The Canadian Nuclear Safety Commission (CNSC) has reevaluated data that it contributed to a 15 nation pooled study of low dose radiation cancer risks published in 2005. The determination was that the data was incorrect, needed further evaluation and should be removed from the cohort used for the 15 nation study. When that action is taken, the reported excess cancer risk disappears into a statistical insignificance.
A little more than a week ago, a reader who used to be a Reactor Operator (RO) at the San Onofre Nuclear Generating Station (SONGS) contacted me for information to support his on-going “battle” with antinuclear activists. In one exchange an activist mentioned a study titled Risk of cancer after low doses of ionising radiation: retrospective cohort study in 15 countries conducted in 2005.
That study reached the following conclusion as documented in the abstract:
Conclusions These estimates, from the largest study of nuclear workers ever conducted, are higher than, but statistically compatible with, the risk estimates used for current radiation protection standards. The results suggest that there is a small excess risk of cancer, even at the low doses and dose rates typically received by nuclear workers in this study.
I initially could not offer any response. However, I just came across a note in a paper titled Low Dose Radiation Adaptive Protection to Neurodegenerative Diseases by Mohan Doss which is available as a prepress article from the Dose-Response Journal. That paper included the following quote:
The 15-country study of radiation workers (Cardis et al., 2005), which was quoted as supportive evidence for the carcinogenic effect of LDR in Appendix V of BEIR VII Report (NRC, 2006), no longer shows increased risk of cancer from LDR. The Canadian data, which was part of the 15-country study, has been withdrawn from use because of newly identified problems with the data (CNSC, 2011).
Here is the footnote information: CNSC. 2011. Verifying Canadian Nuclear Energy Worker Radiation Risk: A Reanalysis of Cancer Mortality in Canadian Nuclear Energy Workers (1957-1994) Summary Report. INFO-0811. Canadian Nuclear Safety Commission. Published June 2011.
Here is the executive summary of that paper:
In 2005, the International Agency for Research on Cancer (IARC) released a 15-country study on the mortality of nuclear energy workers (NEWs). This study showed a statistically significant increase in the risk of mortality from all cancers excluding leukaemia in relation to radiation exposure, with Canadian data the chief driving force behind the worldwide results. The CNSC commissioned a reanalysis of the Canadian portion of the data to understand the unexpected findings. The study is now complete and summarized in this document.
When the researchers looked closely at the data by categories of workers, they identified that the statistically significant increase was driven by a single group of workers. They found that a group of 3,088 former Atomic Energy of Canada Limited (AECL) workers who had been hired prior to 1965 was the only group with “a consistent radiation-associated increase in risk of solid cancer mortality”.
This group of AECl NEWs had a profound impact on the Canadian and 15-country study findings.
It is very likely that these early AECl NEWs have incomplete dose information (i.e., their doses are under-reported).
As a result of the detailed analysis, the CNSC concluded
The CNSC does not have confidence in the historical AECl dose data (1956-1964). The apparent increase in the risk of solid cancer mortality for these early AECl NEWs deserves further investigation.
The CNSC, Health Canada and AECl are further assessing the dose data of early AECl NEWs to resolve the existing outstanding issues. Health Canada has agreed not to share the Canadian cohort for any further epidemiological research until the quality of the data file has been confirmed.
(Note: NEW – nuclear energy worker)
After finding those documents, I wondered if there was any kind of process that would result in changes to the original 2005 study. As you can see if you follow the link, that document is still readily available with no apparent corrections or cautions attached. That’s not surprising; after all, paper copies of published work are also permanently available without any notes or cautions attached.
The process of correcting the historical record is still in progress with some recently published updates. At least some of the additional work recommended in the 2011 paper has apparently been completed.
On November 22, 2013, the CNSC issued a press release titled CNSC Nuclear Energy Worker Study Published in British Journal of Cancer.
That press release pointed to an editorial published on November 13, 2013 in the British Journal of Cancer titled Nuclear worker studies: promise and pitfalls. Here is an important quote:
So, it would appear that the most reliable results from the 15-country study are for the combined 14 countries excluding Canada, which are not exceptional (see above).
The latest Canadian worker study by Zablotska et al (2013) illustrates the care that must be exercised in collating worker data, and the problems that can arise, especially when using data that may have been collected for purposes other than epidemiology.
This post is not one of my most readable or entertaining contributions, but I thought it would be worthwhile to document the path I took — along with all relevant links — to come to the conclusion summarized in the lede. It is not always easy to try to describe the complex process of knowledge acquisition; perhaps someday this documented set of links will help someone answer a challenge from someone who did not get the memo that the original 15 nation study included flawed data.
How can we best use modern communications tools to set the record straight when data errors that affect previously published work get discovered?