Fraud Taints Alzheimer’s Disease Research
The health ministry is investigating suspicions that researchers falsified clinical test data and deleted inconvenient information in a 3.3-billion-yen ($31.5 million) project on treating Alzheimer’s disease. At least one researcher has acknowledged that numbers were deliberately altered at the Japanese Alzheimer’s Disease Neuroimaging Initiative (J-ADNI), which is modeled after the Alzheimer’s Disease Neuroimaging Initiative set up by the U.S. National Institutes of Health.
Other officials involved in the project cited their own lapses in memory when asked about the suspected data tampering.
The central government, through the industry, health and science ministries, has spent 2.4 billion yen on the project, while 11 pharmaceutical companies have established a consortium that has contributed 900 million yen. A major objective of J-ADNI is to study the early signs of Alzheimer’s disease to develop new medicines and treatment methods. The research involves neuroimaging and blood tests to determine the relationship with such symptoms as memory loss.
“In addition to falsification of data, there were also subjects who were not fit to take part in the clinical tests due to the state of their health,” Takashi Asada, a professor of medicine at the University of Tsukuba and a key member of J-ADNI, told The Asahi Shimbun. “As far as we know, close to 20 percent of the data is unusable.”
Asada was part of the group that examined the test data. On Jan. 7, he submitted a report to the University of Tsukuba on tampered data from clinical tests. The project is headed by Takeshi Iwatsubo, a professor of medicine at the University of Tokyo who is a leading authority in Japan on research into senile dementia. Thirty-eight medical institutions around Japan are cooperating in the project.
The Asahi Shimbun has obtained internal documents showing the J-ADNI data center had asked medical institutions to rewrite test data on at least four occasions. The J-ADNI established the data center, consisting of researchers and employees dispatched from the pharmaceutical companies, to summarize and compile test results.
In line with international moves to share research results, J-ADNI tests used the same methods developed in the United States, which had started the research earlier. The use of similar methods is intended to allow for the wide use of the test results.
One suspected tampering of data involves interview tests conducted on 545 senior citizens since 2008, specifically a clinical test at the Kyoto Prefectural University of Medicine on Aug. 27, 2009. According to an internal document, an initial report on a memory test said that subjects were read a story and then asked an hour later, at 5:22 p.m., to recall what the story was about. However, international test methods call for asking the subjects to recall the story 30 minutes after they were told about it.
Several months after the initial test, the data center sent a document to the Kyoto Prefectural University of Medicine, asking that the time of the test be corrected to between 30 and 40 minutes after the story was read.
A university official changed the time of the test to 5:02 p.m., 20 minutes earlier than the initial report.
According to university officials, the individual who changed the time said: “I could not remember the time because so much time had passed since the test. However, since we were told we had used a different method, I thought I had made a mistake so I changed the time to match the established method.”
When asked about another test result that was “corrected,” the individual said, “I do not remember that instance.”
Other internal documents asked the National Center for Geriatrics and Gerontology as well as Yokohama City University to rewrite data about subjects to make them appear to have less serious symptoms.
“I do not remember why I rewrote the data, but it is common to slightly revise data to match standards,” said an official at the national center in charge of the research.
According to sources, the computer system used at the data center does not allow for entry of several different symptoms, so changes had been made to allow for data input.
Officials at Yokohama City University declined to comment on the project. Ministry of Health, Labor and Welfare officials have questioned Iwatsubo and Asada about how the data was changed.
“We would lose international trust if there was tampering,” a health ministry official said. “If there is data that has to be eliminated, it should be deleted and the research should be compiled based on the remaining data.”