Health and Safety
Executive / Commission
Toxic Substances Bulletin
Dear Reader
Autumn is almost upon us - and so is this year's European Week for Safety and Health at Work! In the May issue of TSB, I explained that the theme this year is Dangerous Substances, providing an excellent opportunity to raise awareness of the importance of assessing and controlling risks from harmful chemicals. The Week runs from 13-18 October 2003 and it is still not too late to plan an initiative. For your free action pack phone 0800 085 0050, and let us know what you have done - you may even win an award! Overall winners will be presented with their award at a prestigious event to be held in central London early in 2004. Government is supporting this event at the highest level. Des Browne, Minister for Health and Safety, will be launching the week, when the expanded electronic COSHH essentials website will be available, containing advice for a much wider range of businesses, including hairdressers, foundries and motor vehicle repairers. Log onto http://www.coshh-essentials.org.uk/ in early October and let us know what you think of the new site.
This month's edition of TSB continues our exploration of epidemiology by posing the question: "How do epidemiologists establish whether there are links between exposure to a substance and disease?" - an important element in deciding which substances are the most harmful to health. An article on HSE's developing programme on occupational respiratory disease continues this theme. It sets out how HSE will start by looking at epidemiological studies that have measured the prevalence of respiratory disease in occupational groups - and then goes on to discuss what else is planned to take forward the strategy, pointing out that views from TSB readers would be very welcome! Please have a look at the article and let us know if you can help.
In the last issue, I announced that the Health and Safety Commission had endorsed a chemicals strategy developed by HSE. In addition to the Respiratory Disease Programme mentioned above, a Communications and Education Programme is also being developed. We are interested to hear from any readers who have ideas on how HSE could communicate the risks from harmful chemicals more effectively, particularly to small and microfirms. As part of the programme, HSE will be setting up focus groups of small businesses to examine this question.
Finally, I strongly recommend that you have a look at the article on the proposed new Occupational Exposure Limit (OEL) Framework. This article previews a Consultation Document (CD) setting out proposals for the new OEL Framework which will aim to make it easier for employers to control the risks from harmful chemicals. The CD will be published in October and we want as many businesses as possible to respond.
Many thanks to those of you who sent in comments on the last issue of TSB - we really do welcome your views.
In the June issue of TSB, HSE invited readers to participate in this year's European Week for Safety and Health at Work on dangerous substances. The Week runs from 13-18 October 2003, though campaigns can be run during any week in October. This article highlights some of HSE's own plans to be launched in the Week.
HSE will be adding around 70 new guidance sheets to the e-COSHH essentials website - HSE's free Internet advice and risk assessment tool for business, particularly smaller businesses. From October, there will be new service and retail guidance sheets as follows:
There will also be other guidance sheets on:
http://www.coshh-essentials.org.uk/ contains a wealth of information: good practice, useful advice, employee checklists and background information. From October, guidance sheets can be printed off without necessarily having to complete an on-line risk assessment.
HSE will also be promoting the requirements of the duty to manage asbestos through its network of some 2000+ Partners, and the opportunity will be taken during the Week to launch a number of associated products; these will include updated versions of the Presenter's Pack and the video How are you today? In addition, there will be new 'Good Practice Guidance' aimed at providing further help to dutyholders through a range of case studies. Visit the Asbestos Campaign webpage at http://www.hse.gov.uk/campaigns/asbestos to learn more about what is available to help you to comply with the new asbestos regulations.
There will be a conference on occupational asthma on 14-15 October at Regent's College, Regent's Park, London. The Professional Organisation in Occupational Safety and Health is organising the first day, which is aimed at small businesses and those who wish to learn more about the subject. The British Occupational Hygiene Society (BOHS) is organising the second day which is for those who already have a more in-depth knowledge. Registration details (cost £45 for the first day and £90 for the second) and further information can be obtained from Krissy Robinson, BOHS, Suite 2, Georgian House, Gt Northern Road, Derby, DE1 1LT Tel: 01332 298101 e-mail: krissy@bohs.org
HSE is planning to launch a latex allergy tool kit at the conference. This will take the form of an Internet-based toolkit for training and education purposes, as well as a paper version. Announcements about the tool kit will be made at the IOSH Healthcare Specialist Group Annual Conference during the previous week.
Finally, on 17 October the Advisory Committee on Toxic Substances (ACTS) will hold its first open meeting to coincide with the Week . The theme of the Week is highly relevant to the work of ACTS, whose aim is to stop people being made ill from exposure to substances at work.
For further information, please refer to HSE's European Week website: http://www.hse.gov.uk/campaigns/euroweek/; action packs, posters, stickers, fact sheets are available from 0800 085 0050; HSE InfoLine 0845 345 0055 can provide copies of the Newsletter.
In the last article we said we would look at issues of interpretation in observational studies. In themselves, statistical observations can only establish associations between the incidence of disease and the exposures and individual characteristics that may be involved in its causation. But associations are only really of interest to the extent that they point to causes. So, how can we decide which associations are causal and which are not? Given that some association between an exposure and a disease has been established as statistically significant in at least one set of observations, what characteristics of the observational evidence strengthen the arguments for inferring causation?
The classic discussion of this question was set out by Sir Austin Bradford Hill in the 1965 Cochrane memorial lecture. Bradford Hill described nine characteristics of observational data which tend to support a causal interpretation. The word 'tend' is important here. We are not dealing with hard and fast rules or clear-cut criteria. Rather, these characteristics are an attempt to systematise the kinds of questions we need to ask ourselves in assessing observational data. Essentially, we should be testing whether a causal explanation of the data we are presented with is the best available, or whether other explanations - generally some combination of chance, bias or confounding - are more likely.
The nine characteristics identified by Bradford Hill are: strength of association; consistency; specificity; relationship in time; biological gradient; biological plausibility; coherence; experiment; and analogy. We will discuss these in turn.
The stronger the observed association, the more seriously one must consider the possibility that it is causal. Small differences in risk between groups can be much more readily ascribed to chance or confounding than strong associations. For example, to explain the tenfold differences in risk of cancer between smokers and non-smokers (and still more, the 30-fold excess risk of heavy smokers) by an appeal to confounding would mean postulating the existence of some third influence that is both strongly related to the fact of smoking and is a powerful cause of lung cancer, thus creating a non-causal association between smoking and lung cancer. Some critics of the earliest evidence of the link between lung cancer and smoking did in fact advance such arguments. Once the scale of the lung cancer risk from smoking was established (and in the absence of any plausible third factor of the kind needed) these arguments could not be sustained.
The chance characteristics or special circumstances of a particular study population could be taken to explain an isolated observation, but if a particular association is seen in the range of different populations and in the range of different study types it will be more difficult to dismiss in this way, and the possibility that the relationship is causal becomes much stronger.
Many causes of disease are specific in their effects: beta-naphthylamine only causes bladder cancer, asbestos only causes respiratory cancers, vinyl chloride monomer only causes angiosarcoma. Where a single exposure does have multiple outcomes (for example, smoking or exposure to ionising radiations) there is always a characteristic pattern of response, and usually a 'leading' response, such as lung cancer for smoking or leukaemia for ionising radiations. Associations which are not specific in this sense, where the association seems to be with a wide range of outcomes, rather than being focused on a specific cause or narrow range of causes, raise a suspicion of bias or measurement error as a plausible alternative to a causal interpretation.
Clearly, the cause of a disease must precede the onset of the disease in time. Clarity on this criterion may not always be obvious, particularly with chronic disease, and with retrospective study methods. For example, associations of behavioural characteristics such as diet with chronic disease states (in retrospective study) may be due to the effects of disease (perhaps in its early, undetected, stages) on diet, rather than on a causal relationship leading from diet to disease. Where a putative cause can be clearly established as prior in time to the onset of disease, a causal interpretation is strengthened (or, at least, not ruled out).
