February 13, 2004 The Denver Post by Anne C. Mulkern

In arguing that the risk of a mad-cow epidemic is very low, USDA officials repeatedly offer as proof a November 2001 report by the Harvard Center for Risk Assessment. But experts who reviewed that Harvard work said it ignored important issues and focused too much on less important details.

The USDA withheld the comments from those critical scientists for 15 months, releasing them only after The Post and others filed Freedom of Information Act requests. The documents were put on the agency's website late Wednesday.

In one part of that peer review, the scientists say that the Harvard group appears to have underestimated 'enormously' the number of Swiss mad-cow cases that it uses in its mathematical modeling. The Harvard group conceded that its estimates were low.

The peer review is the latest questioning of USDA claims that mad cow is not a concern in the U.S.

'I don't think the Department of Agriculture has a great deal of credibility,' said Carol Tucker Foreman, director of Consumer Federation of America's Food Policy Institute and a USDA official in the Carter administration. 'The evidence is building that at the very least, the department should go back and consider if it needs to take additional steps' to protect food from mad-cow contamination.

The USDA said it stands behind the Harvard risk analysis and that points made by the peer review group were 'minor.'

'They did not identify any fatal flaws in the (risk assessment) model,' said Lisa Ferguson, USDA senior staff veterinarian. 'We believe the model is effective, and we stand behind its use as a tool.'

The USDA said it does not normally release peer reviews but did so in this case after numerous inquiries.

The agency on Monday ended its official inquiry into the events surrounding the December discovery of one Holstein with the disease in Washington state. The USDA said it was no longer looking for 52 cows that had been imported from Washington with the ill cow.

The peer review analysis does not say how much the first study underestimated the risk, nor does it define the likelihood of people contacting the human form of mad cow, a deadly, brain-wasting illness.

Foreman, the food safety expert, said the risks couldn't be quantified. The peer review takes that position as well and said that even with the flaws in the first report, the risks could still be very low. But, it says it's impossible to know that because of problems with the report.

One of the flaws noted in the peer review is that the Harvard report authors assume that once mad cow is found in the United States, conditions affecting the spread of the disease wouldn't change for two decades.

'This is a huge assumption and probably unrealistic,' the peer reviewers said. 'As with most agents of disease, especially newly discovered agents (emerging diseases), prevalence increases over time largely because of more and improved testing over time. This has not been incorporated into the model.'

It mentions that E. coli O157:H7 bacteria was once thought to be rare and now is widespread.

The peer review report also says the Harvard group failed to look at the risk that infected material could be imported from places other than the United Kingdom. Cattle largely became ill with bovine spongiform encephalopathy, also known as mad cow, when they ate feed that contained the remnants of other cattle.

The U.S. stopped importing cattle from the U.K. in 1989 and in 1997 banned the use of feed containing cattle parts.

'The import of risk material from third countries seems largely ignored,' the peer review found. 'The (European Union) concluded long ago that lots of risk material from the U.K. was transported via third countries. Switzerland, for example, mainly got infected via France, not directly from the U.K.'

USDA's Ferguson and Harvard said the peer reviewers seemed confused as to the original goal of the Harvard analysis. That analysis was not supposed to look at the risk of mad cow entering the U.S. but rather to assume it would enter and then look at the probability it would become widespread.

'In some places the purpose of our report and what we were trying to do was perhaps misunderstood,' said George Gray, executive director of the Harvard risk assessment center. Harvard's risk assessment group revised parts of the analysis in a subsequent 2003 report for the USDA.

The peer review, by scientists in the U.S., London and the Netherlands, was compiled in October 2002.

Two of the five people who did the review said they had difficulty with the Harvard mathematical modeling because it lacked supporting data that could be used to establish credibility.

"We just don't know whether the model is correctly implemented or not," said peer review member Christopher Frey, an associate professor at North Carolina State University, who specializes in risk assessment.