Health and Safety
Executive / Commission
Statistics
Cases of work-related cancer are indistinguishable from cases that are due to other causes. For this reason it is not possible to accurately assess the number of cases of work-related cancer from records of cancer registrations or deaths, or in schemes such as the THOR network.
An approach to overcoming this is to estimate the number, or proportion, of cases of cancer that would not have occurred in the absence of workplace causes (the attributable number or attributable proportion – sometimes referred to as the attributable fraction). Doing so requires there to be available information about the risk of cancer from workplace hazards relative to the general risk, information on the numbers exposed to those risks and information on the total numbers of cases of cancers of interest.
An early estimate of the overall proportion of cancer attributable to occupational causes was that put forward by Doll and Peto in 1981 in a report to the US Congress. In their report, they considered cancer on a site-by-site basis and estimated the possible contribution of occupation to cancer mortality for each site.
Updated estimates of the burden of occupational cancer based on that proportion are presented below but HSE acknowledges that Doll and Peto’s estimates should be revised to take account of more recent research and historical exposures to workplace carcinogens in Great Britain. Work is in hand to do this and initial findings from the first phase of this work which include estimates for six cancer sites have now been published. The proceedings of two workshops associated with this project are also available: November 2004 workshop [PDF 238KB] and June 2006 workshop [PDF 5.08MB].
Doll and Peto’s estimate of the overall proportion of cancer attributable to occupational causes was 4% and they suggested that this would be out by no more than a factor of about two. Applying this proportion to current levels of cancer mortality in Great Britain gives an estimate of 6 000 cancer deaths per year caused by work, with an associated uncertainty range of 3 000 to 12 000. Around 5 000 of these deaths would occur in men and 1 000 in women.
Doll and Peto’s estimates for individual cancer sites can also be applied to current mortality and cancer registration data for Great Britain to give an indication of the current burden of occupational cancer for each cancer site. On this basis, the cancers contributing the largest number of estimated deaths in men are: lung cancer (2 850), other and unspecified cancers (480), pleural and respiratory cancers other than lung (430), bladder cancer (310), leukaemia (230), and mesentery and peritoneal cancer (180). (Other and unspecified cancers mainly consists of secondary and ill-defined neoplasms.) In women the only cancer contributing more than 100 estimated deaths is lung cancer (660). Estimates for the full list of cancer sites are shown in Table CAN01
Two obvious problems with these estimates are that the number of mesothelioma (mainly pleural and peritoneal deaths is severely underestimated - 2037 mesothelioma deaths were recorded in the Mesothelioma Register for Great Britain in 2005) - and the number of other and unspecified cancers might be implausibly large. Another problem with an estimate based on mortality is that the detection, diagnosis and treatment of cancer has improved survival for many cancers, and so any estimate of burden for less fatal cancers is better based on estimates of incident cases.
Asbestos related cancers will make a major contribution to any estimate of the overall burden of work-related cancer for Great Britain. The number of mesothelioma deaths has approximately quadrupled since the early 1980s and in additions there is estimated to be around one asbestos-related lung cancer for each mesothelioma. See mesothelioma and lung cancer sections for further details.
If it is assumed that the attributable fractions used by Doll and Peto are also relevant for cancer incidence data, then the estimated number of incident cancer cases due to occupational causes is around 12 000 (uncertainty range 6 000 to 24 000). The cancers contributing the largest number of estimated cases in men are skin cancer other than melanoma (3 500), lung cancer (3 200), prostate cancer (700) and pleural and respiratory cancers (550). In women, the cancer with the largest number of estimated cases are lung cancer (760), and skin cancer other than melanoma (570). Skin cancers other than melanoma are often under-reported at cancer registries and so the true estimated incidence of these cancers is likely to be higher than estimated here.
Research into the Burden of Occupational Cancer in Great Britain commissioned by HSE is reported at Rushton Cancer Burdens report. So far, only six cancers have been addressed so this research does not yet provide an overall estimate of work-related cancers. The information included in Table CAN02 is presented for comparison with that derived from up-dated Doll and Peto estimates. It demonstrates how the estimates will vary according to whether “established” or “established and uncertain” carcinogens are included, reflecting the importance of presenting clear information on the parameters used to derive such estimates when interpreting and using them.
Certain cancers, which are listed below, are compensable under the Department for Work and Pensions Industrial Injuries and Disablement Benefit (IIDB) scheme. The total number of people of compensated amount to 50-150 cases per year during the last ten years (Tables IIDB01, IIDB03, IIDB05), and thus include only a small percentage of the true number of occupational cancers occurring. In these tables, note that the 2006 figures for disease references D8, D8A and D9 are not directly comparable to previous years. This is because the criteria for eligibility for compensation changed on 6 April 2006.
Reports from the SWORD/OPRA surveillance scheme provide data on the number of cases of occupational Lung Cancer reported to chest physicians and occupational physicians, as shown in Table THORR01. The estimated number of the lung cancers has fluctuated somewhat erratically since 1998 with an average of about 120 cases per year. Table THORS01 shows the number of cases of occupational skin cancer reported to occupational physicians and dermatologists in the EPIDERM/OPRA scheme. The estimated number of skin cancers has also varied considerably, with the largest estimated number of cases – 760 - being reported in the latest year (2006). However, statistical modelling of the EPIDERM/OPRA data by the University of Manchester to allow for various factors that affect reporting levels (including the number and type of participating specialist and occupational physicians, their reporting habits, and seasonal effects associated with the time of year they report) in fact suggests an overall downward trend in incidence.