Sunday, October 18, 2009

Data from Denmark Show No Apparent Effect of Smoking Ban on Heart Attacks

Data from a government study of the effects of the smoking ban in Denmark show no apparent impact of that smoking ban on heart attack hospital admission rates.

The study analyzed rates of hospitalization for heart attacks in all hospitals in Denmark for a period of five years prior to the smoking ban (implemented in August 2007) and 18 months after implementation of the ban. A statistical model controlled for age, month, and secular trends in heart attack rates. No effect of the smoking ban was detected for either 35-64 year-old males, 35-64 year-old females, 65-84 year-old males, or 65-84 year-old females.

The trends in heart attacks among one of these four groups - men ages 35-64 -- is shown in Figure 6 of the report. As can be seen, there is a clear trend of decreasing heart attack rates over time. Seasonal variation in heart attacks is also clear. There is no apparent effect of the smoking ban on the trend in heart attacks. According to the report, the graphs for each of the other three demographic groups are similar, with no apparent effect of the smoking ban on heart attacks in any of the four study groups.

An advantage of this study is that Denmark has an excellent national registry of heart attack admissions. In addition, the report notes that there were few changes in treatment of heart disease in the country during the period surrounding the implementation of the smoking ban.

The Rest of the Story

This is now the fourth country for which data show no short-term effect of the smoking ban on admissions for acute coronary events. The others are England, Scotland, and Wales. None of these findings were included in the Institute of Medicine report which concluded that smoking bans cause an immediate and substantial decline in heart attacks.

An important observation is that the few studies which found large effects of the smoking bans were in very small communities. None of the studies that covered larger populations -- states or countries -- found large effects and many of these studies found no effect. This suggests that the findings from isolated communities - like Helena and Pueblo - are merely random phenomena. The fact that these results cannot be confirmed in any population-based study is concerning.

This pattern of findings also confirms that there is a severe publication bias present. We now have data from four countries showing no short-term effects of the smoking ban on acute cardiovascular events, yet none of these studies have been published. Researchers have a hard time getting excited about negative results and so they are much less likely to decide to publish such data. In contrast, when they find an effect, they are excited about publishing the findings. This is clearly what is going on with these smoking ban and heart attack studies.

(Thanks to Klaus K. for kindly providing the English translation of this report).

No comments: