« What took so long? Scientists consider long-term health monitoring after radiation exposure | Main | Ann Arbor moves to harmonize radiation emergency planning with international best practices »
Thursday
Mar072019

Fukushima at 8: Accusations of scientific misconduct concern city in Japan

Date City produces peaches and dried persimmonsEight years after the Fukushima nuclear reactors exploded on Japan’s Northern coast, spewing radioactive particles into the air, across the land, and into the Pacific Ocean, the country continues to struggle with decontamination and relocation efforts. Determining the health impacts resulting from the nuclear disaster has been particularly fraught. For Date City, about 60 km from the ruined Fukushima reactors, and still blanketed by radioactive contamination from the ongoing catastrophe, the struggle for protection of health continues amid accusations of scientific misconduct and betrayal.

After the nuclear catastrophe began, Date City residents received glass badges that measured radioactivity. About four and a half years of measurements collected from these glass badges were used by Ryugo Hayano, Professor emeritus from the University of Tokyo and Makoto Miyazaki from Fukushima Medical University (FMU) to initially publish two studies in the Journal of Radiological Protection (JRP). Radiation policy makers in Japan often reference the second of these two studies, indicating they trust the data and conclusions it offers. However, earlier this year, Shin-ichi Kurokawa (Professor Emeritus of The High Energy Accelerator Research Organization) and Akemi Shima (resident of Date City) contended that this research and the studies using it, are compromised by serious ethical violations and scientific misconduct.

Date City officials requested the studies subsequent to their adoption of a 5 mSv annual radiation exposure limit, which represents a huge increase of radiation exposure to residents. Date City has also limited decontamination efforts in certain areas, and the former mayor Shoji Nishida, requested that the International Atomic Energy Agency proclaim 5 mSv per year safe, instead of the current 1mSv. More detailed information is coming to light as a new mayor of Date City has been elected.

Kurokawa first raised concerns about the second study in a peer-reviewed August 2018 letter sent to JRP. The JRP, a U.K. journal, has yet to publish Kurokawa’s critique, so he published it on a Cornell University website in December 2018. Kurokawa also published a timeline and further critique of Hayano’s response to the letter in Harbor Business Online in February 2019, original article in Japanese. This research has also been reported on Retraction Watch, a website that tracks published troubled papers, although there are more serious concerns than those RW highlights.

Hayano has admitted (English translation here) to a miscalculation that underestimated doses to Date City residents by three times. Hayano also admits to destruction of the data on which the studies were based, claiming this “deletion” was in accordance with research protocol of the study. But Kurokawa disputes that, pointing out that data destruction is a violation of Japan’s ethical guidelines on handling human data – guidelines that instruct researchers to keep the data as long as possible. This destruction of data, and failure to publish a promised third study, appear to conceal evidence that found very high internal doses of radioactivity in some residents of Date City.

The Date City glass badge experiment

The research used glass badge data from approximately 59,000 Date City residents. These badges, paid for and distributed by Date City, supposedly measured the external radioactivity that each individual was exposed to beginning about August 2011, approximately 5 months after the nuclear catastrophe began, until the summer of 2015. The mayor’s office of Date City provided both the glass badge data and data on internal exposures for individual residents.

According to research protocols agreed to with FMU, Miyazaki and Hayano planned to publish three studies based on these data. The first, comparing individual external doses to survey results of airborne radiation from the Government of Japan, was published in 2016. The second, a prediction of lifetime dose and an evaluation of the effect of decontamination on doses to individuals, was published in 2017. The third study, examining the relationship between external doses and internal doses, will not be published. Instead it has been replaced by a study on a different topic.

Where things went wrong

Bad glass badge data

Perhaps the experiment was doomed from the start as the Miyazaki-Hayano studies admit some residents of Date City may not have worn the glass badges on their bodies or actually lived at the address registered for the badge. Such improper badge use would immediately compromise any conclusions reached concerning individual doses, but the researchers used the data anyway.

Mishandling and destruction of data

In addition to questionable glass badge measurements, Kurokawa contends the Miyazaki-Hayano research suffers from mishandling and destruction of data that violates ethical guidelines:

