I've done a little research as far as baseline lung cancer risk goes. And before someone quotes Mark Twain at me, just let me say that I did most of this to satisfy my own curiosity well before this thread came up, and I did it all partially just for the fun of figuring out how to work the numbers, as well as to better inform myself. And remember to always take statistics with a grain of salt.
According to
National Cancer Institute ~6.6% of men and women will be diagnosed with lung cancer at some point in their live. This number encompasses the marathon runners and the 2 packs a day cigarette smokers. The
Lung Cancer Institute says that 80-90% of lung cancer cases are due to some form of smoking, meaning that nonsmokers account for 10-20% of cases.
15% of 6.6% is
0.99%, meaning that if nobody smoked, about 1% of the US population would contract lung cancer at some point.
85% of 6.6% is
5.61%, meaning that if everybody smoked to some degree and smoking were the only cause of lung cancer, about 5.61% of the US population would get it.
The NCI also says that the 5 year survival rate for lung cancer is 17.4%, so an average nonsmoker has a
0.826% chance of dying from lung cancer.
A fairly substantial
study on pipe smoking showed that men who smoked 1-3 pipe fulls a day had a relative risk (RR) of death by lung cancer of 1.99 (95% confidence interval: 1.12, 3.29; p less than .001).
Back calculate from there, and you can show that, for a 1-3 pipe a day smoker, there is:
0.826% for nonsmokers * 1.99 = a
1.64374% chance of dying from lung cancer, and
1.64% of cases die * 1.174 = a
1.99% chance of being diagnosed with lung cancer. That's 1 in 50 pipe smokers at a 1-3 pipes per day habit. That 1.99 value comes back up because the incidence rate for nonsmokers happens to be 1%, and the effects that has on the related values.
It's worth noting that the risk increases with the number of pipes per day, with the depth of inhalation, and with the number of years a you've been smoking.
Of course these are all generalizations, as is the nature of statistics. And not just that, but then I went a did calculations using values from multiple sources! Though the study I referenced accounts for a decent number of confounding variables there is no way to perfectly represent a specific individual case. There's going to be some skewing of values one way or another.
Feel free to check my math, I'd like to know if I messed up somewhere in there!
