In this column, Cecil came to the following conclusion:
This statement is completely worthless unless we’re given the base percentages in the first place.
For instance, if he’s increasing his chance of getting cancer from 0.01% to 0.02%, while she’s incresing hers from 10% to 15%, this is a far different answer than his chances going from 10% to 20% and hers going from 0.01% to 0.015%.
I’ve come to expect more thorough analysis than this from Uncle Cec. The question was not answered.
Asking to know the “base numbers” is also completely meaningless, not to mention impossible. How are we to state the individual base chance for either of these two individuals who Cecil doesn’t know and has no medical history on?
For the record, however, men have a higher risk of cancer than women (0.2852% vs. 0.2148% for the top three cancers), so that doesn’t help the letter-writer’s argument much.
By using the average numbers for the general population, of course. There are always things you don’t know about any given sample, in any experimental or observational science, so you’re always forced to ignore at least some variables.
Ironically, there have been a number of recent stories about the problems caused by vitamin D deficiencies, presumably because everyone’s taken precautions to avoid “too much” sun exposure.
http://physician-assistant.advanceweb.com/editorial/content/editorial.aspx?CC=71454
http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2009/04/12/DDEH16JQDN.DTL
('Cause I know you all are going to start screaming “cite, cite!”)
I think the article also stopped short of providing the final answer. We really need to know the odds of dying from it. Cecil does mention that skin cancer is more easily treated than lung cancer… but it’s a pretty soft answer. In fact, I’m left thinking that maybe it doesn’t matter whether you smoke or tan, since you’re 99.9% more likely to be killed by something else anyway.
You’re right, we don’t know anything about the specific people involved, but I don’t think the letter writer was asking for a personal health assessment either.
My point is that if you tell me a particular behavior will “double” the risk of a particular bad thing happening, I don’t know what to do with that information unless I have some idea what the risk was in the first place. (0.01-0.02%? Pssh! I’ll takes me chances. 5-10%? 10-20%? Gotta think twice now.)
A more helpful reply, in my opinion, might have been along the lines of: “Historically, lung cancer afflicts x percent of male smokers and y percent of male non-smokers.” Certainly statistics like this are available. *Of course *there are other variables, but this would at least put things in a clearer perspective.
According to wikipedia the risk of lung cancer in non smokers is 1.3% for males and for skin cancer the risk in females is 1.25%. The risks are remarkably similar.
I agree with Wheelz, and Dr Ben Goldacre (www.badscience.net) - who also points out that using percentages in general can be a bad idea, as “20% extra risk” can be interpreted as “risk increase from (say) 1% to 21%” or “risk increase from 1% to 1.2%”. Those of us who are happy with the concept of percentages also assume that everybody else is; its better to say, for instance, that “about four men in three hundred will get lung cancer”.
I was listening to the NS10 Basic Nutrition webcast from Berkley recently, and they very much play down the skin cancer from tanning risk, on the basis that there are two main types, malignant melanoma (very nasty, not affected by actinic exposure) and basal cell carcinoma (nothing like as nasty, affected by actinic exposure). You have to weigh up these risks against the risks of hypovitaminosis D (osteomalacia, depression, psoriasis, prostate cancer).
I’m going to cut and paste the notes that I took from the vitamin D lecture below, because its interesting, pertinent stuff but I’m too lazy to put it into prose. You can listen to the lecture (possibly not the same year as the one I listened to) here.
Light skin better at absorbing vitamin D - theoretically reproductive capability should drop in high-latitude climes if vitamin D not supplemented in diet. Most foods don’t have much vit D (milk is big source)
Parathyroid senses dropping calcium levels and produces PTH which causes activation D3 to calcitriol; this increases intestinal uptake of calcium via gene upregulation. BOTH calcitriol and PTH cause bone to dump calcium.
If you have enough vit D3 and not enough calcium this will eventually decalcify their bones.
Osteomalacia (=rickets in infants) - organic bone matrix normal but not calcified.
Osteoporosis - maxtrix/calcium ratio normal but both reduced, leading to thin and brittle bones. Starts at the age of 30!
If you’re white/asian, old, sedentary, female and small you have bigger risk of osteoporosis. Important to build up calcium when younger.
Most old ladies don’t fall and break their hip - they break their hip and then fall.
Astronauts on their return never strut out - always carried out by people. 2% irreversible bone loss per month at the moment - real issue for 2y trip to Mars!
Vit D3 also needed for reproduction, to stop psoriasis, cancer (?breast, colon and prostate) and autoimmune diseases (T1 DM) and SAD
100,000 IU in one month reduced depression in 8 subjects.
Adequate intake 5-15 µg/d traditionally, but this probably isn’t actually adequate (will also depend on skin colour, skin surface exposed, sunlight exposure etc).
Need 10 mins a day of good sunlight with trunk exposure for good intake. Sun block prevents it!
Malignant melanoma is much nastier than basal cell carcinoma - the first one hasn’t been shown to be sun associated, the second one has but is generally curable.
Toxicity is overexaggerated. Poss up to 4000 units may be useful (based on effect of less than this on rickets)
Rickets becoming quite a problem in inner cities, esp darker skinned people or those who are lactose intolerant.
Beyond the cancer risk with smoking, there are links to heart disease, stroke, hypertension, copd, emphysema,…
Non-cancer-related tanning risks are minimal, besides resembling a reptile after extended tanning for decades.
Good points. This was one of the worst Cecil answers ever. It’s been pointed out that the incidence of lung cancer is similar to that of skin cancer. However, I believe Lung Cancer is much more serious and deadly than skin cancer, on average. I have known so many people who have had skin cancer, requiring simple procedures to remove it, and have been fine afterwards, whereas lung cancer seems to be more deadly. I believe there are stats on lung cancer deaths and skin cancer deaths, which Cecil should have looked up and cited. From there he could have applied the cigarette and tanning bed effect on those stats, and see who would’ve “won the bet.” I don’t know Cecil well enough, but perhaps he’s a smoker who was afraid to state an obvious fact.
Quick follow-up. I googled “skin cancer deaths” and cigarette related deaths. Skin cancer caused about 10,000 per year vs. 50,000 for smoking related deaths.
But he did say who should have “won the bet:”
RR
Yeah, River Runner, I guess he did say the writer “won the bet,” but only in an offhand way, after a few paragraphs of not really good statistical argument, which could have been replaced by a quick statement of deaths caused by the two behaviors.