Wednesday, 19 March 2014

Day 218: Still no bridge to Tadoussac

The Montreal trial adjourned at noon today, after an intensive three-man six-day assault on the plaintiffs' epidemiological expert, Jack Siemiatycki.

There have been lots of dramatic and phoney-dramatic moments in this trial, but there has been nothing to compare to the past two weeks in terms of intensity or importance of discussion. (At least not yet!)

It may be a long time before the winner of this particular battle in this long legal war is declared. However Justice Riordan decides the issue, there is always the likelihood that the Appeal Court and Supreme Court will want to weigh in on how population-level harm is assessed or population-level blame assigned.

Introducing Mr. Price

Mr. Bertram Price was the third and last of the experts hired by the defendant tobacco companies to challenge Mr. Siemiatycki's estimates of the number of Quebecers made ill (with the diseases covered by the Blais suit) as a result of smoking. He was engaged by Imperial Tobacco and has also been retained by its parent company, British American Tobacco.

Mr. Price began to present his expert report (Exhibit 21315) mid afternoon yesterday, and continued this morning. He was on the stand for less than a day.

Mr. Price is an old hand at the numbers game - and not only because he is one of the more chronologically advanced of the expert witnesses (he is 74 years old).

For almost three decades he has managed his own consulting firm, Price Associates, which specializes in using statistics to assess risk in occupational and public health. By his resumé (Exhibit 21315.1), much of his work has been focused on asbestos. This is the first time he has appeared as an expert at a tobacco trial.

At the request of Imperial Tobacco's counsel (Allan Coleman), and over the objections of the plaintiffs, Mr. Price was qualified as an "expert in applied statistics, risk assessment, the statistical analysis of health risks, the use and interpretation of epidemiological methods and data to measure statistical associations and to draw causal inferences." 

Mr. Price is not exactly a new face in this court room: he had attended the testimony of Mr. Siemiatycki last year and was sitting a few seats from me earlier this week as Mr. Marais and Mr. Mundt gave their views on the model used by the plaintiffs' expert. It's hard not to notice when Mr. Price is in the room -- he is several inches taller than 6 feet in height. (Tall enough to have played in the NBA - for the Detroit Pistons.)

Mr. Price may even have had as much court-time as Mr. Allan Coleman, the Toronto-based lawyer assigned to work with him. Mr. Coleman's sole assignment in this case seems to be to work on the epidemiological issues. His pleasant manner and level pace made it an easy-to-follow exercise.

Third on the match 

Mr. Price spoke on the heels of two other long and detailed critiques, but nonetheless had something new to offer.

He suggested another level of analysis that could and should be done: to remove from the body count those people who might have become sick because of smoking, but not because of any misconduct on the part of the companies.

To illustrate this concept, Mr. Price was asked to reflect on the circumstances of a few hypothetical smokers, like that of a man who smoked a pack a day from 1962 until he was diagnosed with cancer in 2000.

If the court ruled "that any smoking after 1963 is not attributable to the defendants' misconduct," Mr Price felt that he would have "only smoked for 2 pack-years in the misconduct period" and therefore not be eligible for a screening which required 5 pack years. (1964 is the year of the Surgeon General's report, which the defendants maintain is the point at which "everyone knew")

And what if the court ruled there was misconduct only until 1972? In such circumstances, Mr. Price looked at the difference in relative risk for having smoked 12 years (in the misconduct period) and having smoked 28 years (in the non-misconduct period). He noted that it did not double, although he did not say what values he was using.

Therefore, he suggested, this smoker should also not be considered eligible under the Siemiatcyki approach.

An encouraging moment for the plaintiffs

This suggestion prompted one of the strongest comments from the bench to date - and one of the few that suggests so clearly that Justice Riordan has an opinion formed on the topic of addiction. The judge asked Mr. Coleman to confirm that in his hypothetical example "this fellow ... was smoking because of the misconduct of the company," before stating a question that he will be in a position to decide:

So isn't his smoking, morally and logically and medically attributable to the whole period of his smoking? How does it stop all of a sudden in 1972 after he's addicted to cigarettes?  

As if that weren't enough to make the plaintiff's leave the courtroom happy (the clock was running out on the day), the Judge also hinted at his views on misconduct. He responded to Mr. Coleman's suggestion that some people may have started smoking before any misconduct, i.e. in the 1940s, by saying "In the 1940s it's probably an easier argument to say there was no misconduct on the companies' part than later."


Overnight, Mr. Coleman must have decided to pursue the idea of "misconduct" no further. Without this element, there was little in Mr. Price's report that had not already been discussed at length by the previous men.

Like those before him, Mr. Price had criticized Mr. Siemiatycki's approach for not having included more information. He felt that in addition to individual smoking histories it was important to know which brands people smoked, what other risks they were exposed to, and whether or not the smoker believed what the companies said about smoking. None of this information was in the analysis and therefore, he concluded, "the Siemiatycki report is not adequate for this purpose." 

Mr. Coleman invited his witness to consider whether there were better methods already established and available to be applied to this problem. (The first two industry experts had been more than reluctant to suggest that the task could be accomplished without considerable additional effort).

Mr. Price gave the impression that there were models available that were better suited for the purpose than the "too simple" dependence on pack years, or the limitations imposed by meta-analysis. Cox Models. Logistic regression models.Survival models. He noted that other scientists had been able to include other variables, such as smoking intensity (Exhibit 20222), or the effects of quitting (Exhibit 20223),

Mr. Trudel was on his feet more than once over the morning, pointing out that Mr. Price should be limited to testifying what was included in his report, and could not introduce new topics. (There are rules, you know!)

Mr. Coleman seemed stuck between the rock of the scope of his witnesses' report and the hard place of Justice Riordan's clear hints that more was needed.

Navigating these constraints, Mr. Coleman nonetheless gave Mr. Price the opportunity to sketch out the approach that he thought could be taken. He identified several steps - the development of a model, the establishment of ways to measure of exposure, the selection of parameters for analysis, the search for existing data and the generation of new data, the application to individuals or groups of individuals.

His option was more detailed than that offered by Mr. Mundt (who had recommended an expert committee) or Mr. Marais (who said a new survey of many thousands of people would be required), but it was not necessarily one that was more readily accomplishable.

Justice Riordan again made clear that he needed something more concrete to work with, drawing on a local metaphor."Everyone says you can build a bridge over the river at Tadoussac, but no one has been able to do it."  

The Toronto-based Mr. Coleman and the Maine resident Mr. Price looked confused by the reference. Justice Riordan explained to the come-from-aways that the highway ends at Tadoussac, and ongoing traffic must use a ferry to cross the Saguenay river.

"I understand Maitre Trudel's comment. It's nice to say you can have it, but nobody does it. They don’t show anything in the literature that says it has been done. So what am I supposed to take from that?"
Minutes later - only an hour after the morning start - Mr. Coleman ended his question.

The cross-examination

The plaintiffs took the same approach with Mr. Price as they had with Mr. Mundt  -- using his own publication record to show that he had at one time supported the analytical approach or principles he was now criticizing.

Given that Mr. Price had watched the trial over the past week, he might have been better prepared for these questions. As it was, he seemed bewildered by his former research, as he was by the questions about it. Confused too, perhaps, by questions expressed in a French accent and using a syntax more common in Montreal than in Maine. "I don't understand the question"  was the frequent reply.

His halting and occasionally conflicting replies may not have affected Justice Riordan's view of his testimony, but it really slowed things down!

One of the first sets of questions put to him by Mr. Trudel regarded mesothelioma and its relationship to asbestos (a subject much studied by Mr. Price). In his own study on how to apportion risk in this area (Exhibit 1715), Mr. Price had not factored in the effect of a potentially confounding risk.

Despite repeated questions by Mr. Trudel (and reminders that he was at the court as an "expert in the inference of causation), Mr. Price steadfastly refused to state a threshold value at which he would infer causation. "There is no bright line cut-off," he said. "No magic number. No normal number." But in his own work on asbestos and lung cancer (Exhibit 1716), he used the same threshold value as the one identified by Mr. Siemiatycki -- 50%.

Mr. Lespérance used the opportunity of cross-examining Mr. Price to sweep away some red herrings. The industry experts had suggested that Quebec smokers' exposure to radon, asbestos and other harmful toxins might be responsible for a large proportion of cancers.

He asked Mr. Price to compare the risks from smoking with those from asbestos -- showing him a recent paper with updated estimates.(Exhibit 1718). Mr. Price confirmed that in this study the relative risks for lung cancer among current smokers (26) and former smokers (6) were all above the threshold value of 2.

How did they compare with risks for asbestos or radon? "The risk with asbestos for lung cancer is smaller than those numbers," Mr. Price said grudgingly, and that for radon "would not be greater."

By not acknowledging the huge gap in risks, or sharing his knowledge about asbestos, Mr. Price opened the door for Mr. Lespérance to rub his nose in them. Only after 20 years of "substantial exposure" to asbestos did the risk exceed 2. (Exhibit 1719). And the number of Quebec homes where radon exposure exceeded the threshold risk level was only around 3,300. (Exhibit 1720).

Soon it was finished, and Mr. Price was thanked for his time and invited to step down.

The defence strategy of offering no positive evidence on epidemiology may prove to be a good one in the long run. But in the short-run of this week, it certainly drew some fire!

Tomorrow, back to some important legal wrangling.

This post was back-timed for consistency in indexing