The Canadian Medical Association grabbed headlines around the world this week by claiming that air pollution is killing citizens of Canada by the tens of thousands each year. Citing staggering new data, the association also predicted that such deaths will increase 83% in the next two decades. Fortunately, the CMA's research methods are fatally flawed and their alarming assertions largely meaningless.

According to association officials, the study was undertaken to quantify the national health and economic impacts of air pollution. There's value indeed in providing policy makers with facts necessary to allocate resources wisely. However, the association has wildly exaggerated the health effects of air pollution and inflated its economic impacts.

The CMA findings are the result of a computer model originated by the Ontario Medical Association in 2000. Major revisions were made to apply the software to all 10 provinces and three territories in Canada. But the likelihood of error in computer modeling increases in proportion to the multitude and complexity of variables in the algorithms. In this instance, the variety and scope of confounding factors has been increased exponentially.

Consider, for example, the sheer number and type of assumptions that would have to be made to predict not just mortality rates presumably related to air pollution, but also the number of visits to doctors offices and emergency rooms, hospital admissions and some 16 other specific categories of health effects experienced by 30 million people. For example, researchers assumed that the risk of illness would not vary among the provinces. Yet vast differences in geography, meteorology and economics — among a multitude of other factors — do combine to produce variations in exposure and risk across Canada.

Researchers also assumed that the level of pollution would hold constant for more than 20 years into the future. This is preposterous, of course. Just retiring older cars will considerably improve air quality. Emissions have declined dramatically in the past two decades, and will continue to decrease as regulation and wealth creation drive development of new technologies.

According to a 2005 Fraser Institute analysis of government air quality data:

• Ambient levels of sulphur dioxide decreased 72.2% from 1974 and 2001

• Ambient levels of particulate matter declined 50.7% from 1974 to 1999

• Ambient levels of carbon monoxide plummeted 82.6% from 1974 to 2001, despite a 30% increase in vehicle registrations during the same period

• Ambient lead levels fell 94% from 1974 to 1998

• Ambient levels of nitrogen dioxide decreased 34.4% from 1974 to 2001

Moreover, there's no uniformity to the transport and diffusion of emissions in a single day, let alone 20 years into the future. This infinite variation in the concentrations of both natural and synthetic chemicals in the atmosphere can dramatically alter one's level of exposure, as does the route and timing of contact.

Indeed, beyond the CMA's press release, the technical report states: The severity of the response to air pollution will depend on many environmental factors in addition to the health status and sensitivity of the individual.

The CMA's predictions about health effects aren't presented in the context of population growth, nor are the projected health costs cast in proportion to economic growth. Consequently, it is entirely possible that the health effects and costs, if any, could actually decline in real terms relative to population and economic growth.

The most problematic aspect of the CMA's research is the assumption that air pollution kills. But hundreds of studies on the health effects of air pollution offer mixed results. Even multiple studies of the same city have returned inconsistent findings. The uncertainty is magnified by the data mining employed in computer modeling. That is, only studies that conform to the CMA's a priori conclusions have been fed into the software. Consequently, the outcome is pre-ordained.

For all the publicity generated by the CMA report, it is curious that the association offers no policy recommendations. Thus, we are left to wonder about the association's motivation in undertaking this quasi-research. Environmental alarmists undoubtedly will seize on the results as proof of the need for more stringent regulation. But a slew of stricter requirements already are in the works. Whatever its intended purpose, the CMA report should be treated with the deference it deserves, i. e., none. As a product of ill-conceived methodology and faulty assumptions, the findings hold no value for policy makers or the public.