Results

QUESTIONNAIRE DISTRIBUTION AND RETURNS

A total of 22 415 subjects were selected for study, but 221 were excluded on their general practitioners' advice, so that 22 194 questionnaires were posted. Of these, 1028 questionnaires (4.6%) were returned as “address unknown”, “moved away” or “deceased”, and 56 were completed by someone other than the intended recipient. Usable responses were therefore obtained from 12 907 subjects, giving a response rate of 58% overall or 61% among those who could be contacted.

The overall response rate by practice ranged from 33.1% in Lambeth to 70.2% in Devon, and in the armed services varied from 60.9% (Army) to 72.7% (RAF). The response rate was higher in women than men (67%v 52%), and tended to be higher at older ages, but was similar in the summer and winter.

Seventy three per cent of respondents (79% of men and 65% of women) were in a paid job or were self employed in the week preceding completion of the questionnaire, and most of these (9084 out of 9368) were at work in that week.

Some subjects had worked in more than one job, but exposure to HTV in secondary jobs was rare (only 37 men and eight women), so further analysis was confined to the main job held. The relative frequency of occupations in the sample was generally similar to that of the 1991 national census,18 except that the sampling design led to overrepresentation of the defence sector.

FREQUENCY OF OCCUPATIONAL EXPOSURE TO HAND TRANSMITTED VIBRATION

Altogether, 2945 men (42.6% of all male respondents) reported having been occupationally exposed to HTV at some time, including 1727 who reported exposure to a source of HTV in the week preceding completion of the questionnaire (31.5% of those in employment). The estimated 1 week prevalence of exposure among men of working age in the national population was 20.5% (95% confidence interval (95% CI) 19.7 to 21.3%) overall, and 31.9% (95% CI 30.7 to 33.1%) in those at work.

Among women, exposure was less common: 475 women (7.9% of the sample) had ever been exposed occupationally, and 225 women (3.7%) had been exposed in the past week. In the national population, the 1 week prevalence of exposure was estimated as 2.9% (95% CI 2.6 to 3.3%) of all women and 6.4% (95% CI 5.6 to 7.2%) of those in work.

For all occupations combined, it was estimated that 4 208 000 men (95%CI 4 044 000 to 4 371 000) in Great Britain were exposed to a source of HTV in the past week. The occupations which contributed the largest estimated numbers of exposed men were those belonging to group 5 of SOC 90 (the craft and related occupations major group). These included: metal working production and maintenance fitters (299 000 exposed, 95%CI 272 000 to 324 000), carpenters and joiners (254 000, 95%CI 242 000 to 267 000), electricians and electrical maintenance fitters (194 000, 95% CI 177 000 to 211 000), motor mechanics and autoengineers (157 000, 95%CI 139 000 to 176 000), plumbers and heating and ventilating engineers (141 000, 95%CI 127 000 to 155 000) and builders and building contractors (131 000, 95%CI 119 000 to 144 000). These six occupations accounted for an estimated 28% of the men with exposure nationally.

The industries which contributed the largest estimated numbers of exposed men were: construction (979 000 exposed, 95%CI 907 000 to 1 052 000), motor vehicle maintenance and repair (276 900, 95%CI 234 000 to 320 000), agriculture (219 000, 95%CI 194 000 to 244 000), manufacture of fabricated metal products other than machinery and equipment (218 000, 95%CI 185 000 to 252 000), and defence (162 000, 95%CI 123 000 to 201 000). These five industries included an estimated 49% of the men with exposure nationally.

The 1 week prevalence of occupational exposure to HTV in women corresponded to an estimated 667 000 women (95%CI 582 000 to 751 000) exposed nationally. The occupation with the largest estimated number of exposed women was cleaner or domestic worker (241 000, 95%CI 193 000 to 288 000), followed by hairdresser (12 000, 95%CI 4000 to 20 000). The education sector gave rise to the biggest estimated number of exposed women by industry (111 000, 95%CI 83 000 to 139 000).

