Tuesday, October 20, 2009

IOM Proceedings Introduced Bias into Report: Only One Side of Smoking Ban Studies Was Presented in Committee's Public Meetings

Today, I reveal that another element of bias in the Institute of Medicine's report on the effect of smoking bans and heart attacks occurred in the proceedings that the Committee undertook in reviewing the relevant data.

The Committee held a public meeting in which it heard presentations by experts in the field covering various topics. According to the report, the topic of smoking bans was only presented by one expert: Dr. Stan Glantz.

The Rest of the Story

Dr. Glantz has a very particular view of the smoking ban studies. I have no problem with the Committee choosing to ask him to present his side of the issue. However, to achieve balance and entertain alternative hypotheses, the Committee should also have invited someone to present the other side of the issue.

If you only hear one side of the issue, then of course you are going to get a biased picture of the data out there. This could easily result in bias in the Committee's analysis and report.

It doesn't appear to me that there was a serious effort to elicit both sides of this scientific question because the Committee was only presented with one side of the picture. How could that not result in a biased analysis?

I hate to use the term, but it appears that the deck was stacked. None of the experts asked to address and provide background materials to the Committee had taken the position that the scientific evidence is not sufficient to conclude that smoking bans result in dramatic, immediate reductions in heart attacks. Thus, it is perhaps not surprising that the Committee failed to consider the multitude of data that refutes its ultimate conclusion.

For example, had I been asked to address the Committee, they would have had in hand all the relevant data from England, Scotland, Denmark, and Wales which showed no effect of the smoking bans in those countries on admissions for acute coronary events. At least they would have had the data in hand and could have considered it as part of its review. These data could also have been provided by any number of other experts in the field. But if you only review data that support a conclusion, you're going to end up drawing that conclusion, regardless of its validity.

The rest of the story is that not only is the report of the IOM committee on smoking bans and heart attacks biased, but the process that led to this report appears to be significantly biased as well.

No comments: