measurement, · measurement,evaluation, andassessmentof occupational exposuresto hand-transmitted...

Occupational and Environmental Medicine 1997;54:73-89

METHODOLOGY Series editors: T C Aw, A Cockcroft, R McNamee

Measurement, evaluation, and assessment ofoccupational exposures to hand-transmittedvibration

M J Griffin

AbstractThe measurement of hand-transmittedvibration converts oscillatory movementsto a form in which they can be evaluatedwith respect to human responses andassessed for their acceptability. This paperpresents methods of measurement, evalu-ation, and assessment currently advocatedin standards and other forms of guidance.The degree to which the methods of evalu-ating different frequencies, directions, anddurations of vibration affect the assess-ment ofvibration on different tools is illus-trated. With the frequency weightingcurrently used to allow for the effects ofdifferent frequencies there is little need tomeasure vibration at frequencies as highas 1000 Hz; this has significant implica-tions to the design and evaluation of pro-posed antivibration devices, includinggloves. Without the current frequencyweighting, vibration at frequencies greaterthan 250 Hz can contribute to the magni-tude of the vibration, but many commoncauses of injury from hand-transmittedvibration have their dominant componentsof vibration below 250 Hz. On many pow-ered tools, although the dominant fre-quency of vibration is the same before andafter frequency weighting, the reportedmagnitude of vibration is greatly affectedby the frequency weighting. On tools withdominant low frequencies, their vibrationis rated as being of far greater importancerelative to other tools when consideringfrequency-weighted acceleration thanwhen considering unweighted accelera-tion. It is shown that the effect ofconsider-ing three axes of vibration as opposed toone axis has a greater effect on some toolsthan on others. The uncertainties andassumptions involved in the measurement,evaluation, and assessment of hand-trans-mitted vibration are reviewed. It is sug-gested that whereas current decisions onhealth and welfare should be based on cur-rent assessment methods, the measure-ment and evaluation of hand-transmittedvibration should involve the collection andreporting of data which allow other inter-pretations in the future.

(Occup Environ Med 1997;54:73-89)

Keywords: vibration; dose-effect relations; vibration-induced white finger; standards

Throughout history, mechanical impacts asso-ciated with accidents and aggression havecaused acute injuries to the human body.During the past century, it has been observedthat repeated impacts (vibration and repeatedshock) with sufficiently low magnitudes so thatthey do not individually cause detectableinjury, can give rise to signs and symptoms ofchronic disorders. The cause-effect relationbetween impacts and acute injury is self evi-dent, although not always easily predicted.The cause-effect relation between vibrationand chronic disorders has been more opaque:the elapse of time between cause and effectand the complexity of both the cause and theeffect have helped to conceal their relation.

In the United Kingdom, one of the firstpublished reports of vibration-induced injuriesappeared in the 1907 report of theDepartmental Committee on Compensationfor Industrial Diseases.' Paragraph 33 of thereport states:

"33. Neurosis due to Vibration-Our atten-tion was called to neurosis due to vibration,caused by the use of pneumatic tools.Tremor and sleeplessness have occasionallybeen observed in individuals, but no evi-dence was obtained of the existence of anynervous disease from this cause which inca-pacitates from employment."

The committee did not recommend that vibra-tion-induced disorders should be added to theschedule of diseases in the Workmen'sCompensation Act. Compensation requiredthat a disease or injury caused incapacitationfrom work for a period of more than one week,and that the disease or injury was so specific tothe employment that the causation fromemployment could be established in individualcases. Deafness, acknowledged to be widelyprevalent among boilermakers in the shipyards,was not accepted for compensation as it didnot prevent a man from continuing at his trade.Early published reports of people affected byhand-transmitted vibration were also publishedin Italy and the United States.2

Compensation for vibration-inducedinjuries to the hand was reconsidered severaltimes in the United Kingdom over the next 80years, but did not come into force until 1 April1985 when compensation was restricted to"episodic blanching, occurring throughout theyear, affecting the middle or proximal pha-langes ... on one hand ... any three of thosefingers" among people in a specified set of

Human FactorsResearch Unit,Institute ofSound andVibration Research,University ofSouthamptonM J GriffinCorrespondence to:Professor M J Griffin,Human Factors ResearchUnit, Institute of Sound andVibration Research,University of Southampton,Southampton SO 17 1 BJ.Accepted for publication1 August 1996

73

on April 21, 2020 by guest. P

rotected by copyright.http://oem

.bmj.com

/O

ccup Environ M

ed: first published as 10.1136/oem.54.2.73 on 1 F

ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

Table 1 Five types of disorder associated with hand-transmitted vibration exposures8

Type Disorder

A CirculatoryB Bone and jointC NeurologicalD MuscleE Other general (for example, central nervous system)

Some combinations of these disorders are sometimes referred toas hand-arm vibration syndrome (HAVS).

occupations.6 More recently, recognition thathand-transmitted vibration from various toolsmay cause disorders as well as finger blanching(table 1) led to a recommendation that thescope of the compensation should be extendedto include neurological effects.7 In other coun-

tries, there are various patterns of compensa-

tion for vascular, neurological, and articulardisorders thought to be associated with hand-transmitted vibration.What type of vibration causes injury? The

provision of an answer is impeded by the com-

plexity of the problem and the need to use jar-gon associated with science, engineering, andmedicine. Knowing that vibration causes

injury leads to attempts to represent the vibra-tion by means of numbers (measurement).Vibration can be measured without knowledgeof the effects it produces. To discern the rela-tive or absolute severity of the vibration (evalu-ation), or to identify the likely consequences ofexposure to the vibration (assessment), it is nec-essary to have knowledge of the cause-effectrelation between vibration and injury.

MEASUREMENT, EVALUATION, AND ASSESSMENTThe measurement of vibration requires that theoscillatory movements are transduced (theenergy is converted to a form representing themovements). The method of measurement haschanged over the years according to the avail-ability of transducers having appropriate sensi-tivity, size, cost, etc. In the past 50 years themost commonly used types of transducershave been accelerometers which, by one

means or another, convert the acceleration ofthe surface to which they are attached into an

electrical signal. No transducer has the capa-bility to fully represent all possible motions:the measured signal is a sufficient representa-tion of the vibration when it follows the rele-vant parts of the motion (the frequencies,magnitudes, and directions of interest) withappropriate accuracy. This involves assump-tions as to which parts of the motion cause thedisorders of interest. Measurements can bestored for later consideration as tables ofnum-bers, as waveforms drawn on paper, as ana-

logue recordings on magnetic tape, or in a

digitized form for use by computers.The evaluation of measurements of human

response requires the use of procedures whichshow the relative or absolute severity of thevibration: it may not be appropriate to assume

that all frequencies or all directions of motionare of equal importance. An evaluation proce-dure will result in one (or a few) numbers so

that the severities of different vibration expo-sures can be compared. This requires knowl-

edge of the relative importance of differentqualities in the measurement (frequencies,directions, durations). The evaluation may beexpressed by values on either an interval or aratio scale (a scale on which differencesbetween intervals on the scale have somequantitative significance). Similar to the fre-quency weightings used in acoustics, it hasbecome common practice to "weight" vibra-tion with a weighting reflecting the assumedeffects of different vibration frequencies, direc-tions, and durations. It is then possible toreport a single weighted value which isassumed to be representative of the severity ofthe complex motion which was measured.The assessment involves a consideration of

the vibration and a judgement about it.Whereas the evaluation results in a numericalvalue representative of the vibration severity,the assessment predicts the outcome of avibration exposure: the type, severity, or prob-ability of disorder, or even the legal conse-quences. An assessment does not necessarilyrequire measurement and evaluation of vibra-tion: a particular type or source of vibrationexposure could be labelled as unacceptableand banned without knowledge of the vibra-tion magnitudes.The boundaries between measurement,

evaluation, and assessment are easily and, toooften, blurred. The ability to make an assess-ment without measurement and evaluationallows the possibility of making this judgementand merely resorting to physical values to sup-port the conclusion. This happens with indi-vidual assessments and also in the process ofstandardisation, when the measurement andevaluation methods may be selected to reachthe desired conclusion rather than being justi-fied in their own right. The separate identifica-tion of a measurement method, an evaluationprocedure, and an assessment criterion mayencourage a more rigorous route to both indi-vidual assessments and the production of stan-dards containing guidance.The objective of this paper is to present the

current method of measuring the severity ofexposures to hand-transmitted vibration andshow the consequences of this method. Thisrequires an initial consideration of the physicalvariables influencing vibration severity, beforesummarising the method advocated in currentnational and international standards. The con-sequences of assumptions implicit in the stan-dards are shown and the implications for thefuture course of standards are considered.

Physical variables and their measurementTable 2 shows some of the principal physical

Table 2 Physical variables relevant to the effects ofhand-transmitted vibration

Magnitude of vibrationFrequency of vibrationDirection of vibrationDuration of vibrationArea of contact with vibrationContact force (grip force and push force)Finger, hand, and arm postureEnvironment (for example, temperature)

74



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Measurement, evaluation, and assessment of occupational exposures to hand-transmitted vibration

variables influencing the severity of hand-transmitted vibration.

