Tuesday, 18 March 2014

Day 217: The variable views of Kenneth Mundt

There are some edge-of-the-seat moments at the Blais-Létourneau trial this week. It is really beginning to feel like play-off season - with every play, every call, every bench switch getting super-charged attention.

Justice Riordan has also upped the intensity, with a much more assertive refereeing than at any previous point in the trial. (Today he actually told Simon Potter to stop interrupting!)

Today was the second-last lap in the defendant's tag-team wrestling match against the conclusions of the plaintiff's expert epidemiologist (Jack Siemiatycki). This morning was the end of (epidemiologist) Kenneth Mundt's turn, and this afternoon statistician Bertram Price began to present his views.

Both these men are in the same line of business -- they are consultants to industries described by Mr. Mundt as having a "dark" past.  Asbestos. Roofing. Chromium. Tobacco. (Mr. Mundt works with Environ, and Mr. Price with his own eponymous consultancy).

Both are experienced expert defence witnesses, based in the United States and initially recruited by U.S. law-firms. Both were given the job of finding fault with Mr. Siemiatycki's estimates of how much one had to smoke before a subsequent illness in the Blais category (emphysema, or cancers of the lung, larynx or pharynx) could be attributed to smoking. (Mr. Mundt's expert report is Exhibit 30217, and Mr. Price's is Exhibit 21315).

Let's pick up where we left off yesterday -- with the cross-examination of Mr. Mundt by plaintiff lawyer André Lespérance.

Not your regular hatchet job?

Mr. Lespérance had seemingly rattled Mr. Mundt last night by showing him that he had once adopted or approved similar methods to those he is now criticizing.

Among those was a citation from an expert opinion written in early 2010 by Mr. Mundt, when he was engaged to help fend off a U.S. lawsuit involving light cigarettes. At that point, he expressed support for the use of pack-years as a measure of how much harmful exposure to which a smoker might have been exposed. But a few months later, in late 2010, when he filed his report in his case (Blais), he said that results based on pack-years were not up to the job.

This morning, Mr. Mundt looked and sounded more confident. He gave a nice-sounding explanation for these differing views in the same year.

He said his own thinking during that period had evolved as he became more "intimately" involved in the topic of exposure to cigarette smoke -- and underlined that the old way of thinking was inadequate. Thsi court would also "benefit from embracing the more recent" perspective.

He quite smoothly worked to keep on his main message track - and to reflect questions about his own choices back to the flaws he saw in the plaintiffs' measurements. "There are simple methods that are attractive because of their simplicity, but sometimes you need a more precise tool. A hatchet works for some jobs, a scalpel is required for others."

Some concessions

But while Mr. Lespérance was mostly unsuccessful at getting Mr. Mundt to agree with his interpretation of published results, he did succeed in having this epidemiologist say some useful things:

* Mr. Mundt said that a relative risk of over 1 would still suggest an association with exposure, even if it did not pass the "magic number" of RR=2 used in this trial. (Mr. Marais had made a great deal about not using any figures whose confidence intervals fell below 2)

* He validated the use of the cancer registry as an appropriate source for information.

* He described the growth of lung cancer as an "epidemic" largely due to smoking.

But no number - no matter how pressed for one

Mid morning, Mr. Lespérance asked the questions that have clearly been on Justice Riordan's mind. "What would be your critical amount. At what point would you be confident that on a population basis the number [indicating causality] is reasonable?"

Mr. Mundt steadfastly refused to give a number. He gave a long methodological rationale for his decision,
citing his desire not to "hamstring" the court, or to make assumptions or inputs that would make the conclusions speculative. "I cannot and I will not give a number."

Justice Riordan stepped up his pressure on the defence witnesses to produce a number. (Each of them has watched the testimony of the others -- so they know what is happening).

He told Mr. Mundt that in his role in this trial he should be prepared to do more than just criticize the numbers of another witness. "An expert witness in Canada comes to assist the Court in the job it has to do. I am looking to you for help, as I looked to Dr. Marais to help. Basically, I am getting 'you can't do it.' I am not getting help."

Mr. Mundt's response to the judge was a differently worded version of that given by Mr. Marais: an estimate or range of estimates can be produced, but only in a long and costly way. He repeated that the Siemiatycki method could not be made to work, and then offered the possibility of setting up a panel of experts to review the matter.

He bristled at the suggestion that he should have anticipated this need. "Had they (the plaintiffs) asked me, I would have done it. But it is hardly my role for this purpose."

There are none so naive as those who will not see?

With an agreement to end this witness by noon, the plaintiffs were racing the clock -- rushing their questions and growing impatient with Mr. Mundt when his answers grew long. (A marked - strategic? - difference from his laconic replies to Mr. Potter yesterday.)

Mr. Johnson had the second round of questions. As he often does, he raised issues that cast a poor light on Mr. Mundt's credibility. And, as he often does, he used analyses from U.S. scholars to connect this witness' testimony to documented industry tactics south of the border.

Almost an hour was spent detailing the role that Mr. Mundt had played in an effort to develop "good epidemiology."

Ong and Glantz
Mr. Johnston gave Mr. Mundt reason to question whether he might have been a willing or unwilling dupe of industry lawyers, showing an analysis of this exercise that linked the effort to a sophisticated tobacco industry strategy to try to avoid restrictions on public smoking.

Mr. Mundt showed neither surprise nor concern. To the contrary, he maintained that furthering good science - no matter at whose behest or on what side in court - was a worthy objective. "All clients have dark past," he said "including governments – we can't let that blind us."

Mr. Johnston gave him several opportunities to consider his work in light of concerns within his profession about the role of the tobacco industry. To each of these Mr. Mundt made it clear that he thought he was part of a respectable -- maybe even laudable -- attempt to improve scientific methods for decision makers.

A flexible witness

The morning had opened with Mr. Mundt being challenged to explain his differing statements to courts on the value of pack years as a measure of smoking. It closed with his being asked to explain a dramatic discrepancy between this week, and his views at a deposition in 2000. (The Falise case).

Fourteen years ago Mr. Mundt did not willingly agree that smoking and lung cancer were causally linked. Instead, he said that causality could not be inferred from observational data. He did not use the term causal without qualifying it: "At best we can make guesses about causality. So we temper it by saying potentially causal."

Although he remained poised and kept his voice on the level, Mr. Mundt seemed stressed by the questions. His face reddened, and his replies became obtruse and a little rambling.

He said that over the years he had changed the way he described things to better communicate in courts, but that the uncertainty he referred to then was nonetheless still part of to epidemiology. He appealed to "the philosophy of science."

Despite good intentions, Mr. Mundt was not freed to go until shortly after the noon recess. Two down, one remain.

A report on Mr. Price's testimony (which began this afternoon) will appear here tomorrow.

A juicy story about Mr. Mundt -- but not from this trial!

Sadly one of the other enlightening stories about Mr. Mundt never made it to this court record. 

Only last year, Mr. Mundt's work was the focus of an enforcement action by the EPA

Mr. Mundt had looked into the relationship between occupational exposure to hexavalent chromium and lung cancer among workers, and had found a dose-response in a study of four sites.  But the study had been repressed by Mr. Mundt's client, Elementis Chromium - and the EPA ultimately imposed a fine of more than $2 million on this company for hiding its results.

Mr. Mundt did not publish the study in its original form. Instead, he split his report into two separate studies -- neither of which had the "statistical power" to demonstrate the strong relationships evident in the four-site study.   They were individually published -- one for Germany, one for the United States. The weaker conclusions of the disaggregated studies were more favourable to the interests of his sponsors,

More on this story can be found in David Michael's book, Doubt is their Product or the Defending Science web-site.