Tuesday, September 12, 2006

Nicotine Yield of Philip Morris Cigarettes Drops in 2005; Yields of Philip Morris Cigarettes Have Actually Declined Since 1997

According to reports submitted to the Massachusetts Department of Public Health (MDPH) by Philip Morris, the nicotine yield of its cigarettes (as measured by a machine method prescribed by Massachusetts regulations that simulates actual smoking behavior) fell from an average of 1.76 mg per cigarette in 2004 to 1.69 mg per cigarette in 2005. The nicotine yield of Marlboro cigarettes fell from 1.84 mg per cigarette to 1.80 mg per cigarette, and over the full period 1997-2005, dropped from 1.85 to 1.80.

These findings, which are being released today on The Rest of the Story and are consistent with data presented by Philip Morris, contradict conclusions presented in a MDPH report which asserted that there has been a significant increase over time in the nicotine yield of Philip Morris cigarettes, including its Marlboro brand.

The Massachusetts Department of Public Health had presented data showing that nicotine yield in Marlboro cigarettes increased steadily from a 1.7 mg average in 1998 to a 1.9 mg average in 2004. What the report failed to disclose is that the nicotine yield of Marlboro cigarettes in 1997 was 1.85 mg, meaning that between 1997 and 2004 there was essentially no change in nicotine yield of Marlboro cigarettes.

In addition, the report did not disclose that in 2005, Marlboro nicotine yield dropped to an average of 1.80 mg, meaning that there was actually a decline in the average nicotine yield of Marlboro cigarettes over the entire period 1997-2005, from 1.85 mg to 1.80 mg.

It turns out that for some reason, 1998 was a particularly low year for Marlboro nicotine yield. While averaging between about 1.8 and 1.9 mg for every other year in the period 1997-2005, the average was only about 1.7 mg in 1998. Thus, using 1997 as the starting period in the report led to the appearance of a substantial increase, when in fact the levels have been fluctuating but relatively steady over the past 9 years.

The MDPH report also concluded that nicotine yields of Philip Morris brands overall increased significantly from 1998 to 2004. It is indeed true that the average nicotine yield of all reported Philip Morris cigarettes increased from 1.68 mg in 1998 to 1.76 mg in 2004. However, the average nicotine yield of Philip Morris cigarettes dropped to 1.69 mg in 2005, meaning that over the entire period 1998-2005, there was no change in the nicotine yield of Philip Morris cigarettes.

Philip Morris did not report the nicotine yield of any non-Marlboro cigarettes in 1997, but given the substantially higher yields of Marlboro in that year and the general relationship between Marlboro and non-Marlboro yields, it is almost certain that the average nicotine yield of Philip Morris cigarettes as a whole actually dropped from 1997 to 2005.

In summary, it seems pretty clear from the Philip Morris reports submitted to MDPH that between 1997 and 2005, the average nicotine yields of Marlboro cigarettes and all Philip Morris cigarettes decreased.

The Rest of the Story

The rest of the story is that in contrast to the report, which gained widespread media attention and led to newspaper articles throughout the nation implicating Philip Morris in a scheme to increase its nicotine levels over the past seven years or so, the truth appears to be that there has been no real change in nicotine yields of Philip Morris cigarettes, including Marlboro, over the past nine years.

There has certainly been some year-to-year fluctuations in the nicotine yields, but what was reported as a clear trend over time turns out to have been fluctuation around a relatively steady level. If anything, nicotine yields in 2005, the most recent year for which data are available, are slightly lower than they were in 1997, the first year in which these data were reported.

This story demonstrates the dangers of leaving out specific years of data when trying to draw conclusions about trends in a phenomenon over time. By failing to consider the 1997 data and beginning the analysis in 1998, the report obtained a falsely low idea of the baseline nicotine yields of Philip Morris cigarettes, and therefore this created the appearance of a significant trend of increasing nicotine yields in recent years.

In addition, by failing to consider the 2005 data, which demonstrate that the nicotine yields of Marlboros and Philip Morris cigarettes as a whole decreased, the report created a false appearance not only that nicotine yields were steadily increasing, but that they were increasing at the present time.

Had the report considered just the 1997 data and ignored the 2005 data, it would have been clear that Marlboro's average nicotine yields have not increased, as the results for Marlboro cigarettes would have been 1.85 mg in 1997 and 1.84 mg in 2004.

Had the report considered just the 2005 data and not the 1997 data, it would have also been clear that Philip Morris' average nicotine yields have not increased, as the average results for Philip Morris cigarettes would have been 1.68 mg in 1998 and 1.69 mg in 2005.

The combination of not considering data from either 1997 or 2005 resulted in the appearance of a substantial increase in nicotine yields of Philip Morris (including Marlboro) cigarettes that does not seem to accurately reflect the true time trend.

While I had previously been critical of the conclusions of the report and the statements made regarding the report by some anti-smoking groups, even in the face of an accurate depiction in the trend of nicotine yields of Philip Morris cigarettes, the revelation that nicotine yields of Philip Morris cigarettes have not been steadily increasing during the past decade and that they have not even increased overall during this time period seems to render my previous criticism moot.

At least for Philip Morris. Data are not available for the brands manufactured by other cigarette companies, so it may well be that conclusions regarding increases in nicotine yields of those companies' brands over time are valid.

This story also illustrates why the conclusions of studies like those reported from Helena, Pueblo, and Saskatoon need to be viewed more critically. There are year-to-year fluctuations in many types of data and it is very easy to be misled into thinking that a particular change from one year to another reflects a true trend, when in fact it is just a statistical fluctuation. And even if it is a real change, it is incredibly difficult to pin down the reason for that change.

One needs at least two years of data before a trend is evident. In Massachusetts, the 2005 data made it clear that what appeared to be a trend of increasing nicotine yields over time was probably not a true trend at all. Imagine how crazy it is to conclude that a trend is present in Helena or Pueblo when there is only data available for about one year following the implementation of the smoking bans.

As I am fond of saying, let's not be hasty. Here, premature conclusions by tobacco control groups led to what turns out to be misleading, widespread newspaper coverage of an important tobacco policy issue.

While I have already explained why the public has been misled by the interpretation of what an increase in nicotine yields would mean for smokers and passive smokers, it now appears that the public has also been at least partially misled about the basic conclusion that nicotine yields of cigarettes have increased steadily and significantly over time.

I think this illustrates my observation about how the agenda seems to have overtaken the science in the tobacco control movement. We seem to be so anxious to disseminate information to advance our agenda that we are no longer being careful and rigorous in our science, and especially in our public communications. I think it's time to slow down and led the science dictate the agenda, rather than the reverse.

No comments: