Friday, October 16, 2009
Analysis Reveals that Institute of Medicine Report Failed to Include Data that Found No Effect of Smoking Bans on Acute Coronary Events in 3 Countries
Bias in Report Results from Failure to Consider Data Which is Inconsistent with Report's Conclusions
I have analyzed the data which the Institute of Medicine included and failed to include in its report and today, I reveal that the report failed to consider data from three countries (England, Scotland, and Wales) which seem to clearly show that the smoking bans in these countries had no significant short-term effect on acute coronary events. These data are all national data which include all hospital admissions at all hospitals in these countries. Thus, they represent a better source of data than what was used in some of the published studies (which only included a sample of hospitals). Moreover, they cover large populations, with a sample size greater than that of all other studies combined. Thus, the data from these countries are critically important an carries much weight in the overall analysis.
National data from the National Health Service show that England's smoking ban had no effect on the trend in the number of heart attack admissions during the first nine months that the ban was in effect.
The National Health Service has reported annual summaries of the total number of emergency room admissions for heart attacks or for admissions in which the patient subsequently suffered a heart attack. These data cover all hospitals in England; thus, there are no gaps in the data. The data cover the fiscal year, starting April 1 of the first calendar year and ending March 30 of the second calendar year. Thus, the 2007-2008 data cover the period from April 1, 2007 through March 30, 2008.
The smoking ban in England took effect on July 1, 2007. Thus, although the 2007-2008 data contain figures for three months during which the ban was not yet in effect (April, May, and June), if the ban had any dramatic effect on heart attacks (such as a 27% which is the figure which anti-smoking researchers are citing), one would expect to see some significant acceleration of the existing secular decline in heart attack admissions during the reported 2007-2008 period.
However, there was only a 2% decline in heart attack admissions in England from 2006-07 to 2007-08, compared to a 2.8% decline in the preceding year and a 3.8% decline in the year preceding that. Thus, these data show no evidence that the smoking ban resulted in any significant, immediate decline in heart attacks.
Also, these data are very important, because the sample size is huge. Unlike the small sample sizes in studies from Helena, Pueblo, and Bowling Green, the sample size in this England study is larger than all of the published studies combined. In addition, the study covers a long period of time, not just the year before and after the smoking ban (as is the case with some of the published studies).
The data above include all hospital discharges with a diagnosis of heart attack, both those in which the heart attack was the presenting reason for admission and those cases in which the patient suffered a heart attack subsequent to admission. While there is nothing wrong with this, someone could argue that it is only the admissions for a presenting heart attack that are relevant, because presumably, subsequent heart attacks experienced in the hospital are not triggered by secondhand smoke exposure.
I re-ran the analysis using only hospital admissions for presenting heart attacks (those which presumably might have been triggered by secondhand smoke, as hypothesized by many tobacco control researchers). The data look like this:
As one can see visually, there is absolutely no change in the trend of declining heart attack admissions in England during the first nine months during which the ban was in effect. There appears to be a relatively steady decline in heart attack admissions from 2002-2008, with no change associated with the smoking ban.
The decline in heart attack admissions from 2006-2007 to 2007-2008 was 3.7%, compared to declines of 3.7% in the preceding year and 3.8% in the year before that.
Thus, this analysis confirms that no matter how you look at it, there was no change in the rate of declines in heart attack admissions in England associated with the first nine months of the smoking ban.
Data from the Scottish national health service, which cover all of Scotland, reveal that while there was a decline in acute coronary event admissions (heart attacks plus angina) in the first year after the smoking ban, the admission rate increased during the second year back to baseline levels. Thus, there does not appear to be any evidence that the smoking ban in Scotland, implemented in April 2006, led to a decrease in coronary event admissions.
National data from Wales reveal that there was no reduction in hospital admissions for myocardial infarction (heart attacks) during the first 9 months after implementation of the smoking ban throughout Wales. Monthly data on heart attack admissions from all Welsh hospitals for the years 2006 and 2007 were examined. The smoking ban went into effect on April 2, 2007.
There were 4,199 heart attack admissions in 2006 and 4,155 in 2007. Thus, there was essentially no change in heart attacks between these two years. In contrast, there was a 6.3% decline in heart attack admissions from 2005 to 2006 and a 10.3% decline in admissions from 2004 to 2005.
When analyzing the data by month and comparing 2007 to 2006, there was an increase in heart attack admissions during the first five months after the smoking ban (April through August) and a decline in heart attack admissions during the next four months (September through December).
It seems quite clear that the data show that the smoking ban in Wales was not associated with a decline in heart attack admissions during the first 9 months it was in effect (and if anything, the decline in heart attacks came to a halt).
The Rest of the Story
The reasons why the report failed to consider these national data are not clear. However, what is clear is that these data were not presented and reviewed in the report.
For this reason, I believe that the conclusions of the report are severely biased. You cannot possibly come to an unbiased conclusion about the effects of smoking bans on heart attacks if you only consider data which support the contention that these bans are having a strong effect. Obviously, if you exclude data which are counter to your conclusion, you are going to bias your analysis.
It appears this is exactly what happened here.
Remember that a meta-analysis or review of data is only as good as its success in reviewing all relevant data that are available. If you only include data that support your position, you are going to end up with a biased conclusion.
By the way, the data for England and Scotland are readily available online and I believe that the committee was made aware of the existence of these data.
The rest of the story is that the report is biased because it fails to include very relevant data from England, Scotland, and Wales which show no evidence that the smoking bans in those countries caused significant short-term reductions in acute coronary event hospital admissions. Because of the size of the populations covered by these data, including them in the overall review would have negated the overall conclusion of the report.