Numbers don't add up

[Vocalek]

The numbers don’t add up. According to the Centers for Disease Control (CDC) smoking prevalence among adults in 1988 was 49.4 million.

• The Campaign for tobacco Free Kids states that 350,000 kids become daily smokers each year.
• Various sources estimate that 400,000 die of smoking related diseases each year.
• The American Cancer Society tells us that 5% of smokers quit each year.

According to my Excel spreadsheet, given those things happening over a 20 year period, we should be down to 16 million smokers.

But smoking prevalence in 2008 was 46 million.

Either a lot more kids are taking up smoking or a lot fewer smokers than 5% are quitting. Setting the number of smoking-related deaths to zero only raises the total to 21 million. How do we account for an additional 25 to 30 million smokers? I’ll put my money on the vast majority of that discrepancy being caused by an inflated estimate of the percent of smokers who succeed in quitting each year.

Which goes to show there is a critical need for more effective tools to help smokers.

Just found the American Heart Association's statement of how many quit each year: 1.3 million. Substituting their fixed number for the 5% gives us 21 million. So we are still missing 25 million smokers--maybe only 16 million if nobody died for 20 years.

 

[tiburonfirst]

well, i don't see how the numbers for smokers gained and lost could have been a constant for the past 20 years, there's probably a good chunk of your discrepancies right there

 

[Elendil]

You are assuming that the only "new" smokers are kids. What those stats don't say is how many adults start smoking each year as well as those who quit but start back up again.

 

[Kate51]

I have to agree with Vocalek, it's probably the death discrepancy! When you constantly see how accurate Dr's can be, they may be jumping to biased conclusions??? (Was that guy who ran through a train barrier smoking? Well, let's test for nicotine!) Or, wait, even better, that 400-pound guy just had Super-Sized Big Mac (FOUR of them??) Test for nicotine! I always liked a cigarette after a big lunch...helped curb my appetite.
Just sayin....using G'ment numbers may be the discrepancy.

 

[geeker]

I believe they count a death from chronic illness as smoking-related if you've ever smoked.

 

[Vocalek]

well, i don't see how the numbers for smokers gained and lost could have been a constant for the past 20 years, there's probably a good chunk of your discrepancies right there

Using the ACS 5% doesn't result in a constant number of smokers lost. The number shrinks as the total number of smokers shrinks.

I tried it another way, thinking that maybe they didn't mean 5% of all smokers, but rather 5% of those who attempted to quit. The exact wording was:

Although 70% of smokers want to quit and 35% attempt to quit each year, less than 5% succeed.

So I tried the spread sheet with "0.0175" instead of "0.05". Left the other numbers the same, and I still came up with only 33.2 million smokers instead of the 46 million CDC says we have.

I have no idea how one would adjust the number of kids taking up smoking. I've never seen it expressed as a variable. I went to the source for the CFTK number http://www.oas.samhsa.gov/NSDUH/2K8NSDUH/tabs/Sect4peTabs10to11.pdf and it shows the number going down: 400,000 in 2007 and 350,000 in 2008. So where would we start back in 1998 to make the number of smokers in 2008 total 46 million? Did a little more digging and found this: Chapter 5: Trends in Initiation of Substance Use - 2001 Summary of Findings

See Figure 5.4. The line under the table reads:

Initiates of daily smoking increased from 1.4 million per year during the late 1980s to 1.9 million in 1997 and decreased back to 1.4 million in 2000. This pattern was mirrored in the incidence rates for 12 to 17 year olds and 18 to 25 year olds.

Even if I plug in 1.4M for the number of new smokers per year (which I think is high, given all the taxes, anti-smoking laws, etc., etc.), I still come up short. We should be down to 43 Million smokers.

I still think they are vastly overestimating the number of people who succeed in quitting. As those who can easily quit leave the ranks of smokers, the percent can quit without difficulty would be bound to be going down.

