Bushs Science Advisor, Congressman Clash Over Computer Models

President Bushs science advisor and a prominent Democratic Member of Congress sparred publicly this week about the administrations stance on computer models.

Representative Henry Waxman accused the administration of invoking scientific uncertainty to discredit evidence counter to its policies.

John Marburger, science advisor to the president, responded that the computer models used to make predictions about climate change and public health were both ambiguous and prone to group-think errors by scientists.

The clash came during addresses to science journalists in a meeting preceding a large scientific conference sponsored by the Association for the Advancement of Science in Washington, D.C.

Health and climate are complex, influenced by thousands of inter-related factors. To predict how current practices predict the future, scientists develop computer models. Given past data, current trends and sets of mathematical relationships, these tools make educated guesses about the future.

Computer models are used to inform many public health and environmental policies by forecasting, for example, how different vaccination programs might slow the spread of disease, how pollution levels might affect the climate, or even how educational policies might change birth and death rates.

/zimages/2/28571.gifTo read how the FDA is using computer models,click here.

But because of lack of computational power and information, models cannot encompass all the complexities of the real world. Most models of climate change, for example, cannot model individual clouds and so estimate cloud coverage over larger regions. And different assumptions built into the model can dramatically affect the results.

Waxman said this liability has been exploited, particularly in a recent controversy about safe levels of mercury in the environment.

“Scientists were told keep remodelling until you get the answer we want,” Waxman said. But he also accused the industry and administration of hiding behind ambiguity when model predictions run counter to policy goals.

A recent data quality act allows vested interests to attack data generated by modeling. Waxman said it was being used as a device to stall regulation.

An article in the Washington Post says that industry groups have used the law to protest dietary recommendations for acceptable levels of salt and sugar, toxic levels of heavy metals, and acceptable logging rates.

Marburger admitted, “There probably are industries that would like to keep arguing” just to keep acceptable levels of substances in food or the environment undefined, but accused Waxman of trying to decrease scientific scrutiny of models. Ultimately, he said, court decisions would be necessary to decide how seriously to take predictions from modelling.

“In fact, there is ambiguity,” Marburger said, adding that models can be used to generate bad information.

In an accusation reminiscent of those launched against the Bush administration, he said that scientists can be misled by their own assumptions and need to be challenged.

“Large coalitions can develop opinions not supported by evidence,” Marburger said.

/zimages/2/28571.gifCheck out eWEEK.coms for the latest news, views and analysis of technologys impact on health care.