COMARE minutes 10th June 1999

COMARE minutes 10th June 1999: extracts.
The material on this page (along with other interesting correspondence)
was received by LLRC 23rd July
in a plain brown envelope


COMARE PAPERS                                                                             COMARE 99-12

FOR MEMBERS' USE ONLY

Meeting:                                                                     Agenda No 2 (iii)

COMARE 56 10 June 1999

For information and discussion

Paper enclosed:
Correspondence with Mr Michael Meacher MP, Environment Minister

Author(s): Mr M Meacher and Prof B Bridges


ITEM 7 Childhood LEUKAEMIA IN WALES
Welsh Office covering paper (COMARE 99-7) the Welsh Cancer Intelligence
Surveillance Unit (WCISU) report (COMARE 99-8)
Report by Chris Busby (COMARE 99-9)

7.1 The Chairman welcomed Dr John Steward and Mr David Adams Jones of WCISU. He noted that Mr Adams Jones used to be the Information Services Division (Scotland) Assessor to COMARE but is now acting as a consultant to WCISU.

7.2 Dr Steward outlined the background to their paper, "Results of a preliminary study to test the Irish Sea proximity hypothesis of Busby et al" (COMARE 99-8). Dr Chris Busby of Green Audit had visited WCISU in 1998. Discussion had touched on a possible protocol for testing the hypothesis that leukaemia rates increase with proximity to the Irish Sea. The subsequent study aggregated small geographical areas into bands of increasing distance from the Irish Sea and identified a reducing gradient of relative risk for the incidence of leukaemia in children under 5 years for the 16 year period, 1974 - 1989. Although the work does not appear to have been published in any scientific or medical journal, the Guardian carried an article reporting the results on 31 December 1998 and the study was also discussed in a BBC Wales documentary on 9 February. The results are also available on the Low Level Radiation Campaign web site.

7.3 In view of public interest in the study, the Welsh Office instructed WCISU to carry out a preliminary analysis to test for evidence of the effect. Dr Steward contacted Dr Busby requesting a copy of his report. In his reply, Dr Busby explained that he was unable to provide a copy without permission of the Irish Government for reasons of confidentiality. However, Dr Steward was able to obtain a copy of the leukaemia and population data used as well as a summary of the methodology.

7.4 Dr Steward decided to repeat the analysis of Busby et al using data from the current WCISU database. Data for each distance band was reconstructed using information on 1991 wards supplied by Dr Busby. The analytical method used was checked and agreed by Dr Busby.

7.5 Dr Steward found no evidence to support Dr Busby's hypothesis. The general level of relative risks for leukaemia were in line with those expected for England and Wales and there was no evidence of a decreasing gradient of leukaemia risk with distance from the Irish Sea

7.6 Dr Steward explained that Dr Busby's work was flawed by a significant error. For urban areas of Wales, Dr Busby had originally identified 400 cases of leukaemia. but corrected this after publication to 118. For rural Wales, Dr Busby found 138 cases of leukaemia in the 0-4 years ago group for the period 1974-89, whereas only 48 cases had been identified on the WCISU database for the same area and period. Dr Steward believed that this discrepancy arose from a data input error which added counts from other age bands to the 0-4 age group. Therefore the problem with the study stemmed not from the epidemiology, but the data. Furthermore he was convinced that Dr Busby was aware of the data problem but would not acknowledge or correct the mistake. In Dr Steward's view, this amounted to scientific fraud. [Dr Steward has denied saying this -- see note below.]

7.5[Note: the non-sequential numbering is in the original]
In conclusion, Dr Steward wondered whether COMARE might be able to assist in obtaining copies of the dataset used by Dr Busby in order to demonstrate how the number of leukaemia cases in rural Wales came to be trebled.

7.6 The Chairman thanked Dr Steward for his presentation and asked Members to consider whether there is a raised incidence of childhood leukaemia in the coastal regions of Wales and, if so, whether there could be any association with Sellafield discharges. The Chairman opened the discussion.

7.7 Dr Draper wished to declare an interest. He had discussed Dr Busby's work with the Welsh Office and had unfortunately been the subject of personal attacks by Dr Busby. Dr Draper was happy with the methods adopted by WCISU to reconstruct Dr Busby's analysis, however, he recommended that Dr Busby be asked to provide a copy of the dataset he had used in order to establish where the error in the number of leukaemia cases had occurred.

7.8 Dr Cartwright reported that he had also been discredited by Dr Busby and agreed with Dr Draper's assessment and recommendation. He mentioned that there had been a number of studies which looked at the incidence of leukaemia in relation to increasing distance from coasts, estuaries and the confluence of rivers.

7.9 Professor Clayton questioned Dr Busby's use of very narrow distance bands and the fact that they were defined by a non-linear scale. He also noted that mean distances of each band from the sea should be included in. the WCISU paper to facilitate a more direct comparison. Dr Steward said that he had replicated Dr Busby's methodology and mean distances of each band could be included to add clarity. In future analyses he would prefer to compare specific data points. Professor Clayton observed that coastal areas tend to be more urban than areas inland and this could give rise to a possible confounding effect.

7.10 Professor Eden questioned whether the Committee should be wasting its time on a scientifically suspect paper which had not been peer reviewed. Professor Little asked whether the Welsh Cancer Registry data had been checked against the National Childhood Cancer Registry. Dr Steward said that it had. Professor Little noted that Dr Busby had redefined his distance bands in terms of 1991 census wards and asked whether 1991 population data had been used. Dr Steward said the study had not been adjusted for 1991 population statistics or the social and economic status of subjects.

7.11 The Chairman explained that soon after the Committee had received a copy of the paper by Dr Busby it became clear that the data was at odds with that held by WCISU. The Chairman therefore decided to undertake an independent, "blind" check. He asked Dr Busby to provide leukaemia data for Dyfed, Mid- Glamorgan and Powys. Dr Busby sent him data for 1979 as requested. The Chairman then asked Dr Draper how many cases were registered for these regions on the National Childhood Cancer Register. Data for Mid-Glamorgan and Powys were consistent, but for Dyfed, Dr Busby reported 48 cases whereas the National Registry reported only 8. There appeared to be an error in the data held by Green Audit and until this problem was corrected, the paper by Dr Busby should be discredited.

7.12 The Chairman summarised the information before the Committee. Dr Busby had noted that the number of leukaemia cases for Wales was too high to be credible and had corrected the figure from 400 to 118. It was likely that the data had been corrupted by the format in which it was transferred. This should have prompted a re-examination of the source data originally supplied by the Welsh Cancer Registry. However, the Chairman suggested that the Welsh Cancer Registry should not escape blame since they failed to keep a record of the data supplied to Dr Busby. In order to resolve the issue and to overcome what appeared to be a conspiratorial mentality, the Chairman recommended that Dr Busby should he urged to provide a copy of the data to COMARE or WCISU who would try to establish how the data came to be corrupted. On the basis of the WCISU study led by Dr Steward, COMARE could find no evidence for a decreasing gradient of leukaemia risk with distance from the Irish Sea coast. Green Audit had therefore aroused needless concern and should be asked to withdraw the report since it was clearly invalid. The Chairman added that a conclusion derived from work which replicated a flawed study would lack credibility. Therefore, the only way to proceed would be to carry out further research. The key question was whether this was justified.

7.13 Professor McMillan asked whether it was possible to be sure that there was no coastline effect on the incidence of leukaemia. Professor Clayton also thought it was premature to say that the coastline effect does not exist. He would support further coastal analysis but not for the reasons put forward by Dr Busby. Dr Adams Jones said that he could supply data and grid references for all cases along the coastal fringe as input to an independent project. Mr Crawley questioned whether the Welsh Office should support more work in this area. The Chairman noted his request for a steer from the Committee. Professor Boddy commented that the public would think that COMARE were not carrying out their duties unless further action was taken to address Dr Busby's hypothesis.

7.14 Professor Little wondered whether there was scope for the issue to be addressed by the Geographical Assessments Group, (GAS). Professor Cartwright was concerned that the group would not be able to respond within the time scale required. The Chairman thought there was little scientific justification for the study but as part of the Committee's remit to keep the public informed, the political pressures could not be ignored. He was also mindful of implications for COMARE of recommending expensive and scientifically unjustified research. Mr Crawley wondered how much it would cost to fund a project to clear up these issues. The Chairman said that this would depend on what the research set out to achieve, but the average cost of a research paper was about 10K.

7.15 Professor Haites wondered whether a study confined to the coast of the Irish Sea would identify any recognisable trend and whether, to be meaningful, research should consider other important coastal areas. The Chairman acknowledged the need for a large enough dataset, however, if the scope of the study was too wide, the costs would be prohibitive. He asked the Committee whether they would wish to recommend a further study to test Dr Busby's hypothesis.

7.16 Dr Hamlet said that this would raise Dr Busby's credibility and would open the door for others to lean on COMARE to recommend research. Professor Eden could not support the recommendation for further research; Dr Busby's work had not been peer reviewed, it contained known errors and there was no scientific justification to test the hypothesis further. Professor Boddy added that by seeking to obtain the data from Dr Busby and comparing this with the National Registry of Childhood Cancers, COMARE had fulfilled its role as watchdog. Dr Sharp agreed that there were no sound scientific reasons to suggest the need for further research. Dr Harrison said that under normal circumstances, the Committee would not concern itself with work that had not been reviewed and published.

7.17 Mr Crawley assumed from the discussion that COMARE would not recommend further investigation of the Busby hypothesis for the 0-4 year old age group because the number of cases was too small. He wondered whether the same argument applied to other age groups and malignancies. Dr Draper said that Dr Busby had referred to brain cancers and noted that other published papers had failed to identified a positive link with coastal effects.

7.18 The Chairman concluded that COMARE's recommendation not to pursue further research was wholly defensible and on the basis of current evidence there was no public health problem.

[Action: Secretariat]

ITEM 8 ANY OTHER BUSINESS

8.1 The Chairman wished to report recent progress of the Advisory Group on Ionising Radiations (AGIR) which he also chairs. He explained that NRPB set up AGIR in 1995 to run in parallel with the Advisory Group on Non Ionising Radiations, (AGNIR) chaired by Sir Richard Doll. AGIR was initially assigned four areas of work for which appropriate subgroups were set up. Some groups were slow to start and the rate of progress has varied.

...

8.4 When the group last met on 18 February a number of key issues emerged for the future work programme. Professor Little will lead a team looking in detail at the risks associated with radon and Professor Goodhead's group will aim to keep abreast of recent developments in genomic instability and its implications for radiation protection. Dr Roger Cox and Dr Alan Edwards of NRPB are currently reviewing Chris Busby's "second event" theory which suggests that cells are more sensitive to chemical and radiological exposure than previously thought. There is limited scientific support for Busby's hypothesis and political pressures have brought this issue before AGIR. The review paper will be considered by AGIR at the next meeting.


Here is an exchange of emails between LLRC and Dr Steward from 11 - 12 July 2000

Dear Mr Bramhall,

> ... regardign the Website, I note that it contains "leaked"
>minutes from a COMARE meeting of March 1999 which I attended as an external
>speaker. I note that the minutes attribute me as saying that there has been
>a deliberate scientific fraud:-
>
>Section 7.6 of COMARE minutes 10th June (received by LLRC 23rd July)
>states:-
>
>Furthermore he (Dr Steward) was convinced that Dr Busby was aware of the
>data problem but would not acknowledge or correct the mistake. In Dr
>Steward's view, this amounted to scientific fraud.
>
>These minutes are quite inaccurate in this respect, I do not hold that view
>and have never stated it. On the contrary, it is my view, and it always has
>been, that there has been some sort of computer error either on the
>extraction of the original dataset by Health Solution Wales or on the
>reading in of the data by Green Audit, but without seeing the original data
>as supplied it is difficult to say which is true (why not send it to us as
>requested?). However, in the light of the aggregated spreadsheet supplied
by
>Dr Busby I now tend towards the data extraction error hypothesis. We have
>very high agreement between us on the conventional data columns, the
>difference is entirely attributable to the nonsensical "All Leukaemias"
>column. This is now a matter for COMARE to investigate.
>
>We continue to investigate the very interesting epidemiological hypothesis
>of an Irish Sea effect on cancer incidence and welcome any information you
>can supply to throw any further light upon the evidence for increased
levels
>of exposure of the population to low level radiation.
>
>John Steward
>________________________________________
>Dr J A Steward MBBCh BA MSc PhD FFPHM
>Director
>Welsh Cancer Intelligence & Surveillance Unit
>14 Cathedral Road
>Cardiff CF1 9LJ
>_______________________________________
>tel:- 029 20 373500
>fax:- 029 20 373511

Dear Dr Steward,
...
I note your disclaimer on the COMARE minutes. What can I say? I assure you
that our transcription is entirely faithful to what we received in the brown
envelope. How would it be if I pasted your comments [i.e. what you have
written in this email ] onto the website? Would that be acceptable to you?
yours
Richard

Thankyou that would be fine my me. I did not say it.
John Steward

This page was last updated May 2001