uols captLrre
n l o d a l i t l s c c n l s t o o l l ' c r t h c b c s t c h ( ) i c c i l o n c n r a k c s u s e o l ' a n c x h a u s t i v c c o t l i n g g r i r i th a t
*ill cnsttre thc ohscrvations conlent veliclilt,. Â scconcl issue coDcerns thc tri(lc-ol1 to b!,
li)und b!'t\\ccn thc ohscrration accuracv (using thc shortest possibl., tinte intL,rvnl) an(l thc
obscrver ntental loild.
A thi(j issrrc is horv to hcsl sample thc rvork activilics: using a task-b1scd or.a ranclomizetl
rllnc sanrplin{ tcchniquc'} r\ccording lo rhc literature. rvi{hin-group lariancc secnrs ro bc
less \\ilh ll lired nunrber ol observation pcriods randomly distributcd over the rvork shill.
When this lrtst rpproach is selccted, a 1ùrther issue rel'crs to thc optimized bllance bctseen
the number ol'$orkers to santple in cach job category and the nuntber of observation
pcrrods pcr lorkcr. Ileccnt strrdics using a bootslrapping techniclue provicle prornising
. , ' l l t t i , r l l . I ' , t l r i s lu r t , 1 t r c . t r , , r r . l . I n t r o d u c t i o n T h e o n - s i t e o b s e n ' a t i o n n r e t h o d o l o g y i s p a r t ol'a workloacl c x p o s l l r c a s s e s s n t e n t s t r a t e g y d c v e ) o p c d l b r a l b l l o u , L t p e p i d e m i o l o g i c s t u d ) , f u n d e d b y t h c B e l g i a n g o v c r n m c n t ( C ) S . ] . C P S i l l i 2 6 ) . T h i s stucl1,, l c d b y s e v e r a l r e s e a r c h c e n t c r s l , a i m s to a s s e s s t h c r e s p e c t i v e r o l c o l ' p e r s o n a l p s y c h o l o g i c a l c h a r a c t e r i s t i c s , a n d w o r k - r c l a t e d p h y s i c a l a n d p s y c h o s r i c i a l l a c t o r s 6 n t h e o c c t l r r c n c e o l ' I o r v b a c k p a i n atronq 972 young ( < 3 I yr) and pain licc rvorkcrs l i 6 1 r c i t h c r h c a l t h c a r c o r t i i s t r i b u t i o n s c c t o r s .
I n d i v r d u a l u ' o r k l o a d i s f i r s t asscssed b y m e a n s o f a s e | l - - a c h r i n i s t r a t c c l q u e s t i o n n r i r e . w h c r e a s a q u e s t i o n n a i r c i s e a s y t o a p p l y , i n a r c p e a t e d m e a s u r e s d e s r g n , t o a l a r y e s l n r p l e o 1 ' subjects, it is rl'ell-knorvn that, in cornparison to thc observations, thc collectcd clata aie not r c l i a b l e e n o u g h t o c n s u r e a c o r r e c t c l a s s i l l c a t i o n o l ' s u b j c c t s i n c l i l l è r e n t c x p o s u r c g r o u p s ( 1 2 ) . A n o b s e r v a t i o n n l e t h o d o l o g v h a s t h L r s b c c n d c v e J o p e d b u t o n l y c o n c e r n s a s r m p l c o l ' t h e s t u d 1 , s t r b . j c c t s i n e a c h j o b catcgofy b e c r t t s e o l ' c o s t li n r i t a t i o n s . T h c d c s i g n 6 l ' s r r c I a n o [ s c l irtiorr p r o t o c o l r a i s e s s c v c r a l n r e t h o c l o l o g i c a l i s s u e s u , h i c h a r c d i s c u s s c c l h e r e . 2 . O h s e r v a t i o n m c t h o d o l o g y T h r c c r e c e n t r c v i o v s (2 ; 7 ; { l ) h a v c tl r s t p r r n c i p l c s l i k e tl r c c h o i c e b e t u ' c e n d i r c c t o r b e c n a n a l y z c d i n o r c l e r t o t l e l i n c th e obscrvatiurr d e l a y e d o b s e r v a t i o n s a n c l b e t r v c e n a c o n t l n u o u s o r } 9
P -R. Sotrrllc dùd I llh'iridtx / tJ!utg O -Sitë OLJsërt utioûs lor Lo* Ba.k Reldtal lotk Lodl Astcs\nkttt
discontinuous capture mode. Oncc thesc principles dellned, six obscrvation mcthods ( | ;4; l0; | | ; l3; I5) lravc bccn sclectcd a n d a n a l y z c d t o l l n d t h c b c s t tr a d c - o l ' l ' b c t w c c n t h c observation accuracy and thc obscrvcr mcntal load and to selcct accordingly thc suitable obscrvation period châraoteristics. A last issuc conccrns the sarlpling strategy of the obscruation pcriods :rvhiclr of citlicr a task-bascd approach or a randornizcd tirnc-slrlpling tcchnicluc is the most ùccurate in assessing wolk activities'?
2.I Obscrt ution pt inciples
It lirst appears in our rcvicw analysis that, whcn using a computerized grid thc choice o1'a s i n g l c o b s c r v c r c a n b c I l a d c t o a v o i d th c in t c r - o b s e r v c r v a r i a b i l i t y b i a s . I I o w e v e r , a c c o r d i n g to Kilbon.r guidelines (7), in that casc thc number ol'variablcs sirnultaneously observcd should be lorvcr than l0 and the trunk posture variables should be assesscd using a maximum of3 categories. Moreovcr, thc single observer rlust bc wcll-trained to its grid and an intra observer rcproducibility test should be made. (l) The grid as uscd in the present study is shorvn in Fig.
L In ordcr to ensure thc grid content validity, this grid takcs into account the low back pain r i s k l a c t o r s ( w l r o l e b o d y v i b r a t i o n , m a n u a l h a n d l i n g , t r u n k b e n d i n g a n d tw i s t i n g a n d s t a t i c work posturc) (3). Five differcnt variables are analyzcd (or l1 if' bcrsic motor actiott is considcrcd as a 7 variablcs group) and the trunk posturc variablcs arc classificd into 2 and 3 c a t c g o r i e s .
Table L Tlrc observlion gt id.1.t used h thc stut|\.'
I t a s i c p o s t u r c llcxio n rotation basic motor action l o a d (k e ) standing s i t t i n g k n c c l i n g / s q u a t t i n g
0-20.
2t-45"
>45"
0" Itot Rot + . no action/holding a )oad . walking /carrying . driving . lifting/lowering . puslting/pulling . throwing . other a c t i o n0
l - l0
l l - 2 s
> 2 5
Whcn clroosing bctwecn clircct or dclayed obscrvations, it appcars tltat, in contparison tq dircct observations, dclayed ones allow a more prccise lneasuremcnt ol' durations and l i c c l u e n c i e s o l ' c a c h a c t i v i t y b u t are time consunring, c o s t l y a n d la c k t h e 3D vision of the workcr posturc bccausc of thc vidco recording. Direct obscrvatrons could thus providc a bcttcr' c o s t - c f ' f c c t i v c n c s s r c l a t i o n s h i p . ( 7 )
I l ' d i r c c t o b s c r v a t i o n s l r c s c l c c t e d , a c h o i c c h a s th e n to b c r r a d e betu'ccn a c o n t i n u o u s o r d i s c t t n t i n u o u s c a p t u r c I n o d e . T h e c o r r t i n u o u s o n c p r o v i d e s b c t t e r a c c u r a c v o f d u r a t i o n a n d lreclucncy tt]cilsursnlcllts but incrsascs thc obscrvcr mcntal load and inrplics thus a rccluction in thc number of obscrvcd variablcs. So lar tlre advantage o1'a continuous observation ovcr d i s c o n t i n u o u s o n e s h a s n o t b e e n dcmonstrated ( 7 ) . So, a discontinuous c a p t u r e n r o d a l i t y ( t i l n c - s a n l p l i n g p r o c e d u r e l i k c s n a p s h o t ) s c c m s t o o 1 l è r t h c b e s t c h o i c c i l ' o n c w a n ( to u s c a n e x h a u s t i v e g r i d .
2.2 Oltservutiotr ltariocl tleJiniliott
when choosing the discontinuous capture way, a tradc-ot1- has to bc lbund betwcen thc observation accuracy (using the shortcst possible tin.re-sampling interval) and the obscr-ver uental load. In l'act, lowcring this intcrval allows the observer to see morc opcrations of each obscrved task but asks to limit the numbcr of observcd variablcs and to dccrease the obscrvatiou pcriod duration.
I R . 9 ' n r i l l . d r d P l \ l t u t r " t \ / L s i ù s O n - S i t . O h \ ù t d t i o k t l o t t . . \ , R d t k R . t i t a ! l l t k t t \ / t ! \ t \ s D r n l
I n t h e li t e r a t u r c , t h e ti r n e - s a r n p l i n g i n t c r v a l u s e d i n p r a c t i c e v a r i c s f r o m l - 5 s e c (l l ) t o I m i n u t e ( l ) . The PIlo methoclologv ( 4 ) u s e s a 3 0 m i n u t e s o b s e r v a t i o n p e r i o d a n c l t h c o w A S m e t h o d o l o g v g u i d e l i n e s ( 6 ) requests a m i n i m u m ol'100 capture p o i n t s lor each observation p e r i o d s . o n t h a t b a s i s , a 3 0 m i n u t e s o b s e r v a t i o n p e r i o d a n d a l 5 s e c ti m c - s a m p l i n g i n t e r v a l ( 120 c a p t u r c p o i n l s p c r p e r i o d ) w e r e c o n s i d c r e d a s b e i n g a g o o d c o n t p r o m i s c .
Moreover, thc computerization of the grid (using a Fujitsu .styli.stit l,T c-j00* sensitrvc screen conlpllter) provides various means to reduce the observer nrental loatl. For inslancc. in c a s e o 1 ' a l o n g d t r r a t i o n r v o r k t a s k , e a c h l5 s e c , t h e s o l i r v a r e u s e s t h c la s t e n c o d i n g a s a t l c l i t u l t value for the next encoding. Nloreover, the observcr is inlorrned of the montent hc^must look a t t h e r v o r k e r b y a b i p and the countdorvn o f t h e l 5 s e c o n d s i s : r l s o s h o w n . T h e r c is a r r , r
t t t ' r t i l t t h l t l i r n c l i o t t l o c l t t r c c l l l r c c n c o r l i n t lr r t l a /rrrrr.sc I i r n c l i o n lo i n l c r n r p t ll r e o b s c r v l l i o r r . w h e n n e e d e < l ( t . c . lo l t l t e r . v o r k c r p r i v a c y ) .
2.3 Sampling ctf ob.sen,atirsn pcriotls
T r v o a p p r o a c h e s a r c p o s s i b l c t o s a m p l c th e r v o r k a c t i v i t i c s o f a g i v c n j o b : u s i n g a task-b a s c d o r a r a n d o r n i z e d t i m e s a n r p l i n g t e c h n i q u e . T h c f i r s t s t r a t c g y i n v o l v c s th e o t r s c r v a i i o n o l ' e a c h o f ' t h e t â s k s 0 1 ' t h e . j o b t h e s a r r e rv a y ( 1 b r in s t a n c e , o n e o b s c r v a t i o n p c r i o c l p e r t a s k ) a n c l t h c n , th e r v e i g l r t i n g o l ' t h e c l a t a c o l l e c t e d t a k i n g in t o a c c o u n t t h c t i m e p r o p o r t i o n o { ' e a c h ta s k :io thal co(lt fiut'rion (n be contpuuble (14). A task-based approach sccms to bc logical antl a c c r t r â t c b t r t i n l l r c t i 1 r c q L r i r e s a J a r g e p r e l i r l i n r r y a n a l y s i s to d c l l n c t l t e t r s k s t c n r p o r l l d i s t r i b L r t i o n . I r o r j o b s in v o l v i r r g v a r y i n g ti t s k s , s u c h a n a l y s i s i s t i n r e c o n s r r r l i n g a n d r v i l l h a v e a direct impact on the actual accuracy ofthe approach.
The second strateg)i consists in a randomized distribution of a llxed numbcr observation p e r i o d s a l l o v e r th e s h i f t . rv i t h i n e a c h j o b or function, r v i t h o u t t a k i n g in t o a c c o u n t t h e t a s k s distribtrtion. According to the literaturc, rvithin-group variance sccl.rls to be lcss rvith a lircd numbcr o1'observation pcriods ranclomly distributed over the rvork shift (9) than when using a t a s k b y t a s k a p p r o a c h . I l e n c e t h e c h o i c e of using a randomizcd o b s e r v a t i o n p e r i o c i s d i s t r i b u t i o n r v a s m a d c i n t h c p r e s e n t s t u d y . Y e t i n s u c h a n a p p r o a c h , i t i s a b s o l u t c l y n e e d e d t o d e f l n e a f i r c d n u n r b c r o l ' o b s e r v a t i o n p e r i o d s a n d a l i x e d n u n r b c r o 1 ' o b s e r v e d s u b j e c t s p c r function.
Recent studies using a bootstrapping technique proviclc scveral options to dellne thesc numbcrs. IJoozemans and al.(5) have detcmrinecl lor a given number o1'observation periods the numbcr of observed subjcct needed to cnsure a ,50% precision level in a -5-95 percentile r a n g c . F o r i n s t i l n c c , t o r c a c h th i s p r c c i s i o n l e v c l , c i t h c r I s u b j c c t s ( i t t l c a s t ) havc to lrc o b s c r v e d d u r i n g 8 p c r i o d s o r a t I e a s t l 2 s u b j e c t s d u r i n g 4 p c r i o d s . T h i s la s t o p t i o n s h o u l d b c an efficient stratcgy lirr it allows obscrving 2 rvorkcrs a tlay so tltat crch job catcgory rs o b s e n c d r v i t h i n 6 d a y s i n s t e a d o f 8 .
3 . C o n c l u s i o n s
The described methodology has been applied to 152 cohort study subjects rvitlrout any problems fbr the obscrvcr and can thus be considered as feasiblc. Yet the methodology has s t i l l to b e v a l i d a t e d . I f t h c c o n t e n t v a l i d i t y o f t h e g r i d is a l r e a d y e n s u r e d , t h e c r i t e r i o n v a l i d i t y must still be tcsted for several âspccts ofthe grid. The cstinration of cluration and fieqrrencres w i t h a d i s c o n t i n u o u s c a p t u r e r n e t h o d o l o g y c a n b e t e s t e d b y c o n r p a r i n g t o s i n r u l t a n c o u s n c a s u r c m e n l s u s i n g a c o n t i n u o u s c a p t u r e m o d e . T h e w e i g h t ( o r exerted l o r c c ) a n d th e tn r n k angle cstimation could also be comparecl to objectives values respectively measured with a dynamometer and an inclinometer (or goniometer) strapped to the subject tnrnk. Horvcver, a second observer (simultaneously taking these measurements) would be needed lor these v a l i d a t i o n t c s t s . F i n a l l y , t h e re s u l t s c o l l e c t c d u s i n g t h e s e l e c t e d s a m p l i n g s t r a t e g y m u s t s h o w a h i g h c r v a r i a n c e b e t r v c c n j o [ r cutcgories t l r a n r v i t l r i n . j o t r c a t c g o r i c s .
1 ' - R S o D ^ i l l c d ù J t ' l r l u t t i u r r / U s t n e O n - S i t c O b s e r w t k ù ! l o t L o v l 3 o t A R e l u t e t l W o t k L o a d A t : i a s s | ' ! n l
4. Rcfcrences
l. Uuchholtz Ll, Paquct V, Punnctt L, Lcc D, Moir S. PATII: a work sampling-bascd approach to ergollomic
j ob analysis lbr constr uçtion and othcr non-rcpctitivc rvork. Appl.lJrgon, lt)\)6;21 (l): l'l'/ -8'/ .
2. Burdorl A. Exposurc asscssment ol risk làctors lbr disordcrs of thc back in occupa(ioritl cpidcmiology.
Scand.J.Work Environ.Health, 1992; l8 (l): l-9.
3. []urdorl'Â. Sorock (i. l>ositivc and ncgativc cvidcncc ol'risk Iàctors lbr back disordcrs. Scand.J.Work
lir! if orr.f lcirltlr, l()97 ; 27 (4): 243 -56.
4. l;ransst:n-llall C, Cloria R, Kilbom Â, Winkel J. A portablc ergononric obscrvation mcthod (Pl1'O) for
corùputcrired or-lir)c rccotdiug ol'posturc and ntanual lt.tndling. Âppl.titgon, 1995;2(r (2): 93- 100.
5. lloozcnrans MJ, Uurdorl'A, Van dcr Bcek AJ, l-'rings-l)resen Nlll, Mathiassen SE. Group-bascd
mcasurcnrcnt strâtcgics in cxposurc asscssnlcnt explored by bootstrrpping. Scand.J.Work Inv iron.l lcalth,
2001,27 (2): 125-32.
6. Karhu O, Kansi P, Kuorinka l. Corrccting working posturc in industry: a praclical nrcthod fbr analysis.
^ p p l . E r g o n , 1 9 7 7 ; E : 1 9 9 - 2 0 1 .
7. Kilbom A. Asscssrncnt ofphysical exposurc in relatiorl 1{) work-related musculoskelctal disordcrs--what
inlormation can bc obtained lrom systenlatic observations'? Scand.J.Work []nviron. ]lcalth, 199.1;20 Spcc
N o : 3 0 - 4 5 .
8. t-i C, Bucklc P. Currcnt tcchniqucs l-or asscssing physical exposure to work-relrtcd musculoskclctal risks,
wilh cnrphasis on posture-based nlcthods. llfgonorDics, 1999; 42 (5): 67.1-95.
9. Mathiassen SIr, Môller T, Forsnran M. Variability in mechanical exposure within and bctwccn individuais
pcrlbruring a strictly controlled industrial work task. Arn.J.lnd.Msd 2001.
10. I{idd J, Nicholson A, lMontan A. A portablc microcornputcr bascd systcnl lbr "on-sitc" activity and posture
rccording. Ir: l aylor and Iraucis, cdilor. Iirgonornics-dcsigliDg proccss. Mcgarv l-D cd. Londol: 1989. p.
366-9.
I l. van dcr lleck A, Van Caalen L, Irrings-Drcscn Mll. Wolkrng posturc rnd activities ol'lorry tlrivcrs: a
Icliability stLrdy oftln-site obscrvation and recording on a pockct cornputcr. Appl.lrrgon.,
I t ) 9 2 : 2 3 ( 5 ) : 3 3 1 - 6 .
12. Van dcl Bcck AJ, Frings-Drcscn Mll. Âsscssmcn( o1'mechanical cxposurc in crgononric cpidcntiology.
O c c u p a t i o n a l & L n v i r o n r n c n t a l M e d i c i n e 1 9 9 8 ; 5 5 ( 5 ) : 2 9 1 - 9 .
13. Wclls I{. Moore ^, Norman I{, Ncunrann I', Andre$s D. Dcvclopment and irDplcnrcntalior ol'SAM, u pcn
bascd conrpulcr systcnr lbr rvork saurpling. 1995.
14. Wclls I{, Nornran R, Neulnantr P, Andrcrvs D, Frank J, Shannon H ct al. Asscssurcnt ol physical rvork load
in cpiclcnriologic stutlics: cournron nrcasurcmcnt nrctrics ibr cxposurc asscssnrcnt. Drgononrics, 1997; 40
( l ) : 5 1 ' ( r l .
1 5 . W i k t o i i n C l , l v l o r t i n r 0 f M , l ] k c n v a l l l - , K i l b o n r  , l l l c l n r l i W . ll  l { l : } 0 , a s i r n p l c c o r n p u l c f - i r i d c d
obscrvalion nrcthod li)r rccordi|]g work posturcs. Scand..l.Work tnviron.llcaltlr. 1995; 2l (6): 440-9.
