Tuesday, 21 January 2014

Day 201: Proving the existence of rational smoking

At the beginning of the week, before Kip Viscusi began his testimony at the Montreal tobacco trials, my colleague Pierre leaned over to share a joke about economists.
Why is economics a theology and not a science? 
Because if it were a science, they would have tried it on animals first.

Mr. Viscusi's second and last day of testimony was given over to his cross-examination by plaintiff lawyers André Lespérance and Philippe Trudel. It often resembled a theological debate. 

This impression of religion-over-science was only somewhat aided by the church-like structure of the courtroom, with its centre aisle and the bar dividing onlookers in the nave from the officiants in the chancel. The medieval clerical robes worn by lawyers and judge didn't hurt either.

What really gave it the air of a debate among Jesuits was philosophical thrust of the questions thrown at Mr. Viscusi, and his apparent willingness to have all facts serve to prove, if not the existence of God, then at least the infallibility of the consumer market. 

For someone who was being put through the epistemological wringer, Mr. Viscusi looked not too bad at the end of the day. He kept to his script and kept his composure. But his conclusions looked considerably the worse for the exercise, with the stuffing knocked out of some of his core assertions. 

Act global. Be fuzzy on the local.

One of the more awkward moments for Mr. Viscusi came late morning, when Mr. Trudel began his round of questions with a focus on when Mr. Viscusi had been engaged, what he had been hired to do and who had been his contact.

Again, this witness seemed unrealistically unable to provide details about his activities. Yesterday's statement that he had testified for "10 to 15" trials was a more precise answer than he seemed able to muster for questions about the process of writing this report less than a year ago.

He wasn't sure when he had first been approached. He wasn't sure if he had received a written mandate from the companies. He seemed unsure of who had framed the questions he would answer until it became clear during discussions that it would be better for his clients if it were lawyers who had done so. At that point he remembered that in fact it was the lawyers.

He knew he had been contacted by John Blain of Freshfields, but wasn't sure which company Mr. Blain represented. (Freshfields is based in the United Kingdom. It's client is JTI-International. Mr. Blain has been previously spotted in this court, during the first appearance of Mr. Hoult). "I don't keep track of clients," Mr. Viscusi said.  

Several times he was unable to say why or how certain documents were included or excluded from his review. Some came from the report presented by Raymond Duch last year, others (like findings from Imperial Tobacco's Viking report) seemed to have appeared in other work prepared by Mr. Duch.

Court rules mandate that source material for expert reports be shared with opponents - so the fact that Mr. Viscusi could not recall how he received source material, or what he might have done with it was significant. Even more fishy-sounding was his explanation that he destroyed all the electronic records that were provided to him by the lawyers. He said this was because he was not allowed to keep too much material on his laptop, and had a "modest allotment in terms of what we are allowed to keep on our computers" by his university. You could almost hear his credibility bang down a few notches with that comment. 

Unravelling the knotted arguments

Although it is padded with many comments on the way people respond to risk communication, the core of Mr. Viscusi's expert report (Exhibit 40494) is a tightly bound theory.

Smokers make rational decisions to smoke. They make these decisions based on their perceptions of risk and benefit. Their perceptions of the risks of smoking are much greater than the true risks. Giving them more information about the risks of smoking would actually reduce their perceptions of risk and would therefore lead to more smoking. 

Mr. Lespérance and then Mr. Trudel challenged this theory by poking at its assumptions and embedded components, and showing where these were inconsistent with other evidence at this trial. Do smokers over-estimate the risks of smoking? Are Mr. Viscusi's statements of the "true risks" accurate? Do smokers really not benefit from warnings? 

They also challenged the credibility of the man by asking questions that exposed his unwillingness to provide frank answers about the events that brought him to this courtroom, and about issues that were excluded from his report.

At times, the exchanges were tense and intense - the questions put with incremental forcefulness and a witness who frequently feigned misunderstanding. More than a few times, it was only after Justice Riordan repeated questions that Mr. Viscusi responded. His squirming away from answers may have left a more lasting impression than the answers he gave.

The few occasions that Mr. Pratte intervened with objections provoked uncharacteristic rebukes from Justice Riordan, who spoke more harshly against "talking objections" than he previously has. (Such objections are disguised hints to the witness).

Do smokers over-estimate the risks of smoking? 

Mr. Lespérance showed Mr. Viscusi measurements of smokers' perceptions of the risks of smoking, including some that were undertaken for Imperial Tobacco. (Exhibits 987.12, 987.6, 1547.1, 40064.61). 

These reports have been much talked about at this trial, but it would appear that no-one informed Mr. Viscusi of them. This lack of familiarity did not stop him from quickly identifying methodological flaws that would explain why all this research on Canadian attitudes and beliefs came to different conclusions than his own.

Mr. Viscusi acknowledges that every cigarette smoked presents a risk - "there is no level of smoking that proses zero risk to all risks." But the finding from Imperial Tobacco's long-running survey that 80% of smokers thought it was possible to smoke three or more cigarettes without risk did not suggest that smokers underestimated this danger. The question was flawed - it was not "probabilistic."

Similarly, the finding that fewer than half of respondents believed that smokers died at a younger age than non-smokers was invalid in his view. It was a "badly worded question – it should be worded, like my questions, in terms of life expectancy."

A 1990 Environics report which showed smokers unable to state many of the diseases caused by smoking was criticized for being an open-ended question format. "That's different than taking people through laundry list. People say one or two things to get the surveyor off their back. It doesn't  mean that people are not aware of the risks, it just means they didn’t bother to mention them." 

Imperial Tobacco's surveys found that smokers believed that they only risked losing a couple of life years to smoking. (Exhibit 987.21) Again, he found fault with the survey methods. "This is not a reliable way to ascertain life expectancy." 

Did Mr. Viscusi underestimate the real risk of smoking?

The fundamental  underpinning of Mr. Viscusi's argument is that health authorities place a lower risk on smoking than smokers do. He cites the Surgeon General as the source of the "true estimate" that between one-sixth and one-third of smokers die as a result of tobacco use, and the Institutes of Medicine that the average smoker actually loses 6.6 years of life. 

He struggled, however, to remember exactly how he got those numbers, or to produce the exact sources from the American authorities. He had a much clearer memory of the flaws of other researchers who reached different conclusions. These included the conclusions of Doll, Peto and others who concluded that one-half of smokers died from tobacco use, and those who did lost an average of 15 years of life. (Exhibit 1545, 977)

Addiction and the rational decision to smoke.

Mr. Viscusi's report does not talk about smokers' knowledge of addiction, or the potential impact of warnings in this area. Given that the Létourneau half of this joint-trial is solely about addiction, it was a curious oversight.

To Mr. Lespérance's questions of why it was excluded, Mr. Viscusi explained that he "did not know it was part of the case."  Moreover, it was not "not a probability of health outcome – it is a property of cigarettes."
He agreed that smoking was addictive -- but that this only happened after 1988, when the Surgeon General labelled it so. He was unwilling to state whether addiction itself could be considered a disease -- only, he said "if it is something that caused death… if addiction to cigarettes independent of any other risks killed you, it would be included."  

Addiction does not seem to figure prominently in his model of the rational decision to smoke. This decision, he says, is made every time a smoker lights up. Addiction plays a role in so far as it increases the costs of quitting. In these cases, the smoker makes "a rational decision that you are not willing to put up the costs of giving up compared with the benefits."
Against these costs the smoker would consider the benefits, and Mr. Viscusi acknowledges that "nicotine is one of the benefits of smoking to smokers." He resisted the use of the term "fix,"  and bristled at comparisons between heroin and nicotine. 
He also resisted the idea that smokers might "lose control". When shown the ruling of the Supreme Court in the Insite decision that supported the understanding that "addiction is a disease in which the central feature is impaired control over the use of the addictive substance," he began to parrot the explanation that this court has frequently heard from tobacco industry witnesses. 

"More people who have quit smoking than those who current smoke. Whatever addiction there is does not mean it is impossible to quit."  He soon went a step further, saying that those who quit are willing to incur the costs of doing so, and those who continue to smoke do so "because they like smoking."

His nonchalance contrasted with the last exhibit shown by the plaintiffs -- the lament of then VP-marketing Bob Bexon, who saw troubles ahead for Imperial Tobacco because the only remaining benefit to smokers was "the “drug” effect of nicotine. ...  Because they are chemically induced through nicotine smokers see these positives as reflections of their own personal weaknesses. They get them but they are not happy about it.  If our product was not addictive, we would not sell a cigarette next week in spite of these positive psychological attributes." 

Mr. Viscusi would not bite. He showed little reaction as he repeated his view that smokers understand the benefits of smoking. "It just means that this is what they enjoy. The reason they smoke is the drug effect." 

Effective communication 

Mr. Viscusi was adamant that people cannot process risk information that is expressed in numerical form. "Things like '40,000 people die from smoking.'"  The reason he gave was that people have to imagine the denominator. Giving a number out of 100 is useful he said -- but when the proportion is smaller than one-hundredth, difficulties arise.

He was similarly adamant that the concept of relative risk was inappropriate as a method to inform smokers. He defended his use of used data from a study of young American smokers that showed a high level of belief about deaths from lung cancer among smokers (six times greater than the "real risk"), while leaving out that the same youth had an even greater overestimation of the risk of death from lung cancer among non-smokers (more than sixty times the real risk). 

"The public doesn’t need to know the relative risk. They need to know the absolute probability. Relative risk is not meaningful." When Mr. Lespérance pointed out that he had stressed the importance of communicating relative risk in his arguments against plain packaging which he prepared for the CTMC (Exhibit 1680), he said it was a "different issue. I was talking about relative risk of lower or higher tar cigarettes."

Warnings don't make a difference

In a trial about cigarettes and tobacco marketing, there have been very few times when cigarette packages have actually been displayed or handled. This afternoon, Mr. Trudel showed Mr. Viscusi a few current cigarette warnings, and asked him to open a carton of cigarettes and comment on the warning he randomly selected. (A good proxy for the experience of a smoker!). 

Although he acknowledged the ways in which these warnings provided informational or sought to change the social acceptability of smoking, he refused to admit that they might reduce smoking rates. The plaintiffs did not, curiously, ask him to comment on recent Canadian studies which have measured the benefit of these new labels, i.e. Azagba et al. and Huang et al.

The Final Words

At the end of the day, Mr. Guy Pratte (JTI-Macdonald) and Mr. Simon Potter (Rothmans, Benson and Hedges) were able to ask their witness questions that allowed him the opportunity to restate their theory of warnings -- that they would have provided no benefit and would possibly have resulted in more smoking.

More health information "from a credible source would cause people to lower their risk beliefs," Mr. Viscusi obligingly repeated. The effect of such aims at deterrence would "decrease the expected cost of smoking, and thus increase smoking prevalence."

Mr. Viscusi faced two more questions on this theme:  Mr. Lespérance asked him what the result would be in a scenario where smokers actually underestimated the risks of smoking, but were then told that they were overestimated these risks. Mr. Viscusi acknowledged that if they were "convincingly told", the result would be "reduced risk beliefs and increased smoking behaviour."

Justice Riordan asked him to compare the likely impact of the specific warnings that were required by the government after 1989, and the general warnings that the industry voluntarily placed on packages in the preceding 17 years. "If you put them side by side at the same time, the specific would have more impact?" he asked. "Without question," said Mr. Viscusi.

Schools of economic thought  

Economists' capacity for debate is almost as legendary as that of lawyers. Both may have been at work today. In addition to Mr. Viscusi, there were two other trained economists at the centre of much of the day's debates.

Much earlier in this trial, Justice Riordan had revealed that he had once studied economics. (As I remember the way he said it, he also strongly deprecated any claim to expertise in the domain!). A quick search on the Library of Canada dissertation data base turfed up the thesis he completed in 1972 to qualify for a masters in economics from McGill University.

The topic was Canada's foreign aid, and what could be done to make it more efficient and more generous. A far cry from Kip Viscusi's concurrent neo-liberal formation at Harvard!.

A decade later, Mr. André Lespérance also graduated with a masters in economics. His was from the University of Montreal, and completed after he completed his law studies.

A change of topic tomorrow. Psychiatrist Dominique Bourget will testify on addiction. She is the first of three expert witnesses scheduled by the defence on this topic.