If a particular exposure causes some disease, then higher exposure levels should cause more cases of disease, and lower levels fewer cases. Where this can be demonstrated, causal interpretation is strengthened. Again, this is not an infallible touchstone. It assumes that we know (and have available) the relevant measure of exposure. If we construct our exposure groups in terms of cumulative exposure (for example, the total number of asbestos fibres inhaled), and the relevant exposure measure involves the intensity of exposure (for example the number of hours exposed above some minimum airborne concentration), then an exposure-response relationship may not be apparent simply because we have not measured exposure in the appropriate way. The elusiveness of disease associations with non-ionising radiations may be due to our ignorance of the appropriate measure of exposure, which means that the appropriate exposure-response comparisons cannot be constructed. Of course, it may also be elusive because exposure to non-ionising radiations does not lead to brain cancer or leukaemia!
This characteristic essentially measures the possibly causal relationship indicated by the observed association against what is known of the relevant biology. If the exposure involved is known to affect biological systems in ways which could reasonably lead to the observed disease outcome, we can be more confident of a causal interpretation. Of course, what is biologically plausible depends on our current biological knowledge. Where epidemiological observation runs ahead of biological knowledge, genuinely causal associations may be observed for which - on current knowledge - biological plausibility is lacking. This would have been the case for many classic epidemiological findings, for example Perceval Potts's 18th-century observation of scrotal cancer in chimney sweeps, or Jon Snow's identification of a contaminated water supply in Soho as the source of cholera infection in the 1850s.
Where biological plausibility looks for active support from current biological knowledge, coherence asks simply that the observed association should not be in contradiction to that knowledge, or with broad trends in the exposure in question and in the disease's natural history. For example, when the issue first arose in the 1950s, a causal link between cigarette smoking and lung cancer was coherent with the known trends in lung cancer incidence and in cigarette consumption (including the sex differences in both these factors).
Sometimes an experimental intervention may be possible (and ethical). Where there is a suspicion that some exposure is causing disease, the exposure may be eliminated. If this intervention is followed by a fall in the incidence of disease, the conclusion that the exposure was indeed causally connected with the disease is strongly reinforced.
A rather weaker form of the biological plausibility argument looks to analogy. We know that the drug thalidomide and the disease rubella can be causes of congenital malformations. The observation of an association between this outcome and exposure in pregnancy to some other drug or some other viral disease can more readily be argued as causal than if these analogous examples were not established.
It should be clear from the discussion above that none of these nine viewpoints can establish a conclusion of causation beyond doubt. Equally, none can be regarded as indispensable. Taken together, they help answer what Bradford Hill describes as the fundamental question: "Is there any other way of explaining the set of facts before us? Is there any other answer which is more likely than cause and effect?"
The interpretation of observational data is essentially not a matter of technical statistics, but the application of common sense and elementary logic to statistical findings in the context of the relevant scientific knowledge.
Technical statistical considerations can help guard against taking too seriously observations that may well be due to chance, and thereby direct attention to those that must be taken seriously. But the later stages of inference are dominated by non-statistical considerations. Epidemiology is an essentially multidisciplinary pursuit.
Occupational respiratory disease has been identified as a priority in HSE/C's developing Chemicals Strategy. As a result, an HSE working group was set up in May this year to develop a strategy to combat this disease area. The Strategy needs to contribute to the government target of reducing the incidence of occupational ill health by 20% by the year 2010.
Occupational respiratory diseases encompass a broad spectrum of ill-health conditions, including occupational asthma. Given the potentially serious health and socio-economic consequences of occupational asthma, this disease area will constitute a major part of the Respiratory Strategy. The previous issue of TSB included an article explaining what is currently known of the incidence of occupational asthma in the UK, and also outlined a programme of activities targeting the top eight causes. Therefore, the rest of this article will focus on the 'non-asthma' types of occupational respiratory disease.
So far, the HSE working group on the Respiratory Strategy has agreed that the starting-point for the Strategy is to characterise the 'evidence-base'. What is the scale of occupational respiratory disease in the UK? Who are the most 'at-risk' working populations? Obviously, activities arising from the strategy need to be targeted at genuine problem areas.
Characterisation of the evidence-base for the 'non-asthma' forms of occupational respiratory disease is not easy. The respiratory group has agreed that HSE's published annual statistics on occupational ill health may considerably underestimate the true scale of respiratory diseases caused by chemicals in the workplace.
This problem can be best illustrated with the example of chronic obstructive pulmonary disease (COPD).
COPD may result from either chronic bronchitis or emphysema. Symptoms include cough, sputum production and difficulty in breathing. COPD impairs the quality of life and can lead to premature death. The main cause is cigarette-smoking. However, occupational exposure to workplace substances such as noxious gases, fumes and dusts can also cause or contribute to the development of COPD. COPD can take many years to develop, and symptoms may be attributed to growing older or lack of fitness. The occupational link may not be recognised.
There are published statistics on the number of coal miners in the UK who receive compensation for COPD. These may be viewed on HSE's website: http://www.hse.gov.uk/statistics/index.htm
However, COPD is not a disease for which compensation is normally available, except for coal miners, and there is little or no specific information on the incidence of COPD in other industry sectors.
However, we do know that in 1996-97, COPD cost the NHS £818 million. The World Health Organisation estimates that COPD is the fourth leading cause of death worldwide. The American Thoracic Society estimates that around 15% of all cases of COPD have an occupational causation.
These observations suggest that COPD merits inclusion in the Respiratory Strategy. However, further detective work is needed to clarify the evidence before decisions can be made on where best to target activities, and what those activities should be. Given the long latency for the development of COPD it will also be a particularly difficult challenge to evaluate the effectiveness of any strategy to combat this type of occupational respiratory disease.
We will need to pursue different lines of evidence. From the published literature we can identify epidemiological studies that have measured the prevalence of respiratory disease in occupational groups. We can also use existing knowledge and experience in HSE. For example, these sources indicate evidence for high dust exposures in construction workers. The dust often contains crystalline silica, a particular concern for lung disease. Construction workers could be a difficult group to target as they often move in and out of jobs. The fact that they work outdoors and engage in heavy manual labour means that they probably have better-than-average lung function, and detecting the early development of COPD could be particularly difficult.
HSE's respiratory working group has also identified electroplating as a possible target, as most electroplating workshops employ few people and use old equipment, with exposures to a number of toxic respiratory hazards such as chrome acids that could contribute to COPD.
As you can see, HSE faces a challenge in developing this Strategy. However, we are committed to taking this work forward and would welcome information and comments from TSB readers. Further TSB articles will keep you informed of progress.
In August, HSC considered a proposal by the ACTS OEL Sub-group to introduce a new OEL framework and agreed to the publication of a Consultative Document (CD).
This follows the discussion document published last year, which set out why a new approach was needed and sought stakeholders' views on three options for a new system. The majority of respondents agreed with the concerns about the present system and supported the concept of a single-limit system linked to good-practice advice.
Research shows that many small firms have limited understanding of OELs and that they need practical advice on what they have to do to control the chemicals they are using. The proposals meet this need by linking OELs directly to advice on how to control the chemical.
The present COSHH Regulations use two types of OEL, which impose different requirements on employers. However, only a very small percentage of firms understand the difference between the two types and the proposal is to replace them with a single type of OEL, called a Workplace Exposure Limit (WEL). This will impose a single duty on employers - the amount of a chemical in their workplace air must not exceed the WEL.
Part of the proposals is that OELs in the present system which are not soundly based should not be transferred to the new system. HSE will provide good practice advice for these which, together with a new approach to adequate control, will mean there will be no overall reduction in standards.
The definition of adequate control would be broadened to include the duty to apply the principles of good practice for the control of exposure to substances hazardous to health (see box below), as well as complying with any relevant OEL. For a substance which carries the risk phrase R45, R46 or R49 (all three indicate the possibility of a substance causing cancer) or for a substance or process which is listed in Schedule 1, the level of exposure is reduced so far as is reasonably practicable.
In most cases firms will not need to worry about understanding the principles of good practice nor about taking air samples to check they are below the WEL. For each substance that has a WEL, HSE will produce practical advice that will tell employers what they have to do. For chemicals that can cause cancer, HSE's advice will explain what As Low As is Reasonably Practical (ALARP) means in practice. Thus employers will not simply be left with a number, as in the present OEL system, but will have information on what they have to do to comply with the law.
The practical advice will be available on the Internet. We propose to link together information on WELs, COSHH essentials - which gives practical advice - and more in-depth information on individual chemicals and the COSHH Regulations. The aim is that employers who just want practical advice on what they have to do will be able to get it, but specialists who want more technical detail will get it through hypertext links. The list of WELs and associated practical advice would also be available in hard copy.
TSB readers are encouraged to respond to the CD, which is expected to be published in October 2003, with 31 December being the closing date for comments. An electronic version of the CD will be available on the HSE website at http://www.hse.gov.uk/consult/live.htm. Alternatively, hard copies are available from HSE Books at http://books.hse.gov.uk
In TSB49 we reported that the EC published its White Paper 'Strategy for a future chemicals policy' in February 2001. The paper proposed a new scheme called REACH (Registration, Evaluation, Authorisation of Chemicals) to manage the supply, manufacture, importation, marketing, classification and labelling of chemicals within Europe.
The scheme would require all chemicals manufactured in quantities of greater than one tonne to be registered, those manufactured at greater than 100 tonnes to be evaluated and certain substances of high concern (carcinogens, mutagens, substances toxic to reproduction) (CMRs) and persistent organic pollutants (POPs) to be authorised.
The EC had originally planned to table draft legislative proposals by the end of 2001; however, this was never likely to be feasible and was delayed until May this year.
On 7 May 2003 the Directorates General (DGs) Environment and Enterprise jointly published a draft proposal for an EU Regulation to introduce the new REACH scheme. At the same time, they launched an eight-week Internet consultation on the 'workability' of the system. The draft proposal is based on the White Paper and takes into account the results of EC working-groups (Member States, NGOs and business), Environment Council conclusions and European Parliament recommendations.
The main themes of the new REACH system are:
: This broad provision requires all those manufacturing, importing or using substances to carry out a chemical safety assessment and take appropriate risk-reduction measures to address any risks identified. The requirement applies regardless of the quantity of the substance being manufactured or used.
: A proposal that industry collects, collates and submits data on the hazardous properties of all substances manufactured or imported into the EU in quantities above one tonne per year. In addition, industry should prepare risk assessments and provide safety information to downstream users.
: There are two types of evaluation proposed:
: A proposal that industry will need to gain authorisations for the use of substances considered to be of very high concern. These are substances that are identified as carcinogenic, mutagenic or toxic to reproduction (CMRs); persistent, bioaccumulative and toxic substances (PBTs); substances that are very persistent and very bioaccumulative (vPvBs); and substances demonstrated to be of equivalent concern, such as endocrine disruptors.
: A proposal that the provisions enable risk-reduction measures to be introduced across the EC where this is shown to be necessary. Member States or the Commission prepare proposals for restrictions.
: A proposal for the creation of an agency for managing the technical and administrative aspects of the REACH system at Community level.
The full proposal and explanatory information can be obtained from: http://europa.eu.int/comm/enterprise/chemicals
We understand that the next stage, following consideration of consultation comments by DG Environment and DG Enterprise, will be consultation with other DGs. The EC anticipates adoption of the proposal in early autumn 2003.
The Department of Food and Rural Affairs (DEFRA) maintain the UK policy lead on the new Chemical Strategy (NECS) and have co-ordinated a UK response to the EC on the workability of the draft proposals:
www.defra.gov.uk/environment/quality/chemicals/reach/index.htm
At its meeting in July 2003, the Advisory Committee on Toxic Substances (ACTS) agreed a biological monitoring guidance value (BMGV) for polycyclic aromatic hydrocarbons (PAHs).
BMGVs are published by HSE in Table 3 of EH40 Occupational Exposure Limits. The 2003 Supplement to EH40 contains new BMGVs for cyclohexanone, glycerol trinitrate and xylene (o-, m-, p- or mixed isomers) but, at the time of publication, it was not possible to include the BMGV for PAHs.
Biological monitoring is the measurement and assessment of hazardous substances or their metabolites in tissues, secretions, excreta or expired air, or any combination of these, in exposed workers. A BMGV is set where it is likely to be of practical value, where suitable monitoring methods exist and where there are sufficient data available. There are two types of BMGV: a Health Guidance Value and a Benchmark Guidance Value.
ACTS has agreed to a Benchmark Guidance Value for PAHs. This is:
4 µmol 1-hydroxypyrene/ mol creatinine in urine measured at the end of a shift.A benchmark guidance value is a hygiene-based guidance value, set at a level where 90% of available validated data, provided by a cross-sectional study of workplaces with good occupational hygiene practices, are below the value. It can therefore be achieved by the great majority of industrial premises which employ good workplace practices.
The BMGV is primarily aimed at those industries that rely on respiratory protective equipment (RPE) to restrict worker exposure. Airborne monitoring in these industries would not therefore necessarily provide an accurate indication of personal exposure. In contrast, biological monitoring would accurately measure the exposures received by individual workers and would enable an employer to verify that the RPE being used was providing an adequate level of protection.
The BMGV is just part of a parcel of measures envisaged by HSE to control exposures of workers to PAHs. Although there are no plans to set an occupational exposure limit, HSE will be incorporating PAHs into its overall strategy for dealing with workplace carcinogens. It will also be looking closely at the need to provide sector-based guidance for those areas of industry where the risk of exposure is considered to be the greatest or the least well controlled.
The latest supplement to update EH40 - HSE's list of occupational exposure limits (OELs) - is available on the Internet.
It includes the new and amended exposure limits approved by HSC in April, and should therefore be read in conjunction with EH40/2002 Occupational exposure limits 2002, which remains current for another year.
Employers have a duty under the Control of Substances Hazardous to Health Regulations 2002 (COSHH) to control workplace exposures to hazardous substances. EH40 contains technical details of any legal changes to OESs and MELs.
There will not be a new version of EH40 this year, because HSC is currently developing an entirely new occupational exposure limit framework. An HSC consultative document (CD) explaining these proposals will be published in the autumn.
Paper copies of EH40/2002 Occupational exposure limits supplement 2003, ISBN 0 7176 2172 3, are available from HSE Books, price £3.00, PO Box 1999, Sudbury, Suffolk, CO10 2WA Tel: 01787 881165 Fax: 01787 313995 Website: http://books.hse.gov.uk
Copies of EH40/2002 (supplied together with the Supplement 2003) continue to be available from HSE Books, ISBN 0 7176 2083 2, price £10.50.
http://www.hse.gov.uk/coshh/index.htm
This web page deals with a host of issues related to the COSHH Regulations and health and safety. These include:
The main title page also contains a useful list of contents and related links to other items which readers may find interesting.
http://www.hse.gov.uk/skin/index.htm
This web page deals with skin disease at work. Many people in Great Britain suffer from skin disease caused by their
work, most of which has resulted from exposure to chemicals. It provides information about a range of free leaflets
on dermatitis at work. Advice for employers and employees.
Ms Carole Sullivan Editor
Miss Naseem Walji Assistant Editor
Mr D Kyle
Dr J Groves
Mrs E Ball
Mr S Campbell
Dr R Rawbone
Ms D Llewellyn