Thursday, April 08, 2010

New Study Shows Huge Decline in Heart Attack Admissions; Demonstrates Why IOM Committee's Report on Effect of Smoking Bans is Such Poor Science

A study published in a recent issue of the journal Circulation reports that from 2002 to 2007, there was a 23.4% decline in heart attack admissions among the Medicare fee-for-service population (see: Chen J, Normand SL, Wang Y, Drye EE, Schreiner GC, Krumholz HM. Recent declines in hospitalizations for acute myocardial infarction for Medicare fee-for-service beneficiaries: progress and continuing challenges. Circulation. 2010 Mar 23;121:1322-8).

After controlling for age and other factors, the study estimates that the average decline in heart attack admissions in recent years has been 5.8% per year. According to the study authors, this secular decline in heart attack admissions is attributable to a number of factors, including improved medication for heart disease (especially the use of statins to lower cholesterol levels), improved surgical treatment for coronary artery stenosis, and decreased rates of smoking.

The Rest of the Story

This study demonstrates why the IOM committee which issued a report on the effect of smoking bans on heart attacks used such poor science in drawing its conclusion. Because of weaknesses in the underlying studies, the committee was unable to quantify the magnitude of the decrease in heart attacks associated with the implementation of smoking bans. But this new study shows that there was a large secular decline in heart attacks occurring anyway, on the order of about a 6% decline in heart attack admissions per year. If the IOM committee was unable to quantify the magnitude of the effect of smoking bans on heart attacks, then how does the committee know that the decline was significantly greater than 6%?

In fact, the IOM committee made no attempt to quantify the secular change in heart attacks that is occurring in the absence of smoking bans. And it made no attempt to quantify the degree to which heart attack declines associated with smoking bans exceed the secular declines that are occurring anyway, as documented in this new study.

In other words, the IOM committee's conclusions were based on very poor science: the absence of any control group and the failure to estimate a magnitude of effect and determine whether the confidence interval around that point estimate includes the secular rate of decline occurring in the absence of smoking bans.

This new study also demonstrates the invalidity of the conclusions of many of the individual smoking ban/heart attack studies.

For example, a study on the effect of smoking bans on heart attacks in New York State concluded that there was an 8% decline in heart attacks in the year following the statewide smoking ban. The study concludes that the smoking ban reduced heart attacks by 8%. But these new data show that in the same year (2004), heart attacks nationwide were declining by about 6%. This suggests, actually, that the New York State smoking ban had no immediate effect on heart attack rates. Certainly, in light of the 6% decline that would have been expected anyway, the finding that heart attacks declined in 8% does not support a conclusion that the smoking ban significantly reduced heart attacks (and obviously does not support the conclusion that the ban decreased heart attacks by 8%).

Nevertheless, these are data that the IOM committee relied upon to justify what now looks like a pre-determined conclusion that smoking bans result in immediate reductions in heart attacks.

As another example, consider the recent meta-analysis out of the Kansas University Medical Center which showed an 8% decline in heart attacks associated with smoking bans in 11 published studies. That 8% decline sounds impressive. Until you consider the fact that heart attacks are declining by about 6% per year anyway. I hate to have to say it, but this really amounts to junk science. The complete absence of any comparison group and the assumption that any change in heart attacks that occurs is attributable to the smoking ban. Any student in my program evaluation course who proposed a study like this without a control group would not pass the class.

The rest of the story is that there is now strong evidence that heart attacks are declining substantially due to improved medical and surgical treatment of heart disease and behavioral changes, including declines in smoking. The magnitude of the secular decline in heart attacks is roughly the same as the declines in heart attacks reported in the published studies which are being used to support the conclusions that smoking bans result in immediate, dramatic reductions in heart attacks. The chief failure of anti-smoking groups and researchers which are using these studies to support the conclusion that smoking bans were the cause of the decline in heart attacks is that they have failed to consider that heart attacks are declining substantially anyway. They have failed to compare the rates of decline associated with the implementation of smoking bans with the secular rates of decline in heart attacks. This has led to conclusions which are unsupported by scientific evidence and based on severely flawed science.

No comments: