Thursday, July 07, 2011

New Published Study Finds No Evidence of a Decline in Heart Attacks Following Implementation of a Strong Worksite and Restaurant Smoking Ban

A new study published in the current issue of Preventing Chronic Disease: Public Health Research, Practice, and Policy finds no evidence of any significant decline in acute coronary syndrome (heart attacks and unstable angina) after implementation of a strong workplace and restaurant smoking ban in Kanawha County in 2004.

(see: Rahul Gupta, MD, MPH; Juhua Luo, PhD; Robert H. Anderson, MA; Anita Ray, RS. Clean Indoor Air Regulation and Incidence of Hospital Admissions for Acute Coronary Syndrome in Kanawha County, West Virginia. Preventing Chronic Disease: Public Health Research, Practice, and Policy 2011. 8(4):A77.)

Kanawha County (West Virginia) had virtually no clean indoor air protection in place until January 1, 2004. Prior to that time, smoking was allowed in most workplaces and in up to 50% of designated seating in restaurants. As of January 1, 2004, smoking was prohibited entirely in workplaces, including restaurants.The investigators examined hospital admission rates for acute coronary syndrome from 2000 through 2008, thus including four years prior to the smoking ban and four years following the smoking ban. Had the smoking ban led to a decline in acute coronary syndrome admissions, one would have seen such a decline during the 2004-2008 period compared to the 2000-2004 period.

The data, however, show no such decline. Instead, the data show a steady decline in acute coronary syndrome (ACS) rates that is consistent throughout the study period and does not accelerate after the implementation of the smoking ban.

Here are the actual figures:

Prior to the smoking ban (2000-2004): Age-adjusted rate of ACS hospital admissions declined by 20.3%, or 4.1% per year.

After the smoking ban (2004-2008): Age-adjusted rate of ACS hospital admissions declined by 21.2%, or by 4.2% per year.

Thus, there was no change in the trend in ACS admissions after implementation of the strong smoking ban.

The authors acknowledge this finding: "We did not find additional significant change between, before, and after the removal of smoking areas in restaurants (the key change in the CIAR revision that took effect January 1, 2004) after accounting for the sustainable decline of ACS hospitalizations since the 2000 regulation revision (Table 1)."

They quantify the change in ACS admission rates associated with the smoking ban. The relative risk of ACS admission associated with the smoking ban was 1.02 (95% confidence interval, 0.92 to 1.12). Since the confidence interval crosses 1.0, this means that there was no significant effect of the smoking ban on ACS admission rates.

In summary, this study demonstrates that a stringent smoking ban implemented in Kanawha County was not associated with any significant change in hospital admissions for acute coronary events, refuting the conclusions of other studies which reported such effects.

The Rest of the Story

Miraculously, this is not what the study authors chose to report in their conclusions. As if pulling a silver lining out of a dark cloud, the investigators instead conclude as follows: "In the presence of a CIAR [clean indoor air regulation], a consistent decline in incidence of hospital admissions for ACS can be demonstrated."

The same conclusion is reiterated later in the paper: "In conclusion, our results demonstrate that from 2000 through 2008, the rate of hospital admissions for ACS has consistently declined in Kanawha County in the presence of an existing CIAR."

This is an incredibly disingenuous sleight of hand.

The analysis tested a specific hypothesis: Was the smoking ban (implemented on January 1, 2004) associated with a significant decline in ACS hospital admissions). The analysis included a specific regression coefficient to test for this effect. The coefficient was close to zero and was not statistically significant, indicating that the smoking ban had no effect on ACS hospital admissions.

But instead of simply concluding that they failed to find any significant effect of the smoking ban on acute coronary event admission rates, the authors chose to spin the paper as having found that in the presence of a clean indoor air regulation of some kind (yes, a very weak ordinance was in effect as early as 1995), the rate of ACS hospital admissions dropped consistently over time.

Well of course the ACS hospital admissions dropped during the study period! These rates are declining everywhere. Cardiovascular disease admission rates have dropped everywhere during the past decade. There are major secular changes taking place in coronary artery disease medical management and treatment, including the availability of cholesterol-lowering statin drugs, better control of hypertension, and improved surgical treatments such as angioplasty.

That coronary syndrome rates dropped in Kanawha County during the period 2000-2008 is a non-finding. I could have told you that without even looking at the data. It would have been absolutely striking to have found an increase in ACS admission rates during this time period.

To associate the decline in ACS admission rates during this time period with the almost non-existent smoking restrictions is ludicrous. First of all, the paper does not show any change in the admission rate from before to after the weak county regulations. They had in fact been in effect since 1995. So unless the paper went back to about 1990, it could not possibly draw conclusions about the relationship between the county regulation passed in 1995 and any change in ACS admission rates.

Furthermore, the so-called clean indoor air regulation in the county is meaningless. Restaurants were allowed to designate up to 50% of their seating for smoking. That isn't a smoking ban. It's essentially no regulation. In almost every other study of this kind, such a regulation would be considered the absence of a smoking ban. Obviously, a 50% designated smoking area policy is not going to protect customers from secondhand smoke exposure. This has been proven in multiple studies.

So this paper has pulled two sleight of hand tricks to make it look like a positive finding when in fact the main finding is a negative one. It's incredibly impressive to me to see this kind of manipulation take place in what is a very simple and straightforward study that fails to find any effect of a strong smoking ban on acute coronary event hospitalization rates.

I think it shows just how seriously the quality of tobacco control science has degraded.

A question: How many anti-smoking groups do you think are going to disseminate the findings of this study? I am so sure that not a single group will share these results with the public that I'm willing to bet $100 that no group [Missouri GASP is not eligible] which previously reported the results of a study showing a positive effect of a smoking ban in reducing ACS admissions will now report the result of this study which clearly shows no effect of a smoking ban in reducing ACS admissions.

If any group does report this negative finding, I'll donate my $100 prize offering to that organization. Believe me, I'm not rushing off to find my checkbook. This is a bet that is a sure thing. No anti-smoking group will report this finding because anti-smoking groups aren't interested in the truth. They are interested in results that are favorable to their cause.

This is the sad lesson that I have learned over the past few years. It is disillusioning to me and to my vision of the importance of scientific integrity in the practice of tobacco control in public health.

No comments: