Disputed study sees the light of day after a lengthy controversy
By Anne Harding | October 26, 2006
The journal Environmental Health has published a study showing abnormally high rates of cancer deaths among workers at IBM semiconductor plants, two years after another journal turned the study down under controversial circumstances.
IBM fought the study's author, Richard Clapp, in court over his right to publish the study, which was based on data the company released during the course of a lawsuit. In March, a New York judge ruled that the material was not confidential and that publishing it would be in the public interest.
IBM and the organization representing U.S. silicon chip makers maintain the study is flawed, and that workers in the industry are safe. "The valid studies that have been done to date do not show any evidence of increased cancer risk among clean room workers," John Greenagel, a spokesman for the San Jose-based Semiconductor Industry Association, told The Scientist.
Clapp, a professor of environmental health at Boston University, said that it's "gratifying" that information that could help prevent harmful environmental exposures is finally being made public. But he added that much more independent research needs to be done into health risks manufacturing workers in the semiconductor industry may face.
"The study done by Dr. Clapp gives us important leads that can be used for more detailed studies of cancer in semiconductor workers," Robert Harrison, a professor of clinical medicine in the University of California, San Francisco's Division of Occupational and Environmental Medicine, told The Scientist.
An unofficial organizing group for IBM workers also welcomed the publication. "We believe this reinforces the anecdotes we've been hearing and talked about for years from IBM facilities," said Lee Conrad, national coordinator of Alliance@IBM in Endicott, New York, where one of IBM's two remaining U.S. semiconductor chip manufacturing facilities is located. The other is in Burlington, Vermont.
"If IBM doesn't like this study, we should do more studies," said Conrad, who worked at the Endicott plant for 26 years but has since retired. "Let's get to the bottom of this."
Under court order in 2003, IBM handed over its Corporate Mortality File on employee deaths to plaintiffs' attorneys in a lawsuit brought by two workers at its San Jose semiconductor plant, who charged that they had developed cancer due to on-the-job exposures. The attorneys hired Clapp to analyze the data, which included nearly 34,000 employee or retiree deaths and covered the period from 1969 to 2001.
Clapp's analysis of 31,941 death records found higher rates of mortality from brain, breast, kidney, lymphatic and hematopoietic cancers, as well as melanoma, among the workers compared to the general public. The judge in the California case ruled the findings inadmissible, after which Clapp sought to publish the study in the journal Clinics in Environmental Medicine, in a November 2004 special issue on the electronics industry.
IBM lawyers sent Clapp a letter warning him not to publish, but he retained an attorney who told him he was within his legal rights. Then the journal rejected the paper. Elsevier, the journal's publisher, later said it refused the study because it only published reviews, not original work. However, the journal had published original studies in the past. Other contributors, including UCSF's Harrison, agreed to withhold their work in protest, at the request of the issue's guest editor, Joseph LaDou, of the University of California School of Medicine.
Chris Andrews, a spokesman for IBM, told The Scientist Clapp's findings "aren't credible, nor are they backed by any kind of legitimate science." According to Andrews, the Corporate Mortality File was an "incomplete human resources database that IBM used years ago in conjunction with providing benefits to beneficiaries of deceased IBM employees."
"It really contained no information that would support the study of and the drawing of scientifically valid conclusions with regard to diseases among IBM workers," he said.
Greenagel, of the Semiconductor Industry Association, said the proportional mortality ratio design of Clapp's study was not a valid approach to evaluating cancer risks.
Clapp has argued that the study design was necessary to control for the "healthy worker effect," which can skew research because worker populations exclude those who are too ill for employment.
Harrison also defended the science behind the study. "Dr. Clapp's analytic methods are used routinely by scientists who study the risk of workplace chemicals and cancer risk," he said. "For example, the proportionate mortality study design was used by researchers in the 1970s who found a link between chemicals used in the tire industry and high rates of leukemia."
Links within this article:
R. Clapp, "Mortality among US employees of a large computer manufacturing company: 1969-2001," Environmental Health, Oct. 19, 2006
A. McCook, "IBM Responds in Study Dispute," The Scientist, July 16, 2004.
Semiconductor Industry Association
A. McCook, "Researchers Boycott Journal," The Scientist, June 23, 2004