VIBRATION MAGNITUDEA vibrating object moves to and fro over some

displacement with alternately a velocity in one

direction and then a velocity in the oppositedirection. The change of velocity means thatthe object is constantly accelerating, firstly inone direction and then in the opposite direc-tion. The magnitude of a vibration can bemeasured by either its displacement, its veloc-ity, or its acceleration. For practical conve-

nience, the magnitude of vibration is now

usually expressed in terms of the accelerationand is measured with accelerometers. Theunits of acceleration are ms-2. Magnitudes ofhand-transmitted vibration are currentlyexpressed as an average measure of the accel-eration of the oscillatory motion: the root-mean-square value (ms-2 rms).

VIBRATION FREQUENCYVibration frequency is expressed in cycles per

second using the SI unit, hertz (Hz). The fre-quency of vibration influences both the man-

ner and the extent to which vibration istransmitted through hand-held machinery, theextent to which it is transmitted through thefingers, hand, and arm, and the responses tothe vibration within the body. The relationbetween the displacement and the accelerationof an oscillatory motion is also dependent on

the frequency of oscillation: a peak to peak dis-placement of only 1 mm corresponds to a lowacceleration at low frequencies (for example,about 0-9 ms-2 rms at 8 Hz) but increasingacceleration at increasing frequencies-forexample, 140 ms-2 rms at 100 Hz, and 14 000ms-2 rms at 1000 Hz. The vibration displace-ment, which can be seen with the eye, does

T Xh

Yh t

Figure 1 Axes of vibration used to measure exposures to hand-transmitted vibration.

not give an indication of the acceleration mag-nitude.

VIBRATION DIRECTIONThe transmission of vibration into the handdiffers according to the direction of theapplied vibration. The effects of vibrationwithin the hand and arm may also depend onthe axis of vibration. Vibration is measured onthe handle of a tool close to the hand in threeorthogonal directions designated x, y, and z(fig 1). The axes may be defined relative to theorientation of the hand or relative to the tool.The position and orientation of a hand on atool may vary. In consequence it is often moreconvenient to quote the vibration magnitudesrelative to three convenient axes of the toolrather than the axes of the hand.

VIBRATION DURATIONSome human responses to vibration dependon the duration of exposure. Also, the dura-tion of measurement may affect the measuredmagnitude of the vibration. The measurementof vibration with rms averaging assumes a timedependency in which doubling the magnitudeof a transient acceleration is equivalent to afourfold increase in the duration of the event.The rms acceleration may not provide a goodindication of vibration severity if the vibration isintermittent, contains shocks, or otherwisevaries in magnitude from time to time.8

OTHER VARIABLESAlthough other variables (the area of contactwith vibration, the grip force, and posture)probably influence the severity of vibrationexposures, there are currently no standardizedmethods of measuring them.

MEASUREMENT METHODSOne or more accelerometers with the capabilityto both withstand and indicate the vibration towhich they are exposed are secured to a sur-face adjacent to the hand.9 10 The mountingmethod should be rigid at frequencies up toand above 1000 Hz. On tools that causeshocks, mechanical filters between accelerome-ters and the tool may be used to attenuate thevery high frequency acceleration which causessome transducers to give erroneous readings atlow frequencies (dc shifts8). The total mass ofaccelerometer and mount must be sufficientlylow so as not to influence the vibration on thetool. Few reported measurements on toolshave identified the adequacy of the mountingmethod or shown that the mass of the mounthas no effect on measurements. Future stan-dards may identify the requirements withmore precision and give guidance on how theycan be achieved.The evaluation of vibration may be

achieved by analogue or digital methods, bydedicated instruments or by general purposeequipment. Dedicated instruments which onlyprovide an indication of the frequency-weighted acceleration may seem to be conve-nient but have drawbacks: the absence ofspectral information prevents the observationof spectra as a "signature" of the vibration giv-

75

...........



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

Examples of tool type (standards identifying type test)

Riveting tools (ISO 8662-2)Caulking tools (ISO 8662-5)Chipping hammers (ISO 8662-2)Clinching and flanging tools (ISO 8662-10)Impact wrenches (ISO 8662-7)Impact screw drivers (ISO 8662-7)Nut runners (ISO 8662-7)Scaling hammers (ISO 8662-10)Needle guns (ISO 8662-10)Nibbling machines and shears ((ISO 8662-10)SwagingPedestal grindershand-held grinders (ISO 8662-4; ISO 8662-8)hand-held sanders ((ISO 8662-4; ISO 8662-8)hand-held polishers (ISO 8662-4)Flex driven grinders or polishers ((ISO 8662-17)Rotary burring tools (ISO 8662-17)Files (ISO 8662-12)Hammer drills (ISO 8662-3)Rock drills (ISO 8662-3)Tampers and rammers (ISO 8662-9)Road breakers (ISO 8662-5)Stone working tools (ISO 8662-14)Chain saws (ISO 7505)Antivibration chain saws (ISO 7505)Brush sawsMowers (ISO 5395)Hedge cutters and trimmersBarking machinesStump grindersNailing gun (ISO 8662-11)Stapling gun (ISO 8662-11)Pad saws (ISO 8662-12)Circular saws (ISO 8662-12)ScabblersEngraving pensShoe pounding up machinesVibratory rollersConcrete vibrothickenersConcrete levelling vibrotablesMotorcycle handlebarsPedestrian controlled machines

ing a check on the validity of the measure-

ments-for example, the absence of dc shifts.Further, without spectral information, thedata are restricted to the current frequencyweighting at a time when its appropriateness isincreasingly challenged. Modern digital signalprocessing techniques allow economic calcula-tion and storage of data which has been digi-tised-this is becoming the method of choicein many situations. With these instruments,alternative evaluations and assessments can bemade with no significant increase in time or

cost.

Evaluation according to currentstandardsIt might be beneficial if standards could beclearly identified as providing guidance on

either measurement, evaluation, or assess-

ment, but current standards have not yetmatured to the appropriate roles. Con-sequently, they tend to have implicationsbeyond their expected scope. For those apply-ing current standards they may be conve-

niently grouped as either being concerned withthe evaluation of the place of work(International Standard (ISO) 5349, 19869),or being concerned with the evaluation of a

tool (parts of ISO 8662, 1986"1).

MEASUREMENT STANDARDS

Repeatability is a prerequisite for useful mea-

surements of hand-transmitted vibration.Practical use of tools in work is highly variableand therefore not easily repeatable. This leadsto two alternative measurement objectives:

testing the tool under defined conditions (typetesting) or measuring the vibration exposure ofa person (or group of people) performing a

particular job.Type testing standards have been evolved for

many tools in recent years (table 3). The objec-tive is to define a procedure in which a tool can

be tested in different laboratories to obtain sim-ilar results. Values derived from such tests can

therefore be "declared" as an indication of thevibration on such a tool. The first requirement isrepeatability: this is achieved by artificially elim-inating some of the sources of variability thatoccur in jobs where the tools are used. The val-ues obtained from these tests should not, there-fore, be assumed to indicate with any precisionthe vibration magnitudes to be expected atwork. The measurements on different tools ofthe same type performing the same test mightindicate the relative magnitudes to be expectedfrom these tools at workplaces. However, withsome tools, a low magnitude of vibration can beachieved by reducing the performance of thetool. The tool with the lowest magnitude ofvibration may not be the most suitable tool andthe magnitude measured in the test cannot beassumed to be the same as that in real work.These measurement standards have been stimu-lated by the requirements of the MachinerySafety Directive of the European Union (dis-cussed later).The measurement of exposures to hand-

transmitted vibration at the workplace oftenrequires consideration of a complex exposure

pattern. The act of measurement should resultin a record of a worker's exposure which is suf-ficient to perform an evaluation or assessmentwith a known or estimated accuracy. Thisrequires measurements which allow evalua-tions representative of the full daily and life-time exposure; it requires consideration ofsampling errors. Sampling procedures foridentifying which portions of exposure pat-terns should be measured are not well devel-oped, so allowing a source of error insubsequent evaluations and assessments.The ISO 8041, 199012 defines some charac-

teristics of instrumentation for measuringhuman exposure to vibration according toother evaluation standards.The current version of this standard is pri-

marily orientated towards the specification of a

commercial meter for the measurement ofvibration. British Standard 7482 parts 1 and 2,19911314 is similar to the intended scope ofISO 8041, but is not concerned with a specificmeter. It offers a more complete means ofspecifying and testing any instrumentationused to measure exposures to hand-transmit-ted vibration.

EVALUATION STANDARDSThe ISO 5349, 19869, the European StandardENV 25349, 1992'5 and many nationalstandards-for example, American NationalStandard S3 34, 198616-use a frequencyweighting (called Wh in British Standard684210) to assess the severity of hand-transmit-ted vibration over the approximate frequencyrange 8 to 1000 Hz (fig 2). The same fre-

Table 3 Some tools and processes potentially associated with vibration injuries

Type of tool

Percussive metal-working tools

Grinders and other rotary tools

Percussive hammers and drillsused in mining, demolition, roadconstruction, and stone working

Forest and garden machinery

Other processes and tools

76



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/


Figure 2 Frequencyweighting (Wh), (shownhere as an asymptoticweighting between 8 and1000 Hz; when evaluatingone-third octave bandspectra the frequency rangeis specified as 6-3 to1250 Hz.