MOST COMMON SOURCES OF HAND TRANSMITTED VIBRATION

Table 1 lists the sources of exposure to HTV in the past week that respondents reported most often, and the estimated numbers of people with such exposure in Great Britain. Among employed men in the sample, the most common exposures were hammer drill (11.9%), hand held portable grinder (10.9%), and jig saw (7.5%). The national estimates for these tools imply that there are more than 1.7 million users of hammer drills, more than 1.5 million users of hand held portable grinders, and more than 1 000 000 users of jig saws. Among women, floor polishers were the source of exposure reported most often, followed by nailing and stapling guns.

Exposure to multiple sources of HTV was fairly common. Thus, 1218 men (22% of employed men in the sample) were exposed to two or more sources of HTV in the week before they completed the questionnaire, including 301 men (5.5%) who reported using six or more vibratory tools. In women, occupational exposure to more than one source of HTV was uncommon (only 39 subjects among 3878 employed respondents).

PERSONAL DAILY EXPOSURES TO VIBRATION IN THE PAST WEEK

The HSE has suggested that when exposures exceed the equivalent of 2.8 ms^-2 for 8 hours (A(8) > 2.8 ms^-2 root mean squared (rms)), control measures and a programme of health surveillance are likely to be required.20 Among 704 exposed men from the sample who supplied full information, 185 (3.4% of male respondents) indicated exposures above this suggested action level, including 77 (1.4%) with estimates of A(8) higher than 5 ms^-2 rms.

Full information on exposure was missing for 1023 of the 1727 men exposed to HTV (either because exposure times or data on representative vibration magnitudes were missing), but in workers who were exposed to several sources, sufficient information was provided to indicate that at least a further 260 men had a minimum A(8) >2.8 ms^-2. On this basis, the minimum number of men in the population exceeding this action level was estimated to be 1 198 000 (95%CI 1 092 000 to 1 304 000). Among women, by contrast, these levels of exposure were uncommon (only nine of 3878 employed respondents had a minimum estimated A(8) >2.8 ms^-2). Figure 2 presents further details in the form of a cumulative frequency curve showing the estimated numbers in the population exceeding different values of A(8) in the past week. Curves are drawn separately for men and women, and are minimum estimates, in so far as exposure information was sometimes incomplete.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Minimum estimated numbers of men and women in Great Britain whose exposure level (average daily A(8) rms) in the past week exceeded the values indicated.

The prevalence of exposure did not differ importantly between subjects who responded at the first invitation and those who required a reminder. In men, for example, 20.4% of early responders reported an exposure compared with 21.0% of late responders; and among men the corresponding frequencies for a minimum estimated A(8) >2.8 ms^-2 rms in the past week were 6.0% and 7.6%.

Estimates of exposure were similar in the summer and winter. Thus, among men from the first posting, 24% reported an exposure in the past week and 6.2% had a minimum estimated A(8) >2.8 ms^-2; whereas the corresponding figures for wintertime respondents were 26% and 7.5% respectively.

Figures 3 and 4 detail, for occupations and industries where exposures were most common, the prevalence and extent of exposures in the past week. Analysis was confined to occupations with >30 respondents and industries with >60 respondents; and minimum prevalences are presented, to the extent that information on exposure was sometimes missing. Estimated exposures >2.8 ms^-2 were found most often in builders and building contractors (63%), followed by welding trades (46%), carpenters and joiners (43%), and motor mechanics (38%). The industries where A(8) values most often exceeded 2.8 ms^-2 were construction (29%), motor vehicle repair and maintenance (23%), agriculture (21%), and manufacture of basic metals (13%). The highest exposures (estimated A(8) >5 ms^-2 rms) occurred most often in the construction crafts (bricklayers and masons (29%), builders and building contractors (24%), carpenters and joiners (15%), and plumbers (9%)); in motor mechanics (16%); welding trades (15%); labourers (15%); gardeners and groundsmen (13%), and farm owners (8%).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

The occupations in which exposures to HTV most commonly arose in the past week among employed men.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

The industries in which exposures to HTV most commonly arose in the past week among employed men.

These occupations and industries also contributed the largest numbers to national estimates of the minimum frequency of exposures exceeding the HSE threshold for control measures (A(8)>2.8 ms^-2, tables 2 and 3), with some 115 000 carpenters and joiners, 90 000 builders and building contractors, 88 000 maintenance fitters, and 72 000 motor mechanics estimated to receive daily exposures above this threshold.

OCCUPATIONAL EXPOSURE TO HTV AND CURRENT EMPLOYMENT

Among men who reported having a paid job in the past week, 913 (16.6%) described themselves as being self employed. Men in this category were more likely to have used hand held powered tools in the past week than all men in work (47.8% v31.5%), and more commonly belonged to one of the occupations with high exposure: thus, 92% of farm owners and managers were self employed, as were 79% of builders and building contractors, 55% of plumbers and heating and ventilating engineers, 40% of carpenters and joiners, 39% of bricklayers and masons, and 28% of gardeners and groundsmen.

Discussion

This survey indicates that in Great Britain over a given week about 20% (4.2 million) of all men of working age and 3% (0.7 million) of all women of working age are exposed to HTV at work. It further suggests, based on self reports of exposure times, that the daily A(8) in 5.8% of men of working age (1.2 million), and 0.2% of women of working age (44 000) exceeds the HSE suggested action level of 2.8 ms^-2 rms The occupations and industries contributing most to estimates of exposure prevalence nationally have been identified, as well as those in which high reported levels of exposure were most commonly found; and the vibratory tools most often giving rise to exposure have been identified. The data also provide an estimate of the lifetime prevalence of occupational exposures—42% of men and 8% of women from the sample having worked at some stage in a job that entailed exposure to HTV for more than 1 hour a week.

POTENTIAL BIASES AND LIMITATIONS

In assessing the implications of these findings for policy, it is important to take account of several potential sources of bias. These include the representativeness of the sample, the accuracy of the information supplied in response to the questionnaire, the reliability of the a_hw values assigned to tool categories, and the validity of extrapolation from the sample to the national population.

Outside the armed forces, nearly everyone in Britain registers with a family doctor, making the age-sex registers of general practices comprehensive and representative enumeration lists. Around a thousand questionnaires from the survey were returned because patients had moved or died before posting, but the factors that influence non-registration and turnover are unlikely to have importantly biased estimates of contemporary occupational exposure to HTV. Representativeness was further assured by the wide geographical base chosen for sampling and by the process of standardisation, in which census information was used to compensate for oversampling or undersampling and differences in response rate between occupations.

The response rate was higher in older subjects, but a comparison of older and younger responders (those over and under 40 years) among commonly exposed occupations did not indicate that age was systematically related to the frequency of exposure. Also, subjects with exposure could have been more highly motivated to participate, a bias that could lead to overestimation of exposures within occupations. This possibility was explored by comparing the prevalence of exposure in people who responded at the first invitation and those who required a reminder, on the assumption that a greater interest in the survey would be reflected in a higher frequency of exposure in the early responder group than the late responder group: in the event, only minor differences were apparent between the two groups.

Information on current exposure to HTV came principally from closed questions concerning a predefined list of sources, but also from answers to an open question about other sources, which were evaluated by a panel of vibration specialists. Errors may have arisen in answers to the closed questions if respondents did not recognise a tool name or confused it with another that they had used; and in the open response section if they described tools in an ambiguous or misleading way. Such errors could lead to an overestimate or underestimate of the frequency of exposure. Another possibility is that recreational activities were sometimes reported as if they were occupational. This risk might be expected to be higher for some tools, such as hand guided mowers, than for others.

To test these concerns the face validity of common exposures was examined by occupation. This was done for users of hammer drills, hand held portable grinders and jig saws (the three commonest sources of exposure), and for users of hand guided mowers, hedge trimmers, and strimmers (three exposures that may arise recreationally as well as occupationally). Exposures in the hammer drills, hand held portable grinders, and jig saws group largely arose from occupations where use was expected or considered plausible, lending support to their reliability. For example, the leading occupations reporting use of the hammer drill came from the construction industry, whereas hand held portable grinders were used most often by metal working and maintenance fitters, and jig saws by carpenters and joiners. However, only 19%–35% of exposures to hand guided mowers, hedge trimmers, and strimmers among men were specifically in gardeners, so a possible bias due to the false reporting of leisure time exposures cannot be discounted in the case of gardening equipment.

