A Cancer Excess South-East of the Hillhouse Site, and 'No Further Cluster Investigation'

From the Lancashire County Council Health Cell’s kidney-cancer report on the Hillhouse area.

This is not a claim that PFOA caused any cancer recorded in this data. The report does not make that claim, and neither do we. It is a description of the report’s own framing and conclusion, set against the figures the same report publishes, on kidney cancer rates around the AGC Chemicals Europe (formerly ICI) plant at Hillhouse, Thornton-Cleveleys.

The Lancashire County Council Health Cell framed its task, in its own words, twice, as a population-level question. In the same document it recorded an excess in one neighbourhood that it deemed statistically significant against England’s rates, and it examined that excess before reaching its conclusion.

The remit, in the report’s words

The Health Cell’s kidney cancer assessment states its own remit in a specific way. In its executive summary, it says the Director of Public Health convened the Health Cell “to consider whether there were any population level health concerns requiring further assessment.” In its conclusion, it repeats the same frame: the assessment “was undertaken to consider whether available NDRS data indicated a population level health concern requiring further investigation.”

That is framed as a population-level question. The report’s two stages show what the assessment covered.

What the two stages covered

The first stage was a screening. The National Disease Registration Service compared observed with expected cases, on England’s rates, for small areas within 5km of the site over 2003 to 2022, for cancers “both overall and by specific cancer type”. The report says that, overall, “the number of kidney cancer cases within LSOAs surrounding the site was similar to that expected”, and that the screening “did not identify higher than expected rates for most cancers”; kidney cancer was higher than expected in two small areas.

The second stage examined those small-area findings. Its purposes were “whether the observed excess of kidney cancer cases existed during 2003-2022 and when compared with alternative comparators” and “whether the observed pattern could, in principle, be linked to a plausible environmental exposure pathway”. Its Results report “no consistent spatial or temporal patterns” and “no evidence of an increasing gradient of risk in relation to proximity to the site”, and its Technical Appendix 2 sets the 14 cases in E01025596 against the 498 kidney tumours recorded in the 79 small areas across Wyre, Fylde and Blackpool that were listed to NDRS, by sex, age, tumour type, grade and stage.

The same report also names two neighbourhoods, and their numbers.

The same report, in the same assessment, also reports specific, named results for two small areas. The first, LSOA E01025596 (a small statistical area, similar to a neighbourhood) south-east of the site, described in the report as close to the site, recorded 14 kidney cancer cases against about six expected during 2003 to 2022. The standardised incidence ratio (SIR), where 100 would mean exactly as many cases as expected, was 218, more than double the expected number, with a 95% confidence interval of 119 to 367. A second area, LSOA E01012665, recorded 17 cases against 9 expected, SIR 181, CI 106 to 290.

The report itself calls both of these results statistically significant, against England’s rates. It states plainly: “these two LSOA were deemed to be statistically significant by the Health Cell.” Confidence intervals are the report’s own test for whether a difference from the expected rate is likely to be real or down to chance: if the range stays entirely above or below 100, meaning even at its least extreme point the difference stays above the expected level, the report says, the difference is considered meaningful. Both intervals here sit entirely above 100.

Two results, side by side

The report records both, in the same pages: an excess against England’s rates, which it deemed statistically significant, in the one small area south-east of the site; and, against other small areas, a rate in E01025596 that “was statistically similar to those in most other LSOAs, although it was higher than the England average”.

The report gives its own reasons for not treating the small-area result as a confirmed cluster: the numbers are small and therefore unstable, and the screening tested many cancers and areas at once so some statistically significant findings are expected by chance alone; and, in its Results and Conclusion, no consistent spatial or temporal pattern; no gradient of risk with proximity to the site, no clustering of cases close to it and no pattern of decreasing incidence with distance; “no alignment with expected environmental air dispersion or soil contamination data”; cases “geographically dispersed”; tumours showing “heterogeneity consistent with background incidence”; and findings “consistent with the level of random variation expected when undertaking multiple small analyses and multiple statistical comparisons”. These are stated, disclosed reasons, set out in the report’s Results, Conclusion and Limitations sections, and they are legitimate considerations in cluster epidemiology. On those grounds the report concluded “that there was no evidence of a kidney cancer cluster, no indication of an environmental association, and that further cluster inspection was not required”, and that, taking its limitations into account, “the observed variation in kidney cancer incidence is consistent with normal statistical fluctuation.”

What this is not

This is not a claim that PFOA caused the cancers recorded in E01025596. The report itself did not attempt to establish that, and neither does this piece. It is a description of how the report’s conclusion sits beside the numbers the same report published for one small area.

The question this leaves open is narrow. Against England’s rates, the excess in E01025596 is statistically significant. The report’s conclusion also rests on comparisons with other small areas and other populations, which it gives as statements and charts but not as figures. What are the comparator rates, and the ratios calculated against them, that led the Health Cell to find the variation “consistent with normal statistical fluctuation”?

We are asking this openly. We are preparing to put it formally to the Health Cell, and will publish any response here.

How we know what we know

Everything above comes from one public document: Lancashire County Council Health Cell, “Assessment of Kidney Cancer Patterns at the Hillhouse Technology Enterprise Zone, Thornton-Cleveleys,” v1.0 Final, last reviewed 12 May 2026, drawing on National Disease Registration Service (NDRS) cancer registration data for 2003 to 2022. The remit is stated in its Executive Summary and repeated in its Conclusion. The overall screening result is in its Background and Methodology sections. The figures for the two named small areas, LSOA E01025596 and LSOA E01012665 (observed and expected cases, ratios and confidence intervals), are in its Background section; the “deemed to be statistically significant” finding, the comparison with other small areas and the gradient finding are in its Results, and the significance finding is repeated in its Conclusion. The comparisons by sex, age, tumour type, grade and stage are in its Technical Appendix 2. We have quoted these directly; the only arithmetic is confirming that both confidence intervals sit above 100. The report is published in full by Wyre Council, so every quotation above can be checked against it.

Status, 28 September 2026: we have asked Lancashire County Council, by Freedom of Information request, for the figures behind this assessment: observed and expected cases, standardised incidence ratios and confidence intervals for every area and every cancer screened, and the comparison populations used in the report’s second stage, with the rates and ratios calculated against them, which the report gives as statements and charts (Technical Appendices 2 and 4) but not as figures. A reply is due by 26 October 2026. Open questions

Corrected 11 October 2026: the heading and text, to give both Health Cell areas, and the confidence-interval explanation. See the changelog.

Corrected 12 October 2026: the two areas in the county’s kidney cancer assessment are now identified by the codes the report uses (E01025596 and E01012665) in place of our own description of where the second one is; the report does not call the second area a comparator. See the changelog.