 

[Kate51]

I wonder where the "succeeded" numbers come from...is it derived from BT's sales or something else. Or is it related to second-hand information, from, again, doctors reporting quit statistics from their patients, or poll driven. Or simple guess-timation, but based on what. I can't seem to find from any links how these reports are in fact statistical. Instead I see a lot of footnotes on "estimate from ___", so these may be fifty years worth of estimating based on progressive cigarette sales from 1960 for all we know.
I know, it's awful to be so suspicious. But I really am slanted to the opinion that obesity is a much larger killer than cigarette smoking ever was or will be simply based on effects. However, death statistics on obesity may be a little more accurate. If a person dies today at 70+ years because of lung complications, the contributor may be many practices and conditions that are no longer present. Using coal or wood for home heating, or kerosene lights, industrial and transportation ventilation and wastes, etc. There again, cigarettes have always been made from 2nd rate smoking materials, pipe, cigar, and chewing tobacco much more 'natural' and 'clean', unadulterated. But I wonder how many death certificates are stamped "Death caused from smoking". My guess: not too many. So, the more one thinks about this, the less confidence I have in these numbers overall. It's a crap shoot. But it's sure relied upon heavily by the proponents of cessation.
As our population grows and ages statistics become very skewed if they're based on estimations. I guess the difference is whether you are questioning them or making the study for a social entity.
Nice work for a Sunday afternoon!

 

[rayneinspain]

Various sources estimate that 400,000 die of smoking related diseases each year.

This line stood out to me for a few reasons:

• "various sources" - This is so ambiguous as to be highly suspect to me. Which sources, and how did they compute that figure?
• "400,000 die of smoking related diseases" - I think, unless there is a stipulation noted elsewhere, that this figure is not exclusive to deaths of smokers. I would guess (until someone corrects me) that this is inflated with presumed SHS-related morbidity figures.

 

[wolflrv]

I think a better comparison would be to compare smoking related deaths and tobacco sales as a trend over the last 20-30yrs. Have deaths and sales gone down steadily? By what percent? Have they risen? By what percent? This would seem to be a much better metric to determine how well smoking cessation is working on the overall population.

Those numbers would also have to be cross-referenced with the population totals over those years as well, because it would include the baby-boomer effect and population rose substantially over those years which would reduce the overall percentage of smokers.

By studying whether or not the trends run parallel to each other, rather than one or the other diverging, you can see whether other methods of cessation are indeed having an impact on the number of smokers still participating.

You could probably also figure the price of a pack of cigs for that year and take the total sales and divide it by (1-1.5 x Pack cost) to give you a ballpark as to the total number of smokers per given year.

Dunno...just off the top of my head..but it seems more reasonable to me...LOL!

 

[PlanetScribbles]

Your question assumes that the numbers are meant to add up Elaine.
The figures are 'massaged' to tell people what they want them to hear.

 

[Thulium]

It's not really valid to compare the numbers because they are both estimates that are calculated by adding up various other estimations, so unless you are getting both sets of numbers (deaths vs. new smokers) from the same study they just won't add up.

How do they get these estimates? See for yourself: Annual Smoking-Attributable Mortality, Years of Potential Life Lost, and Productivity Losses --- United States, 1997--2001

Sex- and age-specific smoking-attributable deaths were calculated by multiplying the total number of deaths for 19 adult and four infant disease categories by estimates of the smoking-attributable fraction (SAF) of preventable deaths. The attributable fractions provide estimates of the public health burden of each risk factor and the relative importance of risk factors for multifactorial diseases. Because of the effect of interactions between various risk factors, attributable fractions for a given disease can add up to more than 100%.

The above is from the 2001 report that they used to create the 2005 report: Smoking-Attributable Mortality, Years of Potential Life Lost, and Productivity Losses --- United States, 2000--2004

(1). In 2005, to assess the economic and public health burden from smoking, CDC published results of an analysis of smoking-attributable mortality (SAM), years of potential life lost (YPLL), and productivity losses in the United States from smoking during 1997--2001 (2). The analysis was based on data from CDC's Smoking-Attributable Mortality, Morbidity, and Economic Costs (SAMMEC) system,* which estimates SAM, YPLL, and productivity losses based on data from the National Health Interview Survey and death certificate data from the National Center for Health Statistics. This report presents an update of that analysis for 2000--2004, the most recent years for which source data are available. The updated analysis indicated that, during 2000--2004, cigarette smoking and exposure to tobacco smoke resulted in at least 443,000 premature deaths, approximately 5.1 million YPLL, and $96.8 billion in productivity losses annually in the United States.

And what did they "discover"?

During 2000--2004, smoking resulted in an estimated annual average of 269,655 deaths among males and 173,940 deaths among females in the United States (Table). The three leading specific causes of smoking-attributable death were lung cancer (128,922), ischemic heart disease (126,005), and chronic obstructive pulmonary disease (COPD)§ (92,915). Among adults aged >35 years, 160,848 (41.0%) smoking-attributable deaths were caused by cancer, 128,497 (32.7%) by cardiovascular diseases, and 103,338 (26.3%) by respiratory diseases (excluding deaths from secondhand smoking and from residential fires). [That's a total of 100% if anybody's counting] Smoking during pregnancy resulted in an estimated 776 infant deaths annually during 2000--2004. An estimated 49,400 lung cancer and heart disease deaths annually were attributable to exposure to secondhand smoke. The average annual SAM estimates also included 736 deaths from smoking-attributable residential fires.

During 2000--2004, on average, smoking accounted for an estimated 3.1 million YPLL for males and approximately 2.0 million YPLL for females annually, excluding deaths from smoking-attributable residential fires and adult deaths from secondhand smoke. Estimates for average annual smoking-attributable productivity losses were approximately $96.8 billion ($64.2 billion for males and $32.6 billion for females) during this period (Table).

Granted, there is still the possibility of risk from carcinogenic TSNA's, but as you know the TSNAs detected in some e-liquids were so low they account for a "attributable fraction" of zero. THAT estimate is consistent with the zero deaths or serious injuries reported from e-cigarette use.

 

[Rocketman]

I've quit four times, one cold turkey, once with the patch, once with the gum, and now once with the e-cig.
Last one has been the longest.
Would any of these show up in the "quit" stats?
Oh, and I'm probably not in the "smoking related deaths" stats just yet.

Rocket,
(delete or move this post if you want)

Thul: I'll 'back post' rather than add another (I like your style).
Statistics is such a wonderful area of math especially since all perturbations in life are statistically related. I feel proud to be included in the cigarette deaths estimate, and 4 times in the successfully quit smoking category. Both of these quantities will eventually drive up the 'lives saved by e-cigs' category. Hopefully never in the died in an automobile accident while refilling a carto group

 

[Thulium]

II'm probably not in the "smoking related deaths" stats just yet.

Except that technically speaking...You are in the "smoking related deaths" moreso than people who are actually dead because all the estimates are based on the number of people who are alive, rather than an actual body count.

Y'see, the way they calculate the number of deaths attributed to smoking is by adding up all the estimated deaths from the different ways that smoking can kill you. ...But they figure out how how many people died by the various routes by taking an estimated fraction of the total number of people who died by that route. ...But they figure out how many people died from each route by an estimated fraction of the total population will die by that route...which is of course an estimate based on the most recent census.

For instance, here's how the ACS calculates the total number of lung cancer deaths (found here: Deaths from Lung cancer - WrongDiagnosis.com ) :

Death statistics for Lung and Bronchial Cancer in various countries:
53.2 men per 100,000 population in the US 2000 (Cancer Incidence, Mortality, and Prevalence Worldwide, GLOBOCAN, 2000 American Cancer Society)
27.2 women per 100,000 population in the US 2000 (Cancer Incidence, Mortality, and Prevalence Worldwide, GLOBOCAN, 2000 American Cancer Society)

Based on that number, they determined that 160,440 people die from lung cancer, and since they estimate that 77% of lung cancer deaths are attributable to smoking, they conclude that 123,836 smokers die from lung cancer. Add that together with the estimated deaths from other types of cancer, cardiovascular disease, respiratory diseases, sprinkle in some estimated deaths from secondhand smoke, house fires, and still births and you get a number around 440,000 deaths attributed to smoking.

Food for thought: Types of tobacco

*Approximately 31,000 new cases of oral cancer will be diagnosed in the U.S. in 2006; nearly two-thirds are male. Estimated deaths from oral cancer in 2004 are at 4,830 men and 2,400 women. (American Cancer Society Cancer Facts & Figures 2004)

...

*"I cannot conclude that the use of any tobacco product is a safer alternative to smoking. This message is especially important to communicate to young people, who may perceive smokeless tobacco as a safe form of tobacco use." (U.S. Surgeon General Richard H. Carmona, MD, June 2003 testimony) While OCF agrees with this statement, especially in relationship to young individuals, there may be some harm reduction benefit when traditional loose leaf chewing tobacco is compared snus style products, and certainly when compared to with smoking tobacco where combustion of the tobacco is present. This does not mean that we endorse the use of smokeless products, as their contribution to disease processes outside the known realm of oral cancer is not completely understood. (OCF)

Elsewhere on that site we find this: The problem death and disease

One study conducted at the University of California, San Francisco, found that more than eight out of ten oral cancer patients were smokers.

 

[IMWylde]

I believe they count a death from chronic illness as smoking-related if you've ever smoked.

Which is a huge point considering the % of the worlds population smoking in the 1950's. Thats just got to skew the numbers.