I , ,,1111111 I,, II

10 100Frequency (Hz

quency weighting, Wh, is applied to themeasurement of vibration acceleration in eachof the three axes of vibration at the point ofentry of vibration to the hand.

These standards use the concept called" equal energy" so that a complex exposure

pattern of any period during the day can berepresented by the equivalent value for anexposure of eight hours (four hours in ISO

JJJ~lWLJXL~d0l ' 5349, 19869). For an exposure of duration t toa frequency-weighted rms acceleration ahWthe eight hour energy-equivalent accelerationahw(eq,8 h) is given by:

ahw(eq,8 h) = ahW I(t/T(8))

Figure 3 Years ofexposure tofour hourenergy-equivalentfrequency-weighted hand-transmitted vibrationrequiredfor the productionoffinger blanching in10%, 20%, 30%, 40%,and 50% of exposed peopleaccording to annex A ofISO 5349, 1986.9

01.' 25

c

0 10

ona)-0

aD 5Ean 20

0 1

xw 4

Figure 4 Accelerationmagnitudes predicted togive 10% prevalence ofvibration-induced whitefinger after eight years fordaily exposure durationsfrom one minute to eighthours (based on ISO5349, 1986).9

Cl

E

CNcnE

C

0

0

4 h energy equivalent acceleration (ms-2 rms)

10000 r,

1000 h

100

Exposure(min)* 1

+ 5* 10o 30

x 60* 120A 240

480

10 [-

10 100

Frequency (Hz)

Table 4 Number ofyears before blanching develops in 10% to 50% of vibration exposedpeople according to ISO 5349, 1986'

Weightedacceleration Percentage ofpopulation affected by finger blanchingahw,(,q,4 h)(Ms- 2nrms) 10% 20% 30% 40% 50%

2 15 23 > 25 > 25 > 255 6 9 11 12 14

10 3 4 5 6 720 1 2 2 3 350 < 1 < 1 <1 1 1

where T(,) is eight hours (in the same units as t).The value of ahw(eq,8 h) is sometimes denoted byA(8).

ASSESSMENT

All current assessments are based on theexpected occurrence of finger blanching(vibration-induced white finger). However,many assume that the prevention of vibration-induced white finger will be sufficient to pre-vent the development of other disorders. Asexplained later, the evaluation method(including the frequency weighting, Wh) is notbased on the development of vibration-induced white finger. The weighting (andother aspects of the evaluation method) mightbe justified as a general representation of thedependence of the hand on vibration fre-quency (and other variables): neither theweighting nor other aspects of the evaluationmethod are proved to represent the extent towhich different frequencies, axes, etc damagethe peripheral vascular system and its controlmechanisms.

Assessment by ISO 5349A relation between years of vibration exposureE, the four hour energy equivalent frequency-weighted acceleration ahw(eq, 40), and the pre-dicted prevalence of finger blanching C was

proposed in an annex to ISO 5349, 19869(table 4, fig 3):

2

C = 100. 1ahw(eq,4h) E] (2)

The values in table 4 and fig 3 refer to fre-quency-weighted acceleration (referenced tothe frequency range 8 to 16 Hz after passingthrough the filter shown in fig 2). BritishStandard 6842, 198710 offers similar guidance,but is restricted to a 10% prevalence of vibra-tion-induced white finger. Figure 4 shows howthe magnitudes required for a predicted preva-lence of 10% vibration-induced white fingerafter eight years are assumed to depend onvibration frequency from 8 to 1000 Hz forexposure durations from one minute to eighthours a day.The frequency weighting, the time depen-

dency, and the dose-effect information used inISO 5349, 19869 are based on less than com-

plete information, and such information as

exists has been interpolated, extrapolated, andsimplified for practical convenience." Con-sequently, the percentage of affected people inany exposed group will not always closely

01

c16 0-01

0*001

0-0001

(1)

77

1 '1



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

match the values in table 4 or figs 3 and 4.Also, the number of people affected by vibra-tion will depend on the rate at which peopleenter and leave a group exposed to vibration.

Machinery Safety DirectiveThe Machinery Safety Directive of theEuropean Community (89/392/EEC) states:"machinery must be so designed and con-structed that risks resulting from vibrationsproduced by the machinery are reduced to thelowest level, taking account of technicalprogress and the availability of means ofreducing vibration, in particular at source".'7Instruction handbooks for hand-held andhand-guided machinery should specify theequivalent acceleration to which the hands orarms are subjected when this exceeds a statedvalue (currently a frequency-weighted accelera-tion of 2-5 ms 2 rms). The relevance of anysuch value will depend on the test conditionsto be specified in type testing standards(already mentioned and table 3). In use, manyhand-held vibrating tools can exceed thisvalue. The specified tests do not encompass allrealistic exposures and so will not representthe exposures received in some jobs. Also, thedifferent exposure durations and other vari-ables involved in different jobs carry differentdegrees of risk.

Proposed European Union (EU) ExposureDirectiveThe opening principles of a proposed CouncilDirective on the minimum health and safetyrequirements for hand-transmitted vibrationare: "Taking account of technical progress andof the availability of measures to control thephysical agent at source, the risks arising fromexposure to the physical agent must bereduced to the lowest achievable level, withthe aim of reducing exposure to below thethreshold level . . . .18

For hand-transmitted vibration, the pro-posed EU Directive currently identifies athreshold level (ahw(eq,8 h) = 1-0 ms-2 rms), anaction level (ahw(eq,8 h) = 2-5 ms-2 rms), and anexposure limit value (ahw(eq,8 h) = 5 0 ms-2 rms.Figure 5 shows the manner in which theseeight-hour equivalent magnitudes correspond

C,,

E

cs'cn 100

04-

X 10a)a)(Dc.c;

s ~~~~~Exposure limit value\ ~~~~Action level>s>\ --~-Threshold level

Short term: 20 ms r2mrrShort term: 10 ms-I rms >

I~ilIIIIHilIIIIHl IIII III

1 10 100 1000 10000 100000

Duration of exposure (s)

Figure 5 Threshold level, action level, and exposure limitproposed in a draft EC Directive on Physical Agents. Aswell as these three criteria, effort shall be made to reduce thehazard when a short term (a few minutes) equivalentacceleration exceeds 10 ms-2 rms. Exposure to a short term(a few minutes) equivalent acceleration equal to or greaterthan 20 ms 2 rms is considered a hazardous activity.

to higher magnitudes at shorter durations.When exposures exceed the threshold level itis proposed that workers must receive infor-mation about the potential risk of exposure tohand-transmitted vibration. The action level isintended to identify the conditions in whichtraining in precautionary measures is required,an assessment of the vibration is to be made,and a programme of preventative measures isto be instituted. The proposed Directive alsoindicates that when the action level isexceeded workers shall have the right to regularhealth surveillance, including routine exami-nations designed for the early detection of dis-orders caused by hand-transmitted vibration.If the exposure limit value is exceeded, healthsurveillance must be carried out and memberstates of the community will be expected tocontrol the harmful effects.The extremely high magnitudes for short

durations (fig 5) should not be considered tobe representative of their safety: these highmagnitudes are a consequence of the equalenergy concept already described. The pro-posed Directive identifies exposures "for ashort-term (a few minutes) equivalent acceler-ation equal to or greater than 20 ms-2" as"activities with increased risk" which must bedeclared to the authority responsible: memberstates would be required to ensure that appro-priate measures were taken to control the risksassociated with these activities. Equipmentwhich transmits to the hand-arm system ashort term (a few minutes) equivalent accelera-tion equal to or greater than 20 ms-2 rms mustbe marked. For lower magnitudes of vibration,the proposed Directive says: "Where the activ-ity involves the use of work equipment whichtransmits to the hand-arm system a short-term(a few minutes) equivalent accelerationexceeding 10 ms 2, increased efforts shall bemade to reduce the hazard, with priority to theuse of low-vibration equipment and processes,including the revision of product design andwork practice. Pending the effective imple-mentation the duration of continuous expo-sure shall be reduced." The ambiguousreferences to short term and equivalent accel-eration leave these as unsatisfactory means ofassessing vibration. Vibration exposures arenot always statistically stationary: work oftenincludes transient exposures to vibration (andrepeated shocks) and variable intermittentperiods without exposure to vibration.The proposed Directive is based on the vec-

tor sum (more correctly the root-sums-of-squares (rss)) of the frequency-weightedaccelerations in three axes, although an axiscan be omitted if the values are less than 50%of the values in another axis at the same loca-tion. This differs from ISO 5349, 19869 whichis based on the axis having the highestweighted acceleration; however, a forthcomingrevision of ISO 5349 is expected to change theevaluation method to the use of rss.

Considerations for the futureA few variables dominate the current measure-ment, evaluation, and assessment of occupa-

78

1 000



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Measurement, evaluation, and assessment of occupational exposures to hand-transmitted vibration 79

100D

..10

:.0 .,.,X --

(A10 100 10 100 1010010100

10

10

0.1

10

I IV/I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

010

10 100 10 100 10 100 10 100

Frequency (Hz)Figure 6 One-third octave band spectra on 20 powered hand tools: unweighted spectra; ------ weighted spectra; (A pneumatic rock drill; Bpneumatic road breaker; C petrol driven hacker compressing road surface after mending; D non-antivibration chain saw; E antivibration chain saw; Fpneumatic metal chipping hammer; G pole scabbler; H needle gun; I random orbital sander; I impact wrench; K riveting gun; I dolly used with rivetinggun; M nutrunner; N metal drill; 0 wire waging; P etching pen; Q electric 9 inch angle grinder; R pneumatic rotary file; S pneumatic 5 inch straightgrinder; Tpneumatic 7 inch vertical grinder. Al spectrafrom the axis giving the highest weighted acceleration.)

tional exposures to hand-transmitted vibra- not widely recognised. They are illustratedtion, yet the manner in which these variables here with examples of vibration measurementsinfluence the currently standardized method is from an assortment of different tools. There



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

Figure 7 Correlationbetween weighted andunweighted rmsacceleration for the 20spectra shown in fig 6.

EI

0a)

0-)o

V

a)

0.0)

C

200 r.

150 t

100

50

U

*~ ~ ~ y = x* U (unweighted = weighted)

U

m m_ * U* U

I

-IF J~~~~~I_

0 10 20 30Weighted acceleration (ms-2 rms)

are doubts in several areas, including the valid-ity of the frequency weighting (Wa), the appro-priateness of the frequency range (8 to 1000Hz), and whether it is sufficient to measure inonly the axis with the highest frequency-weighted acceleration.

FREQUENCY WEIGHTINGFigure 6 shows one-third octave band spectrafor 20 tools sometimes associated with vibra-tion-induced white finger. The tools are onlypartially representative of other types perform-ing a similar function: tools of a specific typetend to have broadly similar spectral shapesalthough they may have very different vibra-tion magnitudes. Figure 6 shows the weightedspectra formed by applying the ISO 5349 fre-quency weighting (Wh). The weighting greatly

200

150

-a

o

~0

a)

0.C

reduces the importance of intermediate andhigh frequency vibration. However, in manycases the dominant frequencies in the spectraare the same before and after weighting: thefrequency weighting changes the magnitude ofvibration but often does not greatly alter thefrequencies contributing most to the rms mag-nitude.

Figure 7 shows the correlation betweenweighted and unweighted acceleration magni-tudes for the 20 tools with spectra illustratedin fig 6. Within tool types, where the magni-tude changes between tools but the spectra aresimilar, a tool with a high unweighted accelera-tion may be expected to also have a highweighted acceleration. However, across tooltypes, where the magnitudes and spectra dif-fer, fig 7 shows a poor correlation betweenunweighted and weighted acceleration. Thetools which differ most from the trend arethose dominated by low frequencies, so thatthey have similar unweighted and weightedmagnitudes-for example, the wacker-andtools dominated by high frequencies so thatthe unweighted magnitude is very muchgreater than the weighted magnitude-forexample, swaging. The frequency weightinghas therefore changed the rank order of impor-tance of several tools: swaging has the highestunweighted acceleration but the fourth highestweighted acceleration, the wacker has the 16thhighest unweighted acceleration but the high-est weighted acceleration.

In so far as tools of a specific type have

100010 100Frequency (Hz)

Figure 8 Integrated unweighted spectra for the 20 tools in fig 6: the accumulation of rms acceleration with increase in frequency from 8 to 1000 Hz(letter denotes the identity of tools as in the caption to fig 6).

n ---n L

80



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/


ECN

0

20C.)

10

0-10 100 10

Frequency (Hz)Figure 9 Integrated weighted spectra for the 20 tools in fig 6: the accumulation offrequency-weighted rms acceleration with increase in frequency from 8to 1000 Hz (letter denotes the identity of tools as in the caption tofig 6).

broadly similar spectra, the tool of a specifictype with the lowest vibration magnitude(weighted or unweighted) will often be theleast likely to cause injury, assuming the othervariables (duration of exposure, grip, etc) areunchanged. Thus, for manufacturers of spe-

cific tools, and those comparing similar tools,a frequency weighting accurately representingthe relative risks from different frequenciesmay not seem necessary. When choosingbetween tools with different spectra this con-

clusion cannot be supported: two tools withdifferent spectra but the same weighted accel-eration will not have the same unweightedmagnitudes and may not have the same poten-tial for causing injury. On this basis, uncer-

tainties about the frequency weighting can bebypassed for some applications if the assess-

ment (for example, predicted prevalence ofvibration-induced white finger, or the accept-ability of vibration) is made specific to a tooltype.'9 This approach must be closely confinedto tools with similar dominant frequencies as

the development of an antivibration mecha-nism for a tool could change its spectral char-acteristics.

Figure 6 shows that, after frequency weight-ing, the peaks in the spectra of all tools fallwell within the range 16 to 250 Hz. The

weighted accelerations are given by the areas

under the spectra when they are plotted on

appropriate linear scales. The use of logarith-mic scales in fig 6 disguises the fact that all ofthe weighted spectra are dominated by vibra-

tion in this narrow range of frequencies. Thefrequency weighting has effectively limited therange of frequencies contributing to evalua-tions and assessments to little more than 10:1,even though current standards require mea-surement of vibration over not less than a

125:1 range, from at least 8 to 1000 Hz.Figures 8 and 9 show, respectively, the

cumulative spectra for unweighted andweighted acceleration for all of the tools shownin fig 6. In the case of the unweighted spectra(fig 8) the acceleration on some tools (forexample, the road breaker, chipping hammer,impact wrench, riveter and dolly, etching pen)increase at frequencies above about 250 Hz.In the case of the weighted spectra, none ofthe tools show increasing values above 250Hz: almost identical values of weighted accel-eration would have been obtained by restrict-ing consideration to frequencies below 250Hz.At frequencies below 16 Hz, the frequency

weighting is such that the rise in the cumula-tive spectra with increasing frequency is thesame for weighted and unweighted spectra.Figures 8 and 9 show that there are large dif-ferences between tools in vibration magni-tudes at frequencies below 16 Hz.

FREQUENCY RANGEFigure 9 seems to show that there is little meritin extending the frequency range above about250 Hz when the frequency weighting is used:the weighting reduces the importance of these

)00

81



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

Liiiy=X7,,'

I-'

A

I

, Uw

- , -. U

,,,,I,,50 100 150

Unweighted(8to 1000 Hz) acceleration (ms-2 rms)

-u

,5

--U-,-4

Kb

10 20 30

Weighted(8 to 1000 Hz) acceleration (ms-2 rms)

Ec

CoI, E

CN

EC0e)

0

0

-rCo

0

0

LO

C)

a)

(1)C2

I3.1200LD

Co

, Co

EC0Co0)0)00

I000LO

0

~0-ca)

.CCD¢

40 C

) D

200 rb

Air150 F

100

50

0

200 Fd

A/Is

,-

,-d ,

A.1I

- X1-I"

I ,-

U

50 100 150

Unweighted(8 to 1000 Hz) acceleration (ms-2 rms)

.

150 H

100

50

.

.

U

,0 _

0

_

*. .-

III

0 10 20 30

Weighted(20 to 500 Hz) acceleration (ms-2 rms)

Figure 10 Correlations between various bandwidths of weighted and unweighted rms acceleration for the tools in fig 6: (a) effect of reducing upperfrequency to 500 Hz on unweighted rms acceleration, (b) effect of reducing upperfrequency to 500 Hz and increasing lowerfrequency to 20 Hz onunweighted rms acceleration, (c) effect of reducing upperfrequency to 500 Hz and increasing lowerfrequency to 20 Hz on weighted rms acceleration, and(d) correlation between weighted and unweighted acceleration in the bandwidth 20 to 500 Hz.

higher frequencies so much that they con-tribute little to the resulting value.

Without the frequency weighting there areseveral tools, the magnitude of which increasesas the frequency range is extended above 500Hz. Some of these-for example, the chippinghammer and the etching pen-have magni-tudes which seem to continue to increasethrough 1000 Hz: higher magnitudes ofunweighted acceleration might be obtained onthese tools if the frequency range wasextended to even higher frequencies. Even so,Figure 1OA shows that, over the 20 types of

Figure 11 Correlationbetween rins acceleration inthe axis with the highestweighted acceleration withthe root-sums-of-squares ofaccelerations in all threeaxes (bandwidth 8 to 1000Hz; no data available fortools D, E, L, or 0).

CoE 40Co

E

30EaxorP

as 20

(D

C.)0Co 10'a0)4-C

.1_

tool, there is a strong correlation betweenunweighted acceleration magnitudes mea-sured between 8 and 1000 Hz and those mea-sured between 8 and 500 Hz.

Although at low frequencies several of thetools have high magnitudes below 20 Hz,Figure 1OB shows a high correlation betweenmagnitudes measured between 8 and 1000 Hzand those measured between 20 and 500 Hzwhen using unweighted acceleration. Theeffect of the frequency weighting in reducingthe importance of high frequencies is tostrengthen this correlation for most tools (figlOC). However, the two tools with dominantcomponents of vibration below 20 Hz (theroadbreaker and wacker) give much lower val-ues when frequencies below 20 Hz areexcluded from the analysis.

Figure lOD shows that restricting the fre-quency range to the band from 20 to 500 Hzresults in a broadly similar order of vibrationmagnitudes for both unweighted and weightedvibration magnitudes. However, the weightinghas raised the relative importance of sometools-such as the scabbler-in which thevibration is dominated by a component at fre-quencies just above 20 Hz compared withtools dominated by vibration at higher fre-quencies.

200 Fa

150 H

100

50

200 c

U

v0o

'a

EI-

4-C0

C)00

-;:I00Co

VU)-c0)0)

Co

Co

0

CoC.)

Us

c

I-

0

c

E0

S04)

-Co

ax

or

4-

TCN._

150 K

200

100

50

I

.1

.1

.1490lilo.,11(l0

0 40

0 k , , ,0 10 20 30 44Weighted acceleration(worst axis) (ms 2 rms)

Is -

, u..

82

e-

- L



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/


DIRECTIONFigure 11 shows the correlation between theuse of the "worst axis of vibration" (the vibra-tion in the axis giving the highest magnitudes)and the use of an overall value obtained fromthe rss of values obtained in all three axes. Themaximum difference possible is that the rss

values are a factor of V3 (1 73) greater. It can

be seen that in several cases the vibration was

dominated by one axis so that there was littledifference between considering the worst axisand considering the rss. However, for some

tools-for example, the drill and the chippinghammer-the vibration magnitudes in each ofthe three axes were sufficiently similar for therss to be almostV3 greater than the magni-tudes in the worst axis.

DiscussionFREQUENCY WEIGHTINGThe currently used frequency weighting (Wh)is based on an extrapolation and a simplifica-tion of the contours representing the manner

in which the sensations of discomfort causedby vibration of the hand depend on vibrationfrequency.8 At high frequencies the subjectivedata on which the standardized weighting was

based were limited to 300 Hz, and merelyextrapolated to 1000 Hz! At frequencies above60 Hz, the shape used in the standardizedweighting is such that it gives relatively greaterweight than was implied by the subjectivedata; indeed, at 1000 Hz the standardizedweighting gives a value 16 times greater thanwould be obtained by linear extrapolation ofthe experimental data. The extent of theextrapolation, the absence of any considera-tion of the physiological or pathological effectsof vibration, and little consideration of biody-namic responses to vibration, provides a basisfor some criticisms of the frequency weighting.Nevertheless, the weighting was adopted bythe responsible subcommittee of theInternational Organisation for Standardisationduring the 1 970s, and this generated a

momentum followed by many national stan-dardisation bodies and, later, by the EuropeanCommittee for Standardisation (CEN).Although it is not uncommon for the weight-ing, and indeed, many other aspects of thestandards to be questioned, it would beimprudent for potential users of the standardsto conclude that doubts allow them to disre-gard current measurement, evaluation, andassessment procedures, or avoid the actionsthey imply.A simple expression of the doubt about the

frequency weighting is to enquire whether therelative harm caused by vibration accelerationover the frequency range 8 to 1000 Hz is betterpredicted by the frequency weighting, or byassuming that all frequencies cause similarharm. Eliminating the weighting would reducethe importance of low frequency vibration or

increase the importance of high frequencyvibration. Without the frequency weighting, a

theoretical tool (A) with an acceleration a at a

frequency of 16 Hz would be as severe as atheoretical tool (B) with the same unweighted

acceleration, a, but a frequency of 1000 Hz.With the frequency weighting, tool A wouldhave an acceleration 62-5 times greater thanthat of tool B and be far more likely to causeinjury. However, real tools do not usually havesuch widely differing frequencies, so that over arange of tools there is a strong correlationbetween unweighted and weighted accelera-tions (fig 1GD). The correlation increases iffrequencies below about 20 Hz are excluded.Epidemiological studies can only be expectedto discern the effect of the frequency weightingwhen they compare the development of disor-ders from tools the relative importance ofwhich is appreciably affected by the frequencyweighting. Some epidemiological studies havefailed to find an improved correlation betweenvibration exposures and the development ofvibration-induced white finger by using theweighting.20 For some tools with a predomi-nantly low frequency percussive action, theassessments of vibration according to currentstandards seem to predict a more rapid devel-opment of symptoms than has been found.2Nevertheless, it must be recognised that thedynamic response of the flesh, finger, handand arm, and the mechanisms of injury mustdepend on the frequency of vibration and notonly the acceleration magnitude. So some fre-quency weighting is required, even if it is notidentical to frequency weighting (Wh) recom-mended in current standards.

Considerations of the adequacy of the fre-quency weighting must be coupled to con-siderations of the different effects ofhand-transmitted vibration: it does not seemlikely that one weighting will be appropriate toeach of the reported groups of effects of vibra-tion (vascular, articular, neurological, etc).The current weighting might be defended asbeing appropriate to some overall (but unde-fined) risk of injury from vibration: the adop-tion of a new weighting optimum forpredicting vibration-induced white fingercould result in failure to protect against neuro-logical and articular effects of vibration.Studies should therefore seek to uncover therole of vibration frequency in the developmentof vascular, neurological, articular, and othereffects of vibration if the current weighting isto be replaced. Such studies need not berestricted to investigations of prevalence andlatency but should also include biodynamicand physiological research. For example, theextent to which the effects of vibration arelocalised close to the point of contact withvibration may indicate whether high or lowfrequencies are responsible: effects restrictedto the area of contact will be associated withfrequencies high enough for the vibration notto be greatly transmitted to distant parts,whereas effects occurring away from the con-tact area are likely to be caused by lower fre-quencies. The localisation may be ascertainedby study of affected people whereas the dis-tinction between high and low frequenciesmay be discerned by biodynamic experiments.

Although modifications to the frequencyweighting, and even its abolition, have beenproposed, the ISO subcommittee responsible

83



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

for ISO 5349 is not currently persuaded thatthe evidence is sufficient to identify the formof any improved frequency dependence. In thelonger term, changes can be anticipated. It istherefore desirable to obtain and report vibra-tion measurements in a form which allows thesubsequent application of other frequencyweightings.22

FREQUENCY RANGEIt has been proposed that the frequency rangebe extended above 100023-26 and that specialconsideration be given to impulsiveness.27 Thiswould require a change to the measurementmethod and also a change to the evaluationmethod.

It is not simple to measure vibration on han-dles at high frequencies. The methods ofattaching accelerometers to handles used insome studies are not even valid at frequenciesas high as 1000 Hz: the mounting sometimeshas a resonance affecting measurements below1000 Hz, and the mass of the accelerometermay affect movements of the handle at thesehigh frequencies. At high frequencies, manyhandles will have modes of vibration such thatthe vibration magnitude will vary from onelocation to another. An extension of the mea-surement range to 2000 or 4000 Hz23 is likelyto increase errors in reported measurementsdue to artefactual resonances. Even if sometools have vibration at frequencies greater than1000 Hz that can cause disorders, it seemslikely that the disorders currently caused bymany tools, possibly most tools, arise fromlower frequencies. In view of the considerabledifficulties, and consequent errors, introducedby the measurement of higher frequencies,some scientists doubt whether current evidenceis sufficient to justify a requirement to measureat frequencies above 1000 Hz for all tools.Measurement difficulties are not sufficient

in themselves to imply that the high frequen-cies are not important, although the currentlyused frequency weighting does imply littleimportance to the high frequencies. Figures 6,8, and 9 indicate that weighted values wouldnot be changed by increasing the frequencyrange beyond 1000 Hz. Consequently, anextension to the frequency range also implies achange to the frequency weighting. Thereseems to be no basis for selecting any particularweighting so it may be unwise to define aweighting arbitrarily: to collect data on thecharacteristics of high frequency vibration ontools is a preferable initial step.

Although some tools may have magnitudesof acceleration at high frequencies which aresufficient to cause injury, it seems doubtfulwhether frequencies above 1000 Hz are amajor contributor to the effects of vibration(particularly vibration-induced white finger)from the use of most tools (table 1). On thisbasis, whereas it currently seems reasonable toadvocate the measurement of unweightedacceleration to frequencies above 1000 Hzwhen this is practicable, it does not yet seemnecessary to include such measurements in theassessment of vibration. As more experience isgained on the measurement difficulties at high

frequencies, and as more experience accumu-lates about the importance of the higher fre-quencies, so their role in revised dose-effectguidance may be reconsidered.

There are problems introduced by theinclusion of frequencies as low as 8 Hz.Voluntary movements when controlling a tool,movements of cables, and other causes of lowfrequency problems-for example, dc shift-can artificially raise the measured vibrationmagnitudes. Restricting the lower frequencyto, say, 20 Hz would reduce the importance ofthese problems. Although there is little reasonto suppose that these low frequencies are anormal cause of vibration-induced white fin-ger, high magnitudes of vibration below 20 Hzmust be undesirable. As long as evaluationprocedures are based on a single frequencyweighting, it may be warranted to include lowfrequencies within the bandwidth of the mea-surement and evaluation procedures. Furtherconsideration of the extent of low frequencyvibration on tools may help to discern theneed for limitations at low frequencies.

MAGNITUDEThe dose-effect guidance given in an annex toISO 5349, 19869 indicates that the incidenceof finger blanching increases in proportion tothe square of the vibration magnitude, if othervariables are unchanged:

Fah~.E 2C = (3)T(4)~ (3

where: C = prevalence of vibration-inducedwhite finger (expressed as a percentage); ah,= frequency-weighted acceleration (ms-2 rms);E = years of regular exposure to vibration;t = duration of daily exposure (in same unitsas T(4)); T(4) = four hours (in same units as t).

This relation between vibration magnitudeand the development of vibration-inducedwhite finger seems to have arisen from consid-erations other than direct evidence of this rela-tion. The full implications are worthy ofconsideration. The heavy weighting placed onacceleration (as opposed to duration of dailyexposure) encourages lower magnitudes ofvibration. However, the weighted magnitude isaffected by the frequency weighting: if theweighting is slightly in error the predictedprevalence will be in even greater error. Thedose-effect guidance is restricted to the range10% to 50% in ISO 5349, 19869 correspond-ing to a 2-2:1 range of vibration magnitudes.For a particular duration of exposure, a 2-2:1increase in vibration magnitude (or a 2-2:1error in the frequency weighting) wouldincrease the vibration from being a problembarely greater than the background level ofvascular problems to more than half of theexposed people developing the condition. It isinconceivable that the frequency weighting isaccurate to within 2-2:1 over the frequencyrange 8 to 1000 Hz. Consequently, in studiesin which the prevalence of vibration-inducedwhite finger has been correctly predicted bythe procedures of ISO 5349, 19869 this musthave arisen from a chance (or restricted) com-bination of variables: such findings do not ver-

84



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/


ify the full process of measurement, evalua-tion, and assessment defined in the standard.

In ISO 5349, 19869 vibration magnitude isassumed to be measured by its rms value. Thisassumes a relation between the accelerationmagnitude and the square root of the durationof exposure. The relation is used to obtain anequivalent four hour vibration magnitude forany combination of continuous, intermittent,or impulsive events. It seems unlikely thatsuch a simple relation will hold over the verywide range of magnitudes and durationsinvolved in hand-transmitted vibration.Research must explore the applicability of thisrelation as much as the relations between themagnitude and frequency of vibration and theyears of exposure. As with whole-body vibra-tion, this time-dependency seems to allowexcessive magnitudes with short durations ofexposure: it may therefore underestimate theseverity of vibrations containing occasionalhigh magnitude shocks.8The possibility of a threshold level for the

occurrence of vascular disorders has been sug-gested.20 A threshold might vary from personto person and depend on some other variables,but above the threshold the outcome would beprimarily related to duration of exposurerather than vibration magnitude. The conceptmight be consistent with some epidemiologicaldata, but should currently be considered ahypothesis suggesting specific lines of investi-gation. The concept could lead to differenttypes of prevention than those implied by acontinuous relation between the incidence ofvibration-induced white finger and the squareof the acceleration magnitude.

DIRECTIONThere has been inconsistency between stan-dards in whether the vibration should beassessed only in the direction with the highestmagnitude (ISO 5349, 19869) or whether theweighted magnitudes in all three axes shouldbe summed together, to form the rss (ISO7505, 198628). The rss is a higher value,although for many tools the increase is notgreat compared with the likely errors causedby the choice of frequency weighting and thefrequency range (fig 10D). However, a differ-ence in magnitude of'/3:1 corresponds to adifference of 3:1 in the predicted prevalence ofvibration-induced white finger according toISO 5349, 1986.9 Considerations of unifor-mity are leading to the use of the summationmethod in most standards and this is alsoadvocated in the proposed Exposure Directivefrom the EU. The method is also consistentwith results of some studies of acute effects ofhand-transmitted vibration.29 Even so, the useof only a single axis is advocated for some typetesting of specific tools (parts of ISO 8662,1986"1). It has been argued that the use of therss has the merit of avoiding erroneously lowvalues arising from misidentifying the worstaxis.Of greater interest, but rarely debated, is the

assumption that all three axes of vibration areequally important in causing injury. The trans-mission of vibration into the tissues of the fin-

gers and hand, and their transmission to otherparts of the body are different for motion nor-mal to the surface as opposed to motion in theshear axes.303' The sensations caused by thevibration also differ in each axis. This suggeststhat a different weighting should be applied toeach of the axes before their effects are com-pared or combined. On this basis, it could beargued that the use of the rss with the sameweighting for each axis will conceal the truevibration severity and introduce errors intosome assessments.When a hand is wrapped around the handle

of a tool, the vibration in one direction can benormal to one part of the hand and tangentialto another part of the hand. It would beimpractical in these situations to use differentweightings for different locations of the hand.On some tools, however, the dominant direc-tion is clearly defined and a handle might bedesigned so that the vibration predominantlyentered the hand in a shear direction, ratherthan the normal direction. This might beexpected to alter the hazards associated withthe vibration, but this would not be reflectedin the measured severity of the rss. The trendtowards the use of the rss method has implica-tions for (a) tool designers, (b) those interpret-ing epidemiological data, (c) the design ofinstruments used to assess vibration on tools,(d) the effectiveness of gloves in attenuatingvibration, and (e) dose-effect guidance whichmight differ from that appropriate for singleaxis measurements.

DURATION OF EXPOSUREIn ISO 5349, 19869 the predicted incidence offinger blanching increases in linear proportionto the duration of exposure to vibration duringthe day, but increases in proportion to thesquare of the duration of exposure to vibrationexpressed in years (equation 3). The influenceof years of exposure was derived from "averagelatency" data, whereas the effect of hours ofexposure arose from a convenience rather thanevidence that this was the case.'2 The linearrelation between years of exposure and accel-eration magnitude seems to be consistent withsome epidemiological studies with percussiveand rotary quarry tools and with chain saws."34It is incontrovertible that the prevalence andseverity of vascular and neurological effects ofvibration increase with increases in the totalduration of exposure to vibration. It is not soevident that there is a large difference betweendaily and yearly exposures.The effect of duration of daily exposure is

important as reducing duration of exposurewill sometimes be more practical than reduc-ing vibration magnitude. Equation 3 showsthat a fourfold reduction in duration of dailyexposure, or a twofold reduction in years ofexposure, is required to achieve the same as atwofold reduction in vibration magnitude.Equation 3 therefore contains unfoundedspeculation that long daily exposures for a fewyears will be less harmful than the same totalduration of exposure spread out over moreyears.The ISO 5349, 19869 specifies how expo-

85



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

sures to different magnitudes and durations ofvibration during the day should be accumu-

lated so as to determine an equivalent dailyexposure. There is no defined procedure foraccumulating exposures which vary betweendays or over weeks or years. For many people,exposure to hand-transmitted vibrationvaries from job to job or with seasonalchanges. It might be thought appropriate to

assess the severity of such exposures by eithersumming the fractions of yearly exposure or

summing the prevalences arising from eachexposure, but these give different results.Mathematically, the most convenient ap-

proach would be based on total hours of expo-sure so that the same time dependency was

used for exposures within the day and across

years. However, this is also unsubstantiatedand gives results very different from thoseimplied by equation 3.

Current standards offer no alleviation of therated severity of exposures to hand-transmit-ted vibration from rest periods, either breaksduring the day or periods away from exposure

during the year. During the day, the totalhours of exposure are used in the assessment,irrespective of whether the exposure is inter-mittent or continuous. During the year themethod is unclear, but often a full year will becounted irrespective of days, weeks, or monthswithout exposure to vibration. With theimportance of the years of vibration exposure

in equation 3, and the lessening of symptomsof vibration-induced white finger sometimesexperienced after absence of exposure, thisplaces a high emphasis on years in a job.

So far, methods of predicting the percent-age of people who will develop vibration-induced white finger have been based on cross

sectional epidemiological studies. However,such studies do not easily surrender therequired data on duration of exposure. Thehistorical exposure of people cannot normallybe estimated with any precision. Also, there is a

need to distinguish between the mean latency(duration between onset of exposure and onsetof symptoms) for an affected percentile in thepublished data and the mean exposure timerequired for the same percentile to develop thecondition.

FORCE AND CONTACT AREA

Consideration of the appropriate frequencyweighting should allow recognition that thesame weighting may not be appropriate forvibration entering the fingers as for vibrationentering the hands. Further, as hand-transmit-ted vibration causes central as well as localeffects, s it could be argued that exposure oftwo hands to vibration is more severe than theexposure of a single hand. Current standardsand guidance, including the proposed EUExposure Directive, assume that it is sufficientto consider only the hand exposed to the high-est magnitude of vibration.

Current methods of evaluation make no dis-tinction between high and low contact forces,or between a force applied by a large area- forexample, the palm of the hand or a smallarea for example, the finger tips. The trans-

mission of vibration into the body and thetransmission of vibration through the body areaffected by these factors in a complex mannerwhich depends on the vibration frequency andvibration direction. The local circulation mayalso be influenced by an excessive grip force.A grip force as low as possible consistent

with safe tool operation may be advocated inmany cases. Increasing the contact area so asto minimise contact pressure might also bebeneficial. These factors have implications forthe design of tools, especially handles.Methods for the standardised measurementand evaluation of these factors are required,and experience with the application of thesemethods may be needed before considerationof how they should influence assessments ofvibration exposures.

POSTUREAs hand-transmitted vibration can result in adisorder associated with reduced peripheralcirculation, other factors which impede circu-lation require consideration. Work with thehands above the head, high grip forces and,possibly, prolonged work in a sedentary pos-ture may reduce peripheral circulation. Again, asimple means of assessing these factors isrequired before their measurement can beadvocated and they can influence the assess-ment of exposures to hand-transmitted vibra-tion.

TEMPERATUREBody temperature has long been assumed toplay a part in the development of the vasculareffects of hand-transmitted vibration (vibra-tion-induced white finger). Although the evi-dence is not unequivocal, there seems littlescope for advocating anything other than awarm environment consistent with maintain-ing good circulation. However, the non-vascu-lar effects may not be alleviated by warmthand so maintaining a warm environment can-not be considered sufficient to prevent disor-ders caused by hand-transmitted vibration.Although warm and dry hands can easily berecommended (including the provision of suit-able gloves) this does not affect the assessmentof the vibration severity, although it maylessen symptoms and might lessen some of theharmful effects of hand-transmitted vibration.

SUBJECT VARIABILITYThere is a large variability between people intheir susceptibility to the effects of hand-trans-mitted vibration. Among users of the sametools the effects may appear in some within ayear or so but not after 20 or 30 years in others.Study of the differences between people mayprovide some clues as to the relative impor-tance of some causal variables already men-tioned, and others, including constitutionalvariables, not yet incriminated in the develop-ment of disorders associated with hand-trans-mitted vibration.The variability in susceptibility to the effects

of hand-transmitted vibration is currentlyreflected in the increased percentage predictedto have vibration-induced white finger as

86



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/


either the duration of exposure or the vibrationmagnitude increases. British Standard 6842,198710 and a recently proposed revision to ISO5349, restrict the dose-effect guidance to thoseconditions causing a 10% incidence of vibra-tion-induced white finger. Continuous dose-effect guidance, such as that in equation 3,should be based on evidence of the relationsbetween the four variables. A 10% criterionreduces the variables by one but also changesthe justification to a simple prediction ofwhether symptoms are likely: an action level.A 10% prevalence of vibration-induced whitefinger is then justified as being at least doublethe prevalence of similar symptoms in a con-

trol group.Observation of affected workers shows that

some may develop neurological disorderswhile others may develop vascular disorders.Further study of the factors associated withthe differences, including the elaboration ofimproved objective indicators of disorders,may help to understand the causal mecha-nisms behind these effects.

OTHER VARIABLESFor a complete understanding of the form andmechanisms of disorders caused by hand-transmitted vibration it may be necessary toconsider other variables when assessing expo-sures to vibration at the workplace, or theseverity of vibration on specific tools. With pre-sent understanding, the above catalogue ofvariables is already overly complex.

In concept, a more complete dose-effectrelation could be given by combining the vari-ous variables into one measure.'9 Although thismay provide a useful philosophical frameworkit may not be helpful when assessing vibration inspecific situations. The process of forming an

overall equation giving a method of combiningthe effects of different variables involves con-

sidering whether variables have continuous or

discrete on-off effects, and whether variablesinteract. A recognition of potential complexityis desirable, but the application of knowledgeto specific situations must be simple. An ele-gant simplicity might be artificially created: itsform being enhanced by understanding thatgreater accuracy of application is not automati-cally achieved through greater complexity.

INERTIA AND MOMENTUMTravellers seeking directions must continuefrom where they find themselves by using theavailable paths. The map maker may standaside and consider new journeys in the future:he will be influenced by the impressions of pasttravellers but concerned not to confuse thosefamiliar with the route: only small changes maybe possible.The enormous number of standards refer-

ring to the evaluation methods-for example,the frequency weighting and frequency range-originally defined in ISO 5349, 19869 mustimpede changes to the standard. A change tothe weighting, the frequency range, or to theuse of rss alters the effective magnitude andtherefore changes the assessment of vibrationon tools. Any such change will also influence

the claimed effectiveness of potential protectivemeasures, such as vibration isolation devices orgloves.The ISO 5349, 19869 has provided a

method of evaluating hand-transmitted vibra-tion and undoubtedly contributed to a reduc-tion in hazardous exposures. For some, thissuccess is a sufficient justification for its publi-cation and even its retention. Others mightargue that greater success would arise from animproved standard based on repeatable dataobtained by scientific method, rather than onthe consent of the members of a small com-mittee. Certainly, an ISO subcommittee isable to promulgate a new standard on thebasis of insufficient objections rather than as aresult of reasoned argument: the reasoningbehind many standards does not reach thelevel of rigor needed for a paper in a reputablescientific journal. At a time when there is littleknowledge, this arbitrary decision making pro-duces a standard when none may be possibleby reasoned argument. When there is knowl-edge the approach is irresponsible if it ignoresor alienates those with expertise. The futurewill show how measurement and evaluationstandards and assessment guidelines have fol-lowed the path uncovered by advancingknowledge of the relations between hand-transmitted vibration and various signs andsymptoms.

ConclusionsThere are many challenges remaining in themeasurement, evaluation, and assessment ofhand-transmitted vibration. The mechanismof injury is not known and even the full range ofinjuries is not agreed. It is therefore not sur-prising that methods of predicting the effectsof vibration from measures of exposure tovibration have significant uncertainties.

At the time of writing there is a need to bal-ance two opposing forces. On one side it isnecessary to express clearly that hand-trans-mitted vibration does cause injury and that themeasurement and evaluation of vibrationexposures according to current methods willusually form a significant and useful part of apreventive programme. On the other sidethere is the need to identify areas in which cur-rent measurement and evaluation proceduresrequire improvement. In some cases this maybe achieved by a greater coherence betweenscience and the standards.32 In other areasthere is a need for research leading to newunderstanding and consequent changes toboth thinking and standards. This desire forprogress is the raison d'etre of the scientist: itshould not be used as an excuse for inactionwhen the measurement and evaluation ofvibration with currently standardized methodsmay reasonably be expected to be beneficial.

Current uncertainties in the relationbetween vibration and its effects have an influ-ence on the collection of data on which futurestandards are to be based. Scientific investiga-tions must reach outside the web formed bycurrent standards if they are to make newdiscoveries: the truth is laid down in nature

87



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

Griffin

and not in standardization committees.Standardised methods of measurement andevaluation should not be confined by crudeand approximate assessment methods: theyshould allow, assist, and encourage the collec-tion of improved data on human exposure tohand-transmitted vibration. This requires thestandardization of methods of collecting andreporting the spectral, axial, and temporalcharacteristics of vibration exposures, andother variables, over a wide range of condi-tions.The analyses of tool vibration presented in

this paper suggest that when using the currentfrequency weighting (Wh) it is likely to be suffi-cient to restrict the measurement of vibrationto frequencies below about 250 Hz. Withoutthe frequency weighting, many tools have sig-nificant vibration at frequencies above 250 Hz,but only a few would be greatly underesti-mated with an upper limit of 500 Hz ratherthan the current 1000 Hz limit. A significantlowering of the upper frequency could appre-

ciably assist the method of measuring hand-transmitted vibration, or even change themanner in which this is performed. At low fre-quencies there are arguments for excludingfrequencies below about 20 Hz, especiallywhen attempting to predict the vascular effectsof vibration. Considering the nature of thevibration on many tools, and currently avail-able epidemiological data relating to vibration-induced white finger, it is difficult to see a

justification for the current combination of thefrequency range (8 to 1000 Hz) and the fre-quency weighting (Wh). It seems that eitherthe frequency range should be reduced or themeasurements should be obtained without thisfrequency weighting: epidemiological andexperimental studies are required to determinethe appropriate combination of frequencyrange and frequency weighting.

Uncertainties in the effects of some vari-ables may have a large effect on the perfor-mance of some proposed preventativeprocedures. For example, in the past, few sci-entists have advocated antivibration gloves as a

useful solution to the effects of hand-transmit-ted vibration. Gloves may attenuate high fre-quencies of vibration but do not attenuate thelow frequencies. It is argued that most toolsare dominated by low frequencies at whichsuch gloves have little beneficial effects. Inpart, this arises from the influence of the fre-quency weighting used in ISO 5349, 19869(fig 2). The use of unweighted accelerationover the range 8 to 1000 Hz would increasethe perceived effectiveness of such gloves. Anextension of the frequency range to frequen-cies above 1000 Hz (or raising the low fre-quency limit to, say, 20 Hz) could furtherincrease the apparent value of gloves. Whilenew standards for antivibration gloves (forexample, ISO 10819, 199636) make use of thecurrent frequency weighting, their future evo-

lution should include consideration of the con-

sequences of changes to the frequencyweighting.

In the past, a lack of knowledge has allowedthe production of standards for the measure-

ment, evaluation, and assessment of vibrationbased on committee consensus. Consultationwith the scientific literature, or carefullyargued debate over the interpretation or appli-cation of published data is not always part ofthe process in forming a personal, national, orinternational view on human vibration stan-dards. It is to be hoped that future standardswill identify the foundations on which they areoffered and that this will include informationpublished in scientific, medical, and engineer-ing journals available for peer review. Whenthe limitations to knowledge prevent consen-sus based on such foundations, the decisionsand their justification should be declared. Inall cases, methods proposed by standardisa-tion committees would benefit from trialapplication and thorough analysis before theyare offered for ritual approval by membercountries. Extraordinary as it may seem, dis-tinguished bodies have been prepared to agreeto proposed new vibration standards withoutbeing aware of the basis on which they havebeen formulated, without seeing any exampleof the application of the proposed standard,and without being provided with informationon the consequences of the standard. This,and the anonymity involved in the productionof standards, allows the publication of arbi-trary standardized methods without this beingmade clear to users of a standard.

