Wednesday, April 15, 2009

Don't forget the sunscreen...or data to prove your point

Commentary on: South Dakota sun lovers smarter than most By Jon Walker, Argusleader.com, April 14, 2009

This article actually details how South Dakota residents are really bad at wearing sunscreen. Considering that many S. Dakota residents work outside (i.e. farming, ranching, etc) the risk for skin cancer is quite high. The article's title implies that South Dakotans are actually good at protecting themselves with sunscreen, "...sun lovers smarter than most," but the author never provides data of sunscreen usage in other states for a true comparison...so I did!

The American Cancer Society has some really great information on this from 2008 (source). Here are states that have counts below South Dakota:

State -- New cases of melanoma (skin cancer)
South Dakota -- 160
Wyoming -- 120
North Dakota -- 110
Alabama -- 80
Dist. of Columbia -- 50

Now, the numbers above are frequencies (just counts) and they can also be misleading unless you compare the rates of new cases. Rates are more representative of the situation since it takes into account population size. Rates are important to understand what's going on.

So, I took into account each state's population and calculated the rates of new cases below:

State -- Rates of new cases of melanoma (count/total pop) x 100
South Dakota --0.21%
Wyoming -- 0.02%
North Dakota --0.02%
Alabama -- 0.002%
Dist. of Columbia --0.01%

Now -- looking at these rates, we have a completely different picture than we did with just the count data. According to the rates, South Dakota is the WORST at preventing new rates of skin cancer. This is discordant with what the article proclaims.

Bottom line:
  • South Dakota needs to be better at applying sunscreen as a state.
  • Be careful of data without denominators and
  • Beware of misleading titles (like the one in this article).
  • Always look for data to support the author's claims.
  • Finally-- do a quick google images search of 'skin cancer' or 'melanoma' and you will see just how ugly it really is. Eghh. Apply sunscreen liberally!


If you want to know the skin cancer rates in your state or do your own fact checking, check out the American Cancer Society's 2008 data: http://www.cancer.org/downloads/STT/2008CAFFfinalsecured.pdf

3 comments:

  1. I still disagree with your argument. You cannot just compare rates of skin cancer in a vacuum to judge each state. People in Florida get a lot more sun than people in Maine, so they would automatically have more skin cancer. I think the data needs to be normalized more....maybe by # of sunny days? By average temperature? Not sure what would work, but you see my point right?

    ReplyDelete
  2. The assessment posted in the blog was not trying to establish causation, correlation or behavioral factors associated with new cases of melanoma. That would require a logistical regression model that would take into account other confounding factors such as: age, gender, socioeconomic status, number of sunny days, etc.

    The message in the blog is to really consider the data being presented in an article, ask yourself if it makes sense, and if it doesn't to then do some additional research and make your own conclusions.

    ReplyDelete
  3. I agree with the message presented in the blog that data presented in articles should be taken with a grain of salt, but don't agree with the comparison of state melanoma incidence using crude rates. Age plays a big factor in cancer diagnosis; any comparison of cancer rates between states should at the very least take age into account (i.e. age-adjusted rates). Here is a nice link that lays out issues that should be taken into account when comparing cancer rates between states: http://www.cdc.gov/cancer/npcr/uscs/data/00_guidance_include.htm
    You can also query state-level age-adjusted melanoma rates at this link: http://apps.nccd.cdc.gov/uscs/Table.aspx?Group=4fb&Year=2005&Display=n

    ReplyDelete