An independent test of the accuracy in reporting sources of exposure was conducted by visiting workplaces and comparing reports of exposure with direct observations. The findings, which are reported elsewhere,21 provide empirical evidence that self reports are generally accurate as to whether exposure does or does not occur, although workers do sometimes confuse the names of vibratory tools or omit to report some sources of exposure.

Quantitative estimates of exposure are subject to several potential sources of error and bias—in assigning a_hw, in self assessment of durations of exposure, and in the method of dose calculation. The task of assigning a_hw was based on a qualitative appraisal of data that were available from both published and unpublished reports (see ). Generally, several sources of information were consulted for each source of exposure, but a different choice could have been made. One value was ascribed to each family of tools, but in practice variations of up to fivefold have been found in some published series,22 reflecting differences in the model type, eras of design, and source of manufacture; and differences in the task and fitted accessories. Single values represent a convenient simplification. However, to test the appropriateness of the choices made, we separately made field measurements on sources of exposure that were found to be common in this survey, and have reported our findings.23 In general, the median measured values were close to the values listed in the and chosen for analysis.

In a few cases the lack of dependable data prevented any value being assigned, and missing exposure details required the calculation of minimum rather than absolute A(8) values in many subjects. Personal daily exposure to vibration was calculated according to a widely promulgated method of dose summation, but there is some uncertainty about the best method.24-26 Reliance was also placed on self estimates of the duration of daily exposures—an approach traditionally assumed to be adequate, but seldom tested. Elsewhere we report a study showing that workers tend systematically to overreport their exposure durations (by around 2.5-fold),21 but, because of the assumed square root relation between A(8) rms and daily exposure time according to ISO 5349, 198619 this would translate into an exaggeration of calculated A(8) of around 60%, and is comparatively small in relation to the variation in magnitude of vibration between apparently similar sources.

It should be emphasised that the various limitations of quantitative assessment of exposure are not particular to this investigation, but general to the field of inquiry. The epidemiological studies that underlie suggested action levels have been based on a similar means of exposure assessment.

The questionnaire collected information during two calendar periods. The responses included people who were employed but absent from work, and encompassed various typical and less representative exposure patterns. Short cycle variations and periods of atypical exposure might be expected to even out in a large sample, and do in any case contribute to the broad dynamic picture of national exposure. But estimates in occupations where seasonal differences in vibratory tool use exist could convey a misleading impression of the extent of annual exposure. In practice however, these effects are likely to be small: few men and women reported that their exposure in the past week was unusual, and a cross comparison of responses within occupations identified only minor differences between summer time and winter time respondents in the proportions reporting exposures.

With occupational (or industrial) populations to derive national estimates of prevalence of exposure, two simplifying assumptions were made—that there was little heterogeneity between industries within an occupation (and vice versa), and that any geographical variation in exposure within occupations and industries was adequately represented in the broad geographical sampling base, or unimportant as a source of error. Given the small numbers of exposed people in the various occupation-industry combinations (no more than 71 in any occupation-industry pair) the validity of these assumptions cannot readily be tested. However, it seems unlikely that the overall estimates of exposure, and particularly those for larger occupational groups, will have been seriously biased as we found the leading occupations with exposure to be similar countrywide in an analysis that divided Britain into six separate regions.

The aggregated estimates of exposure, by virtue of their large sample size, have narrow 95% CIs, but in other cases population estimates were derived from small sample sizes within an occupation or industry. This represents a limit on precision and results in relatively wide 95% CIs for some occupations.

COMPARISON WITH OTHER STUDIES

This study indicates that exposure to HTV is far more common than estimated previously.12-14 The discrepancy seems to have arisen mainly because sampling in the earlier studies was selective as to industry and source of exposure. No data were collected from at least 28 of the 38 industries in the SIC 92 code groups where we found exposure; within an industry, often only restricted aspects of exposure were assessed—for example, only chain saws in agriculture and forestry—and many of the sources of exposure that we included in our questionnaire were omitted or incompletely surveyed (including floor polishers, nut runners, impact screwdrivers, jig saws, circular saws, hand guided mowers, hand held hedge trimmers, brush saws, barking machines, stump grinders, tampers, scabblers, hammer drills, needle guns, nibbling machines, motorcycle handlebars, and nailing and stapling guns). The emphasis previously seems to have been on sources reported to cause health problems and exposures that were explicitly reportable and compensatable under statutory rules, but other sources are now considered to be potentially hazardous, and many of these now are in the Industrial Injuries Advisory Committee's revised list of compensable exposures.27 These extra sources may account for an estimated additional 1.5 million users nationally. Finally, differences also arise because we sampled more completely from small businesses and the self employed, where the proportion of exposed workers was higher.

While our survey was underway, an independent estimate of community exposure was reported in the Office of National Statistics omnibus survey.28 A questionnaire on self reported working conditions was administered to a sample of adults in private British households, selected from a national post code address file. Face to face interviews were conducted in August and October 1995. A question on exposure to HTV was posed as part of a much wider survey of economic activity that had no particular focus on use of vibratory tools, blue collar work or work related illness. Exposure was defined as “the use of power tools which transmitted vibration to the hands”, but no predefined checklist of sources of exposure was used, as in our survey. None the less a similar estimate of exposure was obtained. Among 2230 employed people, 27% of men (95%CI 24% to 30%) and 7% of women (95%CI 5% to 9%) reported using hand held powered tools in their job. A more detailed comparison between our survey and the self reported working conditions survey shows striking similarities in the respective estimates of exposure frequency by occupation (table 4). Hence, estimates that were collected independently, during different calendar periods, by different methods, and in a different context, show a high degree of agreement. This provides reassurance about the validity of exposure information in this study and suggests that response and sampling biases have not caused important errors in the numbers estimated to be exposed.

IMPLICATIONS

In the United Kingdom, HSE recommends a programme of preventive measures and health surveillance at an A(8) of 2.8ms^-2rms20 (An annexe to British Standard 6842, 1987 predicts that 10% of the workforce might be expected to develop finger blanching over an 8 year period at this threshold29). Elsewhere, it is advised, because of the uncertainty about the exposure-response relation, that it would be prudent to conduct health surveillance on all regularly exposed workers.30 Such exposures are surprisingly prevalent, and preventive measures and health surveillance may be warranted for many men in Britain. There could, however, be major practical difficulties in providing appropriate health surveillance, especially for the self employed and in small businesses. This emphasises the strategic need to provide technical and engineering solutions that limit exposures at source.

The distribution of exposures to HTV is influenced by Britain's industrial base. Hence, in certain occupations and industries we would expect the pattern to differ from other countries (for example, mining is no longer an important national source of exposure, by contrast with some other nations). However, it seems likely, at least in the industrialised countries, that similar patterns of exposure will prevail in common occupations.

UNDER-RESEARCHED TOOLS

Many sources of HTV that are common in Britain have been overlooked in previous surveys of health and exposure to vibration. These may confer health risks, but the evidence on this is limited. At present it rests almost entirely on the measured vibratory characteristics of a selection of these tools. This has led to proposals for some sources to be included in parts 2–17 of ISO 8662 defining vibration type tests (to demonstrate tool safety standards and relations with safety labelling)31; and to such sources being included in the Industrial Injuries Advisory Committee's recent list of potentially compensable exposures.27 By contrast, there is little direct evidence of adverse health effects arising from some of these sources. A search of the EMBASE (Excerpta Medica) database for 1980–98 and Medline (National Library of Medicine, USA) database for 1966–98 failed to identify any articles in which keywords—such as mower, circular saw, jig saw, impact screwdriver, nailing gun, and stapling gun—were used in conjunction with search criteria for Raynaud's phenomenon and VWF; and very few published studies on health risks from these sources or from metal drills and hammer drills could be discovered in the Institute of Sound and Vibration Research's international library of information. Such common sources of exposure seem largely to have been overlooked in health inquiries, but are worthy of formal study, and some exploratory health analyses will be reported separately.

Conclusions

We conclude that occupational exposure to HTV is surprisingly prevalent in Great Britain, and that average self estimated A(8) exposures in a 1 week period exceed the HSE's suggested action level of 2.8 ms^-2 in >1 000 000 British men. The scale of exposure, including exposures in small businesses and self employed people, suggests that higher priority should be given to technical and engineering controls. The data provide information on the occupations, industries, and sources which should be the priorities for corrective action, and highlight common exposures in Britain that warrant further investigation.