According to a recent report by the Ontario Medical Association, nearly 6,000 Ontarians die each year from air pollution and tens of thousands more are sent to hospital. But the OMA was able to manufacture its scary death toll only by ignoring weaknesses in studies that supported its predetermined conclusions and excluding countervailing evidence.
Rather than providing a sound basis for Canada's public health policies, the OMA has shown what happens when health experts choose the warm glow of feel-good activism over critical thinking about the real factors than affect human health and welfare.
Researchers have known for decades that very high air pollution levels can kill. Thousands died during the 1952 "London Fog," when pollution soared tens of times higher than the highest levels experienced in Western countries today.
Nowadays, air pollution is far too low to cause obvious acute increases in mortality, and doctors can't tell if any particular death was caused by air pollution. Instead, the evidence for air pollution health effects comes from statistical analyses that try to tease out any effects of air pollution from a thicket of other health-related factors, such as weather, diet and physical activity.
Based on these statistical studies, the OMA claims about 5,800 Ontarians will die prematurely this year due to air pollution, rising to more than 10,000 in 2026. The OMA attributes most of the deaths to long-term exposure to fine particulate matter (FPM), based mainly on a study known as the American Cancer Society (ACS) study. The ACS study followed hundreds of thousands of Americans for 16 years and looked for correlations between FPM levels and risk of death.
Although the study reported that higher FPM was associated with greater mortality risk, the OMA ignored the study's biologically implausible results. For example, the ACS study reported that FPM kills men, but not women; those with no more than a high school education, but not those with at least some college; former smokers, but not current or never smokers; and people who say they are moderately active, but not those who say they are sedentary or very active. Results like these indicate the study hasn't uncovered a genuine cause-effect relationship.
The OMA also ignored a separate long-term study of 50,000 veterans with high blood pressure. Although this group should have been more susceptible to air pollution than the general population, the study reported no mortality risk due to long-term exposure to FPM or other pollutants.
The OMA attributes about 1,800 deaths each year to daily fluctuations in air pollution levels. However, studies of air pollution and daily mortality have their own statistical challenges. British scientists recently evaluated uncertainties in these studies' statistical models, fortuitously with Toronto data. With a combination of scientific and British understatement, they concluded "statements of the form: 'ozone has no effect on mortality' receive [the] most support from the data." They drew similar conclusions for other pollutants. Weather was the only factor with a statistically significant relationship to mortality.
The OMA claims air pollution will cause 17,000 hospital admissions and 60,000 emergency room visits this year. But similar methodological concerns apply here as well. The OMA was also once again selective about which studies it used for its estimates. For example, for ozone pollution OMA sources its estimates to a California Air Resources Board review of ozone health effects. But CARB's review excluded several studies that failed to find any harm from ozone exposure. Data on asthma attacks should also create skepticism about the ostensible health effects of air pollution. Hospital visits for asthma are lowest in July and August when air pollution is highest.
After creating thousands of fictional cases of death and disease, the OMA compounds its transgressions by making it appear that air pollution health risks will increase in the future. For example, the OMA estimates that among people over 65, air pollution deaths will rise from about 4,800 in 2005 to 8,900 in 2026. This gives the impression of an 85% increase in the risk of dying from air pollution. What the OMA fails to mention is that the total number of people over 65 will also increase--by almost a factor of two according to Ministry of Finance projections. If the population rises by 95%, but pollution deaths rise by only 85%, then the risk of dying from air pollution must decrease with time--just the opposite of what the OMA implies.
Health professionals normally talk about disease trends in terms of age-adjusted rates, rather than absolute numbers. This nets out any trends in the size and age of a population, revealing real trends in health risks. The OMA might respond that its report is about trends in the health care costs of air pollution, which requires an estimate of absolute numbers of deaths and hospitalizations, rather than just rates. But such a claim would still be specious. All else equal, all needs will increase with an increasing population, as will the total aggregate income available to provide for those needs. The policy-relevant question is what will happen to the age-adjusted per-capita health care costs of air pollution. The OMA's own estimates imply these costs will decrease, but the OMA report asks the wrong question and obfuscates the answer to the right one.
Unlike current air pollution levels, extreme weather clearly does kill. Toronto Public Health (TPH) estimates that heat killed an average of 120 Torontonians per year between 1954 and 2000. Based on this past relationship, and presumed global warming in the future, TPH projects a tripling of heat-related deaths by 2080. TPH also claims increasing temperatures will increase air pollution.
Like the Ontario Medical Association, TPH reports only changes in the absolute number of deaths, rather than age-adjusted rates. But TPH's results suffer from more serious flaws. Even if we assume that temperatures will rise substantially in the future, past experience suggests that heat-related mortality will nevertheless decrease. Between the 1960s and 1990s, average summer temperatures rose nearly 1 C in U.S. cities, but the risk of dying from extreme heat declined 75%. The improvement was due to better health care, better warning systems and, most critically, increased availability of air conditioning. TPH doesn't discuss trends in heat-related mortality in Canada but they are likely similar to the United States.
TPH is also wrong about future air pollution. Already-adopted requirements will eliminate most remaining pollution emissions over the next few decades. For example, Environment Canada estimates that the average automobile and diesel truck will be more than 80% cleaner in 25 years, due to progressively tougher emission standards adopted in the last decade. All else being equal, higher temperatures are associated with higher ozone levels. But large upcoming declines in ozone-forming emissions ensure that, at worst, any climate change that does occur will merely slow the rate of improvement. For fine particulate matter, the prognosis is even rosier. Much FPM is "semi-volatile" material that evaporates as temperature rises. As a result, to the extent global warming occurs, it will reduce FPM levels.
The formulation of sound environmental and energy policies depends on the input of doctors and scientists who can think deeply and objectively about the real factors that affect people's health and welfare. Unfortunately, with their tendentious and misleading analyses, both the Ontario Medical Association and Toronto Public Health have crossed the Rubicon from science and critical thinking to fear-mongering and political activism.
Joel Schwartz is a visiting fellow at AEI.