Health and Safety
Executive / Commission
Statistics
Injury and dangerous occurrence statistics given in this report for 1996/97–2006/07 were compiled from reports made to all enforcing authorities, namely HSE, local authorities (LAs) and the Office of Rail Regulation (ORR), under the Reporting of Injuries, Diseases and Dangerous Occurrences Regulations 1995 (RIDDOR 95). These Regulations came into effect on 1 April 1996. Since 1 April 2006, enforcement of health and safety on railways has been the responsibility of ORR, and they have provided relevant figures for the latest two years that fall within the scope of RIDDOR. Prior to this, enforcement was the responsibility of HSE’s Railways Inspectorate. More information on railway safety can be found on the ORR website.
Deaths of all employed people and members of the public arising from work activity are reportable to the relevant enforcing authority. There are three categories of reportable injury to workers defined under the regulations: fatal, major and over-3-day injury. Examples of major injuries include: fractures (except to fingers, thumbs or toes), amputations, dislocations (of shoulder, hip, knee, spine) and other injuries leading to resuscitation or 24 hour admittance to hospital. Over-3-day injuries include other injuries to workers that lead to their absence from work, or inability to do their usual job, for over three days. A non-fatal injury to a member of the public is reportable if it results in the injured person being taken from the site of the incident to hospital. Reporting requirements generally exclude incidents that occur to persons travelling in a vehicle, as part of their work, whilst on a public on a highway.
Selected incidents that have a high potential to cause death or serious injury are reportable under RIDDOR 95 as dangerous occurrences. A dangerous occurrence is reportable whether or not someone is injured.
In 2001/02, HSE introduced new guidelines to improve the quality of recording of kinds of accident and give more detail on equipment and material agents involved. As a result, there was a small change in the percentage share in each kind, predominantly for major and over-3-day injuries.
Injuries which are not reportable under RIDDOR 95 are: road traffic accidents involving people travelling in the course of their work, which are covered by road traffic legislation; accidents reportable under separate merchant shipping, civil aviation and air navigation legislation; and accidents to members of the armed forces. Although fatal injuries to the self-employed, arising out of accidents at premises which the deceased person either owned or occupied, are technically not reportable under RIDDOR, any such incidents are presented in the published figures.
While the enforcing authorities are informed about almost all relevant fatal workplace injuries, it is known that non-fatal injuries are substantially under-reported. Currently, it is estimated that just under half of all such injuries to employees are actually reported, with the self-employed reporting a much smaller proportion. These results are achieved by comparing reported non-fatal injuries (major as well as over-3-day), with results from the Labour Force Survey- Injuries (see below). Additionally, RIDDOR places a requirement on employers to report certain types of disease, although as these reports are
The Labour Force Survey (LFS) is a national survey of households living at private addresses in the UK – consisting currently of about 52 000 responding households each quarter. The survey is managed by the Office for National Statistics in Great Britain and by the Central Survey Unit of the Department of Finance and Personnel in Northern Ireland on behalf of the Department of Enterprise, Trade and Investment (DETINI).
The HSE commissions questions in the LFS, to gain a view of work-related illness and workplace injury based on individual’s perceptions. The HSE questions are included in two survey modules - ‘The Workplace Injury survey’ module and the ‘Self-reported Work-related Illness (SWI) survey’ module. For consistency the questions have, with few exceptions, been included in the winter quarter of the LFS (covering December to February). However, mostly due to an EU requirement under regulation, in May 2006 the LFS moved from seasonal quarters to calendar quarters, with the HSE survey module now included in quarter 1 (covering January to March). Whilst this small change in survey period may have potentially introduced a discontinuity in the data series, preliminary investigations suggest that this change in survey design has not affected the top level injury, ill health and working days lost data. Further work is planned to explore the possible effect of this small change in survey period. Because of the potential discontinuity, comparison of 2006/07 data with earlier years at more disaggregated levels should be made cautiously.
The workplace injury survey module was first included in the LFS in 1990, with a limited question set included annually since 1992/93. The LFS gives estimates of the levels of workplace injury by a range of demographic and employment-related variables and a broadly consistent time series is available from 1998/99.
The SWI survey module has been included in the LFS annually since 2003/04, and periodically prior to then (earliest results are from 1990, although only results prior to 2001/02 are not directly comparable with later time periods). This survey module provides an indication of the overall prevalence and incidence of work-related illness and its distribution by major disease groups and a range of demographic and employment-related variables. Responses obviously depend on laypeople’s perceptions of medical matters, but such perceptions are of interest and importance in their own right. However, they cannot be taken as an indicator of the “true” extent of work-related illness. People’s beliefs may be mistaken: they may ascribe the cause of illness to work when there is no such link; and may fail to recognise a link with working conditions when there is one e.g. because of the possible multifactorial nature of ill health or the delay between exposure and ill health (several decades in the case of cancer).
Both the workplace injury and the SWI survey modules have since 2003/04 (and periodically prior to then) also provided information about the number of working days lost due to workplace injury and work-related ill health. Estimates of working days lost for both workplace injuries and work-related ill health are expressed as full-day equivalent days to take account of the variation in daily hours worked (for example part-timers who work a short day or people who work particularly long hours). This information is available by a range of demographic and employment-related variables.
Since estimates derived from the LFS are based on a sample (rather than the full population), they are subject to a margin of error. The main factor which determines the width of an estimate’s margin is the number of sample cases an estimate is based on. In published reports and tables, the sampling errors are often expressed as 95% confidence intervals. Each of these represents a range of values which has a 95% chance of containing the true value in the absence of bias. Confidence intervals should be quoted in preference to the prevalence or incidence central estimate or rate whenever there are less than 30 sample cases. In order to reflect some of the variability in the days lost estimates (measure from person to person) as well as the sample numbers involved, confidence intervals should be quoted for days lost estimates and rates based on fewer than 40 cases taking time off. Estimates based on fewer than 20 sample cases are deemed unreliable and not published.
More detailed information about the survey design and methods used are given in Technical note Workplace injury and work-related illness survey modules of the Labour Force Survey: Background and methods. Published reports for SWI surveys from 1995 onwards can be accessed via the publications/release schedule.
In 1989 the first of several clinically-based reporting schemes for occupational disease was developed at the National Heart and Lung Institute in London. This scheme, the Surveillance of Work-related and Occupational Respiratory Disease (SWORD), relies on systematic, voluntary and confidential reporting of all new cases seen by consultant chest physicians. EPI-DERM, a scheme for surveillance of occupational skin disease by dermatologists, was begun in 1993, followed in 1996 by Surveillance of Infectious Disease At Work (SIDAW), by consultants in communicable disease control (CCDCs).
Three other schemes were started more recently, OSSA (Occupational Surveillance Scheme for Audiologists), MOSS (Musculoskeletal Occupational Surveillance Scheme) and SOSMI (Surveillance of Occupational Stress and Mental Illness). In MOSS and SOSMI, unlike the other specialist schemes, the physicians are advised to report cases either caused or made worse by work. In most of the schemes a large proportion of physicians in the relevant specialities participate systematically and voluntarily.
Occupational physicians reported to SWORD from its inception and to EPI-DERM from 1994. In 1996 the Occupational Physicians Reporting Activity (OPRA) was established as a separate scheme for all types of work-related disease, and in 1998 all seven schemes were brought together and run from the University of Manchester as constituents of the Occupational Disease Intelligence Network (ODIN). From April 2002 the network, still run from the University of Manchester, has been known as The Health and Occupation Reporting network (THOR).
In most of these schemes (SIDAW and OSSA being the exceptions), there is a sampling process whereby most participating doctors are asked to send in reports for one month in each year, and the numbers of cases that they report are multiplied by 12 in arriving at the estimated annual totals. (In MOSS, all participants were originally included on this 'sample' basis, but from January 2002 some rheumatologists reported cases every month throughout the year). To avoid any systematic seasonal biases the sampled doctors are randomly allocated their reporting month, and this allocation changes from year to year. Not all reporting doctors are sampled; some are so called 'core' reporters, who report cases every month throughout the year. Cases reported by them are included in the estimated annual totals without any scaling up. The estimated annual totals are generally based on smaller (often considerably smaller) numbers of actual reported cases, and are subject to random variation due to sampling error.
Many cases of work-related disease will fall outside the catchment of the THOR schemes, since many workers will not have access to an occupational physician at their place of work, and other specialists such as chest physicians, dermatologists, psychiatrists, etc will largely see only the more serious or difficult-to-resolve cases that are referred to them by other doctors. (They do, however, see patients over a wider age range than the occupational physicians, who almost exclusively see patients who have not yet retired). Therefore, figures from the THOR schemes should be regarded very much as minimal estimates of the true incidence of work-related disease.
Figures published by HSE relate to Great Britain only, although the THOR schemes do collect reports from doctors throughout the UK.
The incidence rates for THOR cases, per 100 000 workers in each occupation or industry, are calculated using denominators from the Labour Force Survey (LFS). The analyses by occupation now use the Standard Occupational Classification (SOC) 2000 rather than SOC 92, on which previous annual statistics have been based.
The fact that in many industries few, or even no, sufferers will have access to occupational physicians means that incidence rates based on or including OPRA reports cannot be used as a fair basis of comparisons between industries or occupations which have different degrees of coverage by such doctors. Comparisons between industries or occupations are best made by using rates based only on reports by 'disease specialists' (e.g. dermatologists, chest physicians, etc). Such specialists are accessible via the NHS to patients with all kinds of employer (including small businesses and the self-employed).
Any analysis of the THOR data currently presented on the HSE website in order to identify trends over time should be undertaken with caution. Those wishing to draw inferences regarding apparent changes in reported numbers of cases should be aware that there can be several potential explanations for differences between one year and the next. For example, participation by specialist doctors in the schemes is voluntary and so the number of reporters may vary with time. In addition, there is evidence that some reporters may be less inclined to report as time goes on.
A more sophisticated longer term statistical analysis is being undertaken to take account of the kinds of factors identified above which complicate the measurement of trends. This has involved the use of a multi-level statistical model (see report on ‘Trends in ill health data from THOR [417KB] 
’). Within this model data is analysed in a process which effectively calculates the trends over time in the level of reporting by individual reporters and then summates these individual trends as part of the process of calculating the overall trend. This modeling approach takes full account of changes in the number of reporters over time. It also enables some allowance to be made for the fact that individual reporters may vary in factors such as the density of cases they see and the stringency of the criteria which they apply when deciding whether particular cases are work-related.
For more information on THOR, please visit the University of Manchester website.
THOR GP is a UK-wide surveillance scheme covering work-related ill health that was initiated in 2005. Participating General Practitioners (GPs) report anonymised information about newly diagnosed cases to a multidisciplinary team at Manchester University. Details are recorded on a central database and the collated information is starting to provide a powerful resource for investigating the increased risks of particular types of ill health in relation to occupations, industries and causal agents or work activities.
The pool of voluntary reporters currently participating in this project consists of around 270 GPs already trained at a postgraduate level in Occupational Medicine by the Centre for Occupational and Environmental Health (COEH) of the University of Manchester. The specific course is offered by distance learning and COEH is one of only a very few sites in the UK that offers this type of specialist GP training. Consequently volunteer GPs reporters practice in areas widely distributed across the UK. The GPs reporters are instructed to make their decisions as to whether a new case should be identified as being attributable to work on the balance of probabilities (i.e. whether it is more likely than not). Reports are collected via web forms each month. When reporting a case the GPs are asked to classify it into a broad disease category and to provide information on age, gender, job, industry, type of exposure, and absence from work.
An audit of the accuracy of the recording of sickness absence within the surveillance scheme revealed that there was a considerable level of underreporting. This was primarily because some reporters tended to forget to arrange for updating of the database on occasions when they signed off patients for further sickness absence over-and-above the initial period of sickness absence.
Since the scheme only covers a small fraction of the total number of GPs, there are plans to obtain detailed demographic information about the patient make-up in the practices of the participating GPs. To this end, a breakdown is being obtained of the age, gender and postcodes of all the patients in each practice. Once this data has been collated together it will be possible to establish what proportion of the UK population is covered. Moreover, estimates will be made to determine how the employment mix of the sample of the population covered by THOR-GP compares to that of the country as a whole.
Given that the surveillance scheme is only just getting off the ground it is not possible to use the data from this sample of GPs as yet to make a reliable estimate of the overall incidence of work-related ill health.
For more information on THOR GP, please visit the University of Manchester website.
The Industrial Injuries Scheme, administered by the Department for Work and Pensions (DWP), compensates workers who have been disabled by a prescribed occupational disease. The self-employed are not covered by this scheme. Diseases are prescribed in connection with defined occupations or occupational conditions. They are only prescribed if an occupational cause is well established, and if terms of prescription can be framed in such a way that the majority of cases falling within the terms of prescription will be of genuine occupational origin.
Where there is a long delay (latency) between the cause of a disease and its appearance, it may be difficult both to identify and prove occupational causes, and to frame satisfactory terms of prescription. Even when this is done, the numbers of awards probably understate the disease's incidence, because individuals may be unaware of the possible occupational origin of their disease or the availability of compensation; this applies to shorter latency diseases as well.
Respiratory diseases are assessed by Special Medical Boards, and there are also separate arrangements for assessing occupational deafness. Figures for these diseases are published on a calendar year basis. Figures for other prescribed diseases (PDs) are published for years starting 1 October. For most diseases, benefit is payable if the extent of disability (from a single PD or from a number of PDs together) is assessed at 14% or more. However, figures are available for all newly assessed cases including those assessed at 1-13% disability. This so called '14% rule' was introduced for all claims lodged after 1 October 1986, for all diseases except pneumoconiosis, mesothelioma, and byssinosis (where benefit is still payable for lesser degrees of disability) and deafness (where the benefit threshold is 20% disability). For pneumoconiosis, byssinosis and mesothelioma, benefit continues to be paid and statistics are collected for all cases assessed at 1% or more disability. For deafness the available figures do not identify those assessed at less than 20% disability, who do not qualify for benefit.
In April 2002 a new method of collecting statistical information on claims and assessments was introduced by DWP, making the data more accurate. The apparent increase in some figures is believed to be largely due to this rather than reflecting a true rise in claims and assessments. It will also reflect the fact that, as of April 2002, the data include cases where the assessment results in "0%" disability being recorded, i.e. where the condition is accepted but where there is no loss of faculty. This category also includes cases where the percentage disability is missing (not coded at the time of publication) due to the provisional nature of the data.
Care needs to be taken in interpreting the annual totals for all prescribed diseases and their trend. Prescribed diseases are a mixture of different types of disease, and they do not represent the full spectrum of work-related illness. Individual components of the total are liable to be strongly affected by changes in prescription and factors affecting the take-up of claims (e.g. the contraction of traditional industries where the availability of compensation is well known, and the shift in employment to newer industries where it may be less well known). Much of the total is accounted for by lung diseases, vibration white finger, and deafness, and many such cases are a legacy of past working conditions which would be judged inadequate or in some cases illegal by today's standards.
The current set of data have been rounded to the nearest 5 cases, or to "-" if less than 5 cases. This has been done to maintain the anonymity of DWP customers.
For more information on the IIDB, please visit the website of the Industrial Injuries Advisory Council.
Mesothelioma and asbestosis death statistics for Great Britain are derived from the two registers of deaths due to asbestos related disease maintained by HSE:
The mesothelioma register comprises deaths where the cause of death on the death certificate mentioned the word 'mesothelioma'. For a substantial proportion of cases, it also contains information about whether the site of the mesothelioma was pleural (affecting the external lining of the lungs), peritoneal (affecting the external lining of the lower digestive tract) or both.
The asbestosis register comprises deaths where the cause of death on the death certificate mentioned the word 'asbestosis'. The information on the registers from the death records includes date of birth, date of death, sex, last occupation and postcode of residence at death.
Mesothelioma and asbestosis death records are supplied to HSE electronically by the Office for National Statistics (ONS) - for deaths in England and Wales - and the General Register Office for Scotland (GROS) - for Scottish deaths. Records are currently selected by ONS and GROS from their data collection systems via the mesothelioma cause of death code. ONS also search for strings 'meso', 'mesa' and 'asb' within the cause of death text descriptions. This combined approach helps to ensure that any deaths in England and Wales that may have been miscoded are identified. In addition, processing within HSE of asbestosis deaths is carried out before the mesotheliomas to enable identification of a small number of additional mesotheliomas via the string 'asb'. This is for situations where mesothelioma was spelt incorrectly on death certificates.
Some death certificates mention both asbestosis and mesothelioma. Such deaths are included on both registers in order to keep track of cases where both diseases were present. The mesothelioma and asbestosis mortality statistics are updated annually to include figures for the year two years behind the current year. The delay is a result of the substantial time periods that can be involved in the death certification process. When we publish a figure for the latest available year it will include deaths for that year, which are registered up to 15 months after the year end. This means that the data will be approximately complete when first published. However, there may eventually be a small number of further registrations after this 15-month period, in which case figures are updated during subsequent annual updates.
A series of validation checks are carried out on the annual death data before their incorporation into the registers. Validation includes checking for important missing information, such as date of birth or death, and checking for duplicates. Any queries are followed up with ONS and GROS. Coding of mesothelioma site is also carried out at this stage along with categorisation of asbestosis deaths according to the diseases mentioned on the death certificate.
As a further check on the completeness of the mesothelioma register, HSE receives details of all cancer registrations in Great Britain where the morphology code is associated with mesothelioma. Although the latest year for which cancer registration data are complete is less recent than for death data, cross-checking cancer registrations with the register provides a way of identifying a small number of additional deaths from mesothelioma. Any individuals having a cancer registration of mesothelioma, but who are not listed on the register are flagged for death notification with ONS or GROS. Any death certificates are then supplied to HSE and checked manually to identify mesothelioma deaths, which are then added to the register. Published statistics of mesothelioma deaths for the most recent years, which have not yet been validated in this way, are marked as provisional.
Before 1993, if there was insufficient information on a death certificate to accurately classify the death, the ONS sent a 'medical enquiry' to the certifying doctor for further information. This procedure was discontinued for deaths registered from 1993 onwards, but ONS hope to reintroduce it sometime in the future. This discontinuation has affected the site coding of mesothelioma deaths: the proportion coded as 'site not specified' was typically around 10-20% before 1993 and over 45% thereafter.
Under the Control of Lead at Work Regulations (CLAW) 2002 and the former 1980 and 1998 Regulations, all workers with significant exposure to lead are required to be under medical surveillance by an appointed doctor or one of HSE's medical inspectors. The surveillance includes measurement of each worker's 'blood-lead level', the amount of lead in samples of their blood, expressed in micrograms per 100 millilitres (µg/100ml). Annual returns give summary statistics for each workplace based on the maximum blood-lead level recorded for each worker under surveillance.
The Approved Code of Practice issued with the Regulations lays down levels of blood-lead concentration above which the appointed doctor is required to decide whether to certify that the worker should no longer be exposed to lead. If a worker's blood lead level reaches or exceeds this 'suspension level' a repeat measurement must be made, and if this is still at or over the level the worker should be suspended from working with lead. The number of such workers suspended is also recorded annually and analysed in the statistics. Under the 1980 Regulations the suspension levels were 70µg/100ml for males (80µg/100ml up to 1986) and 40µg/100ml for females of reproductive capacity (to protect the health of any developing foetus). The suspension levels were lowered in the 1998 Regulations (and remain unchanged in the 2002 Regulations) to 60 and 30 µg/100ml respectively, with new 'action levels' of 50 and 25 µg/100ml. The 1998 Regulations also introduced suspension and action levels for young persons aged under 18 years of 50 and 40 µg/100ml respectively.
Because of the time required to collate and validate the returns from all the appointed doctors and medical inspectors, the latest year for which data are available is 2004/05.
Enforcement notices cover improvement (requiring employers to put right a contravention of health and safety legislation within a specified time limit); prohibition (stopping work activity that gives, or will give, rise to a risk of serious personal injury); and deferred (stopping a work activity within a specified time) prohibition notices, as issued by all enforcing authorities, namely HSE, local authorities (LAs) and the Office of Rail Regulation (ORR),. Offences prosecuted refer to individual breaches of health and safety legislation; a prosecution case may include more than one offence.
Since April 2002 inspectors from HSE’s Field Operations Directorate (FOD) have, as part of routine inspections, provided a rating of a workplaces’ level of risk control against various ‘Risk Control Indicators (RCIs).Recent movements in these RCI ratings have been analysed (see ‘Trends in risk control [PDF 390KB] ’) with a view to providing supporting information for the judgement of progress against the Revitalising target.
HSE developed the WHASS programme of surveys as a valuable new source of information on work-related ill health and injuries, introducing more detailed information on other less direct but relevant measures of health and safety in the workplace. Whereas injuries and ill health are relatively uncommon and are therefore difficult to measure by survey with sufficient precision, other facets of the relationship between work and health affect all workers. Measuring precursors of adverse health and safety will generally provide more information to analyse and break down by factors such as industry sector and workplace size. Further, the WHASS programme of surveys will obtain useful information about the perceptions of employers and workers of how well health and safety is managed and risk is controlled within their own working environment.
Future surveys under the WHASS programme directed at separate samples of managers and workers will be combined with the similar employer and worker surveys developed as part of HSE's programme monitoring arrangements. Plans for a linked survey that would directly compare the views of workplace health and safety managers with that of workers employed in their workplaces have been put on hold for financial reasons.
The first WHASS employer survey [PDF 420KB] , that questioned health and safety managers, has been completed and the first findings report published in November 2005.
The results from this survey, the 2005 WHASS employer survey, are based on responses from just under one thousand employers with responsibility for health and safety management in their workplace, a response rate of 63%.
The first WHASS worker survey [PDF 450 KB] was published in May 2006. The results from this survey, the 2005 WHASS worker survey, are based on responses from 10 016 British workers, a response rate of 26%. The survey was administered by telephone, with households selected by random digit dialing and a respondent selected randomly from household members who worked at some time in the last 12 months prior to interview.
Further details about the methodology can be found in the Technical report [PDF 265KB] . Since non-responders outnumber responders by three to one in the worker survey there is potential for the responses to be unrepresentative of the population as a whole. For work injuries and work-related illnesses, where we can compare responses in the survey with the responses to the same questions in the LFS, this comparison implies that people who have suffered a work injuries in the previous year were 40% more likely to respond that people who have not. And people with a work-related illness are more than twice as likely to respond than those without similar biases could apply to other measures, such as exposure to particular kinds of hazard, or reporting your working conditions as stressful. In these cases we do not have the same kind of direct comparison that can be made with the LFS. However the Third European Working Condition Survey (run in 2000) has some similar hazard questions, and a comparison with the UK data from this survey does not suggest that the worker hazard responses are substantially biased. A module included in the Office for National Statistics Omnibus Survey in April and May 2005 had some overlap with the WHASS questions on stress. Here again the findings were similar, encouraging the view that -- apart from reports of injury and illness -- the worker responses can be regarded as reasonably representative.
All estimates derived from the 2005 WHASS employer and worker survey are subject to a margin of error. The main factor which determines the width of the margin is the number of sample cases an estimate is based on. In published reports, the sampling errors are often expressed using 95% confidence intervals. Each of these represents a range of values which has a 95% chance of containing the true value in the absence bias.