  • Residents (research subjects) of Date City were not informed of the content of the research prior to the research commencing, and were not given opportunity to refuse use of their data. Miyazaki, being a municipal advisor on radiation to Date City as well as a study author, should have known how to handle this properly, yet he did not.
  • Miyazaki and Hayano failed to note that some residents had not consented to use of their data, a fact obvious in the data supplied to them by Date City. They further failed to obtain consent from those residents prior to use of their data.
  • Hayano presented data before the research protocol was submitted to, and approved by, an FMU Ethics review committee.
  • Residents were not told of the papers once they were published, nor were they told that the Mayor’s office of Date City had requested the papers be published. This presented a conflict of interest since the Date City Mayor’s office had an agenda (see slides 21 & 26) of encouraging residents to increase “resilience” while living in a contaminated environment. For residents, this means consuming contaminated food and restricting decontamination efforts per Date City’s new 5 mSv annual exposure limit. A few months after Date Mayor Shoji Nishida announced this “resiliency” policy, Miyazaki was hired as radiation advisor to the city.
  • Miyazaki and Hayano violated research protocol by replacing the third study originally agreed to, with a study that said nothing about internal versus external doses.
  • At the conclusion of the research, all of the data were destroyed. According to records obtained by an information request filed by Shima, Kurokawa’s co-author of the Kagaku article, Hayano created an integrated database at the request of Date City, but did not share this database with the city. Therefore, when the database was destroyed, Miyazaki and Hayano knew that Date City could not replicate it or the data it contained.

Kurokawa points out that research conducted in Japan must follow the ethical guidelines based on the Declaration of Helsinki for proper protocols in handling medical and health research involving human subjects, such as valuing welfare of the research subjects over that of scientific results. FMU approved the Miyazaki-Hayano research papers under these protocols – protocols this research seriously violated by not allowing people to control use of their own data and by destroying the data after publication so that neither researchers nor the research subjects, can access it or replicate the studies.

Underestimation of dose

In addition to the mishandling of data, Kurokawa has discovered discrepancies in the values of cumulative doses in paper 2, which appear to underestimate actual doses. Hayano has, by his own admission, underestimated individual doses by three times. Professor Hayano says that he will issue a correction (corrigendum) for this dose underestimation, but has failed to completely answer the additional serious discrepancies, and the ethical violations of mishandling and destruction of data Kurokawa notes.

Why the “phantom” third study matters

The missing third study was supposed to investigate correlation between external and internal individual doses – a correlation Miyazaki and Hayano had already hypothesized would not exist. However, upon reviewing other data in Date City reports, the opposite was found: “[there was very] clear correlation between the external and internal doses…some cases with very high levels of internal exposure measurements.” Kurokawa offers his own hypothesis as to why Miyazaki and Hayano never published a paper on this third research question:

The true reason for not publishing Paper 3 could be the discovery of a clear correlation between the external and internal doses with some residents showing internal exposure measurements of several thousand Bq even since 2015. Not publishing inconvenient results despite receiving the internal exposure dose data from Date City would have to be considered a violation of the Ethical Guidelines. (emphasis added)

This correlation also reveals that Date City’s “resiliency” plan is not protecting its residents. Miyazaki and Hayano’s unwillingness to address internal dose evidence in the Date City data also calls into question Hayano’s other research on internal doses issues such as monitoring of food and whole body scans of children, the last publication of which appears to be in 2015.

Mistaken assumptions based on faulty studies

Japan’s Radiation Council (JRC) on setting standards for protecting people from radiation often references this ethically and scientifically compromised research in discussions, particularly the second paper, which was the focus of Kurokawa’s critical letter. Hayano’s work is often mentioned by other scientists and press as indication that doses from Fukushima radiation are low, that decontamination efforts paid for by Date City funds, might not have been necessary, and that living in an environment contaminated by “low” levels of man-made radiation is acceptable.

Where was the peer-review?

For its part JRP has now determined at this time that a correction for the dose underestimation is all that is needed, while an investigation into the consent issue is conducted. JRP claims to adhere to the Declaration of Helsinki for proper protocols in handling medical and health research involving human subjects. However, data misuse and destruction should require retraction of the papers, not correction.

Kurokawa contends that underestimating 70-year lifetime doses by three times is a severe enough miscalculation that a mere correction will not suffice, implying the conclusions of the papers are now in jeopardy. Hayano is claiming, falsely, that JRP wants a rewrite of the paper. Even if JRP did want a rewrite, it is unclear how Hayano intends to accomplish this since the Date City data on which the original papers were based have been destroyed. Kurokawa states:

There is no way to rewrite a paper when the research has already completed and all the data have been destroyed. Even if Date City were to re-supply the data to FMU, it would be considered new research and a new research proposal would have to be submitted to the Ethics Review Committee at FMU. A resulting paper would no longer be a revised version, but an entirely different paper based on new research. A scientist should never conceal such information, let alone pretend as if what was requested by JRP was a rewritten paper when it was a corrigendum that was actually requested. (emphasis in original English translation)

To date, neither Miyazaki nor Hayano have responded in the customary fashion, which would be to answer Kurokawa’s original letter criticizing their published research point-by-point. Kurokawa has published an analysis of Miyazaki-Hayano paper 1 in the March issue of Kagaku in Japanese, and will be publishing detailed analysis of paper 2 in April 2019.

Thanks to Yuri Hiranuma for input and review of this article and for the translations used to write it. See Yuri's blog.