Science and secondhand smoke: the need for a good puff of skepticism.(Report)

THERE IS NOTHING MORE POWERFUL THAN A lie whose time has come. Thus, the smoking bans.

The experts proclaim that 63,000 Americans are killed yearly by secondhand smoke, more than the victims of AIDS, drank drivers, the Iraq war, and Hurricane Katrina, put together. If it's true here, imagine the death toll in China.

One would think such a strong assertion would be followed by demands for strong evidence by the scientific establishment, the medical community, the media and the man in the street. Instead, with very few exceptions, scientists and doctors have remained silent, the media have led the orchestra and the citizens have waltzed to the music.

But the claims about the deleterious effects of second hand smoke are based on nothing more than cooked statistics--there are no bodies, no autopsy reports. But they have led to draconian smoking bans imposed by governments from California to New York to Ireland to Israel to Australia to England. In the process, civil liberties have been trampled and smokers demonized, driven into the streets and lately, in some places, off the streets as anti-smoke zealots promote the notion that outdoor smoking is virtually as insidious as indoor smoking.

The blueprint for this campaign dates to 1975 when British delegate Sir George Godber instructed the World Health Organization on how to get smokers to quit. (1) As reported in "Passive Smoking: How Great the Hazard?", Sir George said, "it would be essential to foster an atmosphere where it was perceived that active smokers would injure those around them, especially their family and infants or young children who would be exposed involuntarily to the smoke in the air." Eleven years had passed since the U.S. Surgeon General stunned the tobacco industry with a 387-page report linking cigarette smoking and lung cancer. It was an historic event, ranking among the top news stories of 1964. (2) But people weren't kicking the habit because they didn't think cancer would happen to them. Sir George understood what it would take to overcome this mindset. Make the nicotine addicts believe they delivered death to innocent bystanders.

For the mother of all guilt trips to take hold, what was needed was an official imprimatur. The U.S. Surgeon General delivered it in 1986 with a report concluding that secondhand smoke "can cause lung cancer in nonsmokers." The data "suggest" that nonsmokers are exposed to levels of environmental tobacco smoke (ETS) that "would be expected to generate a lung cancer risk," wrote Surgeon General C. Everett Koop (emphasis added). Put together it spelled, "maybe," but two paragraphs later, Dr. Koop wrote: "It is certain that a substantial proportion of the lung cancers that occur in nonsmokers are due to ETS exposure." (3)

This sleight-of-hand went unnoticed by the popular media, which promoted the report as truth itself. But throughout the next decade, experts repeatedly criticized Dr. Koop's conclusions. After reviewing the Surgeon General's report, the International Agency of Research on Cancer (an offshoot of the World Health Organization) concluded that, as far as the risk of lung cancer was concerned: "The observations on nonsmokers that have been made so far are compatible with either an increased risk from passive smoking or an absence of risk." (4)

Dr. Ernst Wynder, president of the American Health Foundation, a pioneer who in 1951 had connected active smoking and lung cancer, asked a question unanswered to this day: "If passive inhalation in fact increases our risk of lung cancer, there should have been a steady increase in the incidence of lung cancer among nonsmokers. This would have been observed in the female population for the last forty years. As more and more men smoke, more women passively inhale the smoke of their husbands. [Yet] there has been no significant increase of lung cancer in male or female nonsmokers. In this case we have a non-fit." (5) Exactly. Data from national mortality surveys show that lung cancer rates among never-smoking women remained stable between the 1950s and the mid-1980s; lung cancer rates among women didn't rise until decades after women started smoking.

Why this emphasis on married women? Because the statistical studies relied upon by the Surgeon General (and ever since) were mainly based on the incidence of cancer among nonsmoking married women who lived with smoking husbands. As compared to nonsmokers married to nonsmokers. Linda Stewart, in her seminal article "How to Read a Study," lays out the flaw in this approach:

 
      All it does is, it counts things up, and then 
   relates one set of numbers to another: 
      In a group of 10 people, 6 have a cold. 
      Of the 6 who have a cold, 3 own a cat. 
      Of the 4 without a cold, 1 owns a cat. 
      Epidemiology can then give you a formula: 
   People who own cats have twice the rate of colds. 
   What it can't do is offer a connection: "Cats cause 
   colds." (6) 

Which brings us to the first role of epidemiology: Correlation does not prove causation. The late Dr. Alvan Feinstein, for 40 years Sterling professor of medicine and epidemiology at Yale, testifying before the Congress on secondhand smoke, related a challenge he gave his seminar each year: "Go into the statistical abstract of the United States ... and pluck out data to support the most outrageously silly contention you can come up with." The winner was somebody who found a strong statistical relationship between the sale of VCRs and the incidence of AIDS. "A wonderful statistical relationship," said Dr. Feinstein. "And then if you want plausibility, well, what are they doing while they're watching VCRs, and so on and so forth." (7)

If Feinstein had published this "proof" that VCRs caused AIDS, he'd have been …