A new weapon against the tobacco industry : class action lawsuit using a novel epidemiologic parameter

Epidemiologyand public health

Lawsuits against thetobacco industry, if successful, have the potential to compensate victims ofsmoking and to diminish the capacity of the tobacco industry to continue tofunction with impunity. Up to now, the only successful legal actions againstBig Tobacco have been those brought on behalf of states or provinces to recovernational health care costs associated with tobacco diseases, and those broughton behalf of individual victims, in which financial damages were sought.

Class action lawsuitsto recover damages on behalf of the huge numbers of victims have not previouslybeen successful. The problem is that the tobacco industry has successfullyargued that such a lawsuit requires a demonstration of “more likely than not”causation for each plaintiff. Depending on the jurisdiction, there may be manythousands or millions of incident cases annually, and it is impossible to bringthem all into court individually to determine whether smoking was “more likelythan not” a contributing cause of each case.

Fifteen years ago,class action suit was launched in Quebec on behalf of all lung cancer patientswhose disease was caused by cigarette smoking. The plaintiffs’ lawyers asked meto estimate how many cases of cancer were caused by smoking. This sounds likethe classic attributable fraction in epidemiology. But upon refinement, thequestion became one that has not previously been addressed in epidemiology:What proportion of lung cancer cases in Quebec, if they hypothetically could beindividually evaluated, would satisfy the “more likely than not” criterion?

The novel methodologyI developed is based on two stages. First I estimated the dose-responserelationship between smoking and lung cancer, for which I use the pack-years asa measure of smoking. I define the amount of smoking that is required to inducea two-fold risk as the “critical amount” that makes it more likely than notthat smoking contributed to the cancer. Depending on the model used andconsidering statistical variability, I estimated that the critical amountrequired to double the risk was somewhere between 5 and 12 pack-years ofsmoking.

Then I had toestimate what fraction of Quebec lung cancer patients had smoked more than thecritical amount. I estimated this to be over 90%.

The Quebec trial, atwhich I testified and was cross-examined at length and criticized by a paradeof defense witnesses, ended in 2014. The judge has made his judgement. Hesupported the plaintiffs. I will describe some aspects of the trial.

Cette présentation a été donnée dans le cadre du BRIO SIRICscientific day 3 organisé annuellement par le SIRIC BRIO et qui a pour but deréunir tous les acteurs du SIRIC BRIO et plus largement de la cancérologie àBordeaux.