Joined
·
1,537 Posts
NOTE: I am posting Cigar Science VII at the same time which looks at a particular study but I would recommend that you read this first.
PREFACE: As promised, I am going to take a detour from my other posts about taste perception to consider the potential health risks of cigar smoking. There is quite a lot of research out there and it is fairly complicated so I have decided to review it study by study. At the end I will try to summarize it all.
For my own personal interest I am primarily concerned with the effects cigars will have on the health of the smoker so I will start with that topic. In the future I will also look at research on the effects of second hand smoke.
Before getting into specific studies, I thought I would post a little bit about the statistics that are used because they can be very misleading. Sorry for all the math in this post but it is really the only way to get the point across. Hopefully reading this will make you a better consumer of the research in this area
ODDS RATIOS: If you have read any articles about the health effects of smoking you probably noticed the most common statistic reported is the Odds Ratio (OR) and most of the studies I am going to talk about use this as well. Before we get into the actual study results I think it is important that we all understand what an OR is and isn't.
An OR expresses the relative odds of an event happening for two different groups. An OR of 1.0 means both groups have the same odds of the event occurring, an OR of 2.0 means that one group is twice as likely to have the event occur, and so on. I will use an example from the sinking of the Titanic to show you how an OR is calculated and how to interpret it. Here are the basic stats from the sinking of the Titanic:
Female passengers: Alive 308, Dead 154, Total 462
Male passengers: Alive 142, Dead 709, Total 851
Total passengers: Alive 450, Dead 863, Total 1313
How much more likely was it for a male passenger to die on the Titanic than a female passenger? To calculate the OR you take the relative proportion Dead/Alive for Males and divide it by the relative proportion of Dead/Alive for Females:
OR= (709/142) / (154/308) = 4.99/0.5 = 9.98
So males had roughly a 10 times greater odds of dying on the Titanic. This sounds like a really big number and in the case of anti-tobacco lobbyists they can throw it around and scare the hell out of people. But…
WHY OR's CAN BE MISLEADING:
Problem #1: OR's do not take into account the actual probabilities of the events happening.
If we look at the percentage chance of dying (DEAD divided by TOTAL) in each group from the above numbers we get:
Males: 709/851 = 83.3%
Females: 154/462=33.3%
The 83.3% for males is indeed a high probability. But imagine if instead the survival statistics from the Titanic were the following:
Female passengers: Alive 432, Dead 2, Total 462
Male passengers: Alive 811, Dead 40, Total 851
We would now get:
OR= (40/811) / (2/432) = 0.049/0.005 = 9.86
Still about the same OR as before but now the probabilities of dying are only:
Males: 40/851=4.9%
Females: 2/462=0.4%
So men do have a 10 times greater odds of dying but still have an over 95% chance of surviving. So the large OR is pretty meaningless in this case. This will be an important point because a lot of the studies on cigars have very low numbers of subjects and the actual probabilities of risks are low.
Problem #2: OR's don't fit with our natural interpretation of relative risk.
-If I told you that 10% of non-smokers get disease X while 90% of cigarette smokers get disease X most people would say that you are 9 times more likely to get disease X if you smoke cigarettes. But no in this case the OR is equal to 81!!! Again this larger number is good for politicians to throw around as it has a much more dramatic impact.
Problem #3: OR's do not fall on a linear scale so they are not easy to compare.
-Imagine that a study found that 10% of non-smokers get disease X, 20% of cigars smokers get it, 40% of pipe smokers get it, and 80% of cigarette smokers get it. Logically we would say that you are twice as likely to get disease X if you smoke pipes as compared to cigars, and are 4 times as likely to get disease X if you smoke cigarettes as compared to cigars. But if you break this down to odds ratios you get:
Cigar OR = (20/80) / (10/90) = 2.3
Pipe OR = (40/60) / (10/90) = 6.1
Cigarette OR = (80/20) / (10/90) = 36.36
The relative relationship between the different groups is totally lost when you convert to OR's and can look a lot worse than it is. Now you are 3 times as likely for pipes versus cigars and 18 times as likely for cigarettes versus cigars.
CONCLUSION: So you need to be very careful when interpreting Odds Ratios. A large number is not always as bad as it looks. Next time we will consider a specific study on the health effects of cigar smoking and be sure to keep this in mind
Cheers!
Rob
PREFACE: As promised, I am going to take a detour from my other posts about taste perception to consider the potential health risks of cigar smoking. There is quite a lot of research out there and it is fairly complicated so I have decided to review it study by study. At the end I will try to summarize it all.
For my own personal interest I am primarily concerned with the effects cigars will have on the health of the smoker so I will start with that topic. In the future I will also look at research on the effects of second hand smoke.
Before getting into specific studies, I thought I would post a little bit about the statistics that are used because they can be very misleading. Sorry for all the math in this post but it is really the only way to get the point across. Hopefully reading this will make you a better consumer of the research in this area
ODDS RATIOS: If you have read any articles about the health effects of smoking you probably noticed the most common statistic reported is the Odds Ratio (OR) and most of the studies I am going to talk about use this as well. Before we get into the actual study results I think it is important that we all understand what an OR is and isn't.
An OR expresses the relative odds of an event happening for two different groups. An OR of 1.0 means both groups have the same odds of the event occurring, an OR of 2.0 means that one group is twice as likely to have the event occur, and so on. I will use an example from the sinking of the Titanic to show you how an OR is calculated and how to interpret it. Here are the basic stats from the sinking of the Titanic:
Female passengers: Alive 308, Dead 154, Total 462
Male passengers: Alive 142, Dead 709, Total 851
Total passengers: Alive 450, Dead 863, Total 1313
How much more likely was it for a male passenger to die on the Titanic than a female passenger? To calculate the OR you take the relative proportion Dead/Alive for Males and divide it by the relative proportion of Dead/Alive for Females:
OR= (709/142) / (154/308) = 4.99/0.5 = 9.98
So males had roughly a 10 times greater odds of dying on the Titanic. This sounds like a really big number and in the case of anti-tobacco lobbyists they can throw it around and scare the hell out of people. But…
WHY OR's CAN BE MISLEADING:
Problem #1: OR's do not take into account the actual probabilities of the events happening.
If we look at the percentage chance of dying (DEAD divided by TOTAL) in each group from the above numbers we get:
Males: 709/851 = 83.3%
Females: 154/462=33.3%
The 83.3% for males is indeed a high probability. But imagine if instead the survival statistics from the Titanic were the following:
Female passengers: Alive 432, Dead 2, Total 462
Male passengers: Alive 811, Dead 40, Total 851
We would now get:
OR= (40/811) / (2/432) = 0.049/0.005 = 9.86
Still about the same OR as before but now the probabilities of dying are only:
Males: 40/851=4.9%
Females: 2/462=0.4%
So men do have a 10 times greater odds of dying but still have an over 95% chance of surviving. So the large OR is pretty meaningless in this case. This will be an important point because a lot of the studies on cigars have very low numbers of subjects and the actual probabilities of risks are low.
Problem #2: OR's don't fit with our natural interpretation of relative risk.
-If I told you that 10% of non-smokers get disease X while 90% of cigarette smokers get disease X most people would say that you are 9 times more likely to get disease X if you smoke cigarettes. But no in this case the OR is equal to 81!!! Again this larger number is good for politicians to throw around as it has a much more dramatic impact.
Problem #3: OR's do not fall on a linear scale so they are not easy to compare.
-Imagine that a study found that 10% of non-smokers get disease X, 20% of cigars smokers get it, 40% of pipe smokers get it, and 80% of cigarette smokers get it. Logically we would say that you are twice as likely to get disease X if you smoke pipes as compared to cigars, and are 4 times as likely to get disease X if you smoke cigarettes as compared to cigars. But if you break this down to odds ratios you get:
Cigar OR = (20/80) / (10/90) = 2.3
Pipe OR = (40/60) / (10/90) = 6.1
Cigarette OR = (80/20) / (10/90) = 36.36
The relative relationship between the different groups is totally lost when you convert to OR's and can look a lot worse than it is. Now you are 3 times as likely for pipes versus cigars and 18 times as likely for cigarettes versus cigars.
CONCLUSION: So you need to be very careful when interpreting Odds Ratios. A large number is not always as bad as it looks. Next time we will consider a specific study on the health effects of cigar smoking and be sure to keep this in mind
Cheers!
Rob
