The research was carried out by the Small Area Health Statistics Unit (SAHSU), which was set up in 1987 to develop statistical methods for investigating reported clusters of disease around nuclear and industrial installations.
In November 1990, the Department of the Environment announced that SAHSU was to investigate the incidence of mortality and cancer around solvent incinerators, benzene and vinyl chloride works, coke ovens, smokeless fuel plants and municipal incinerators (ENDS Report 190, pp 7-8). The studies were due to be completed by the end of 1991, but the work has proved more difficult than expected. Only the solvent incinerator study has been completed.
The study was triggered by an apparent cluster of laryngeal cancers around an incinerator at Charnock Richard, near Chorley, Lancashire, which was operated by Waste Incineration between 1972 and 1980.
The cluster came to light after the local authority and local health authority commissioned studies in response to public complaints about the plant. The studies identified four cases of the disease within two kilometres of the incinerator. A subsequent statistical analysis of the data concluded that there was a statistically significant association between the incidence of laryngeal cancers and distance from the plant.
The SAHSU study checked the data and applied more sophisticated statistical techniques. It was also extended to nine other incinerators burning similar wastes. These were operated by Associated Portland Cement, Cleansing Services, Evode, Glaxo (3), Kodak, Robinson Brothers and Chemical Manufacture and Refining. Some are now closed.
SAHSU looked at the number of cancer cases recorded in ten concentric circles around each plant. A lag of five years and ten years after the start-up of each incinerator was allowed for any effect to show up in the health statistics.
The research failed to turn up anything of note. The ratio of observed to expected cases of laryngeal cancer within radii of three and ten kilometres from each works did not deviate from one to a statistically significant extent. The same result was reached when data for the ten sites were pooled.
SAHSU also applied a more complex analysis using all 10 concentric circles around the sites. This was designed to detect any deviations from expected frequencies of disease, and any trend with distance from the incinerators. No significant differences or trends were found.
The report concludes that the number of laryngeal cancer cases around all the incinerators was within limits that could readily be attributed to chance variation. The statistical methods used in the earlier Charnock Richard study were more sensitive to local variation in risk, according to SAHSU. However, its conclusions had been devalued by the use of poor controls. SAHSU was able to overcome this problem because it had access to cancer data which were related to the areas under study by postcode.
However, SAHSU has acknowledged four weaknesses in its own analysis which may have resulted in an under-estimate of the effects of the incinerators. Cancer registration data are incomplete. The ten-year gap between censuses allows no interim adjustments to be made for local population changes. Even the ten-year period allowed after the incinerators began operation may be too short for the detection of many cancers. And its approach took no account of stack heights and wind directions and their effect on plume dispersion.