1 Departmental Committee on Compensation for IndustrialDiseases. Report of the departmental committee on compensa-tion for industrial diseases. London: HMSO, 1907:Cd3495.

2 Corsini A. Gli strumenti ad aria compressa in rapportoall'igiene dell'operaio. II Ramazzini 1907;2:437-52.

3 Loriga G. II lavoro con i martelli pneumatici. (The use ofpneumatic hammers.) Boll Ispett Lavoro 1911;2:35-60.

4 Cargile CH. Raynaud's disease in stone cutters using pneu-matic tools. JAMA 1915;65:582.

5 Bureau of Labour Statistics. Industrial accidents andhygiene series No 19. Effect of the air hammer on stone cut-ters. Final report, bulletin 236. Washington, DC: BLS,1918:PB-254-601.

6 Industrial Injuries Advisory Council, Department of Healthand Social Security. Social Security Act 1975: Vibrationwhitefinger, London: HMSO, 1981: Cmnd 8350.

7 Department of Social Security, Social SecurityAdministration Act 1992. Hand arm vibration syndrome(vascular and neurological components involving the fingersand thumb), Report by the Industrial Injuries AdvisoryCouncil in accordance with Section 171 of the SocialSecurity Administration Act 1992 on the questionwhether hand arm vibration syndrome should be pre-scribed. London: HMSO, 1995:CM 2844.

8 Griffin MJ. Handbook of human vibration, London:Academic Press, 1990.

9 International Organisation for Standardisation. Mechanicalvibration-guidelines for the measurement and the assessmentof human exposure to hand-transmitted vibration. Geneva:ISO, 1986:5349.

10 British Standards Institution. Measurement and evaluation ofhuman exposure to vibration transmitted to the hand.London: BSI, 1987:6842.

11 International Organisation for Standardisation. Measure-ment of vibrations in hand-held power tools-Part 1: general.Geneva: ISO/DIS, 1986:8662/1.

12 International Organisation for Standardisation. Humanresponse to vibration-measuring instrumentation. Geneva:ISO, 1990:8041.

13 British Standards Institution. Instrumentation for the mea-surement of vibration exposure of human beings. Part 1.Specification for general requirements for instrumentation formeasuring the vibration applied to human beings. BritishStandard. London: BSI, 1991:7482 Part 1.

14 British Standards Institution. Instrumentation for the mea-surement of vibration exposure of human beings. Part 2.Specification for instrumentation for measuring vibrationtransmitted to the hand. London: BSI, 1991:7482 Part 2.

15 European Committee for Standardisation. Mechanicalvibration-Guidelines for the measurement and the assess-ment of human exposure to hand-transmitted vibration.Brussels: ENV, 1992:25349.

16 American National Standards Institute. Guide for the mea-

88



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/


surement and evaluation of human exposure to vibrationtransmitted to the hand. New York: ANSI 1986:S3-34.(ASA 67.)

17 Council of the European Communities. On the approximationof the laws of the member states relating to machinery.Council Directive 89/392/EEC. Brussels: Official Journalof the European Communities, 1989: June, 9-32.

18 Commission of the European Communities. Amended pro-posal for a Council Directive on the minimum health andsafety requirements regarding the exposure of workers to therisks arising from physical agents-Individual Directive inrelation to Article 16 of Directive 8913911EEC. Brussels:Official Journal of the European Communities, 1994: NoC 230, 19 8.94, 3-29.

19 Griffin MJ. Foundations of hand-transmitted vibrationstandards. Nagoya 7Med Sci 1994:57(suppl): 147-64.

20 Nelson CM, Griffin MJ. Comparison of predictive modelsfor vibration-induced white finger. In: Dupuis H, ChristE, Sandover J, Taylor W, Okada A, eds. Proceedings of the6th International Conference on Hand-arm Vibration.Bonn, 1992. Sankt Augustin: 1993:875-83. (HVBG, D-53754).

21 Walker DD, Jones B, Ogston S, Tasker EG, Robinson AJ. Astudy of white finger in the gas industry. Br Ind Med1985;42:672-7.

22 Nelson CM, Griffin MJ. Improving the measurement andevaluation of hand-transmitted vibration. In: Dupuis H,Christ E, Sandover J, Taylor W, Okada A, eds. Proceed-ings of the 6th International Conference on Hand-armVibration. Bonn, 1992. Sankt Augustin: 1993:613-23.(HVBG, D-53754).

23 National Institute for Occupational Safety and Health.Criteria for a recommended standard: occupational exposureto hand-arm vibration. Cincinnati: US Department ofHealth and Human Services, National Institute forOccupational Safety and Health, DHHS (NIOSH),1989. (Publ No 89-106.)

24 Pekkarinen J, Starck J, PyykkO I. High-speed digitalmethod to measure impulsive hand-arm vibration.Vibration at work. In: Proceedings of 3rd InternationalSymposium of the International Section of the ISSA forResearch on Prevention of Occupational Risks. Vienna,19-21 April, 1989, 47-50.

25 Lundstrom R. Health effects of hand-arm percussive tools.An overview. In: Okada A, Taylor W, Dupuis H, eds.Proceedings of the 5th International Conference on Hand-Arm Vibration. Kanazawa, 23-26 May, 1989. Kanazawa:Kyoei Press, 1990:113-6.

26 Dandanell R, Lundstrom R. Alternative methods forassessment of riveting hammer vibration, In: Dupuis H,Christ E, Sandover J, Taylor W, Okada A, eds.Proceedings of the 6th International Conference on Hand-armVibration. Bonn, 1992. Sankt Augustin: 1993:925-31.(HVBG, D-53754.)

27 Starck J, PyykkO I. Impulsiveness of vibration as anadditional factor in the hazards associated with hand-arm vibration. Scand Work Environ Health 1986;12:323-6.

28 International Organisation for Standardisation. Forestrymachinery-Chain saws-Measurement of hand-transmictedvibration. Geneva: ISO, 1986:7505.

29 Stelling J, Dupuis H. Beanspruchung des Hand-Arm-Systemsdurch multiaxiale mechanische Schwingungen. (Stress to thehand-arm system caused by multiaxial mechanical vibra-tion.) Hauptverband der gewerblichen Berufsgenos-senschaften, 1995.

30 Reynolds DD. Hand-arm vibration: a review of three years'research. In: Proceedings of the International OccupationalHand-Arm Vibration Conference. 1975. Cincinnati:National Institute for Occupational Safety and Health1977:99-128. (DHEW (NIOSH) Pub No 77-170.)

31 Reynolds DD, Standlee KG, Angevine EN. Hand-armvibration, Part III: subjective response characteristics ofindividuals to hand-induced vibration. Journal of Soundand Vibration 1977;51:267-82.

32 Griffin MJ. Hand-transmitted vibration and its effects: acoherence between science and standards? In: Proceedingsof the Stockholm Workshop 94, Hand-Arm Vibration Syn-drome: Diagnostic and Quantitative Relationships to Exposure.May 25-28, 1994. Arbete och Halsa 1995;5:47-58.

33 Bovenzi M. Italian Study Group on Physical Hazards in theStone Industry. Hand-arm vibration syndrome and dose-response relation for vibration-induced white fingeramong quarry drillers and stonecarvers. Occup EnvironMed 1994;51:603-11.

34 Bovenzi M, Franzinelli A, Mancini R, Cannava MG,Maiorano M, Ceccarelli F. Dose-response relation forvascular disorders induced by vibration in the fingers offorestry workers, Occup Environ Med 1995;52:722-30.

35 Bovenzi M, Griffin MJ, Ruffell CM. Acute effects of vibra-tion on digital circulatory function in healthy men. OccupEnviron Med 1995;52:834-41.

36 International Organisation for Standardisation. Mechanicalvibration and shock-hand-arm vibration-method for themeasurement and evaluation of the vibration transmissibility ofgloves at the palm of the hand. Geneva: ISO, 1996:10819.

Correspondence and editorialsOccupational and Environmental Medicine wel- minimum. Letters are accepted on thecomes correspondence relating to any of the understanding that they may be subject tomaterial appearing in the journal. Results editorial revision and shortening.from preliminary or small scale studies may The journal also publishes editorials whichalso be published in the correspondence are normally specially commissioned. Thecolumn if this seems appropriate. Letters Editor welcomes suggestions regardingshould be not more than 500 words in length suitable topics; those wishing to submit anand contain a minimum of references. Tables editorial, however, should do so only afterand figures should be kept to an absolute discussion with the Editor.

89



.bmj.com

/O

ccup Environ M


ebruary 1997. Dow

nloaded from

http://oem.bmj.com/

measurement, · measurement,evaluation, andassessmentof occupational exposuresto hand-transmitted...

Documents