Sometimes I find I have important data that I can’t quite explain. For example, cancer rates in the US vary by more than double from county to county, but not at random. The highest rates are on the rivers, and the lowest are in the mountains and deserts. I don’t know why, but the map shows it’s so.
Cancer death rates map of the US age adjusted 2006-2010, by county. From www.statecancerprofiles.cancer.gov.
Counties shown in red on the map have cancer death rates between 210 and 393 per 100,000, more than double, on average the counties in blue. These red counties are mostly along the southern Mississippi, the Arkansas branching to its left; along the Alabama, to its right, and along the Ohio and the Tennessee rivers (these rivers straddle Kentucky). The Yukon (Alaska) shows up in bright red, while Hawaii (no major rivers) is blue; southern Alaska (mountains) is also in blue. In orange, showing less-elevated cancer death, you can make out the Delaware river between NJ and DC, the Missouri heading Northwest from the Mississippi, the Columbia, and the Colorado between the Grand Canyon and Las Vegas. For some reason, counties near the Rio Grande do not show elevated cancer death rates. nor does the Northern Mississippi and the Colorado south of Las Vegas.
Contrasting this are areas of low cancer death, 56 to 156 deaths per year per 100,000, shown in blue. These appear along the major mountain ranges: The Rockies (both in the continental US and Alaska), the Sierra Nevada, and the Appalachian range. Virtually every mountain county appears in blue. Desert areas of the west also appear as blue, low cancer regions: Arizona, New Mexico, Utah, Idaho, Colorado, south-west Texas and southern California. Exceptions to this are the oasis areas in the desert: Lake Tahoe in western Nevada and Lake Meade in southern nevada. These oases stand out in red showing high cancer-death rates in a sea of low. Despite the AIDS epidemic and better health care, the major cities appear average in terms of cancer. It seems the two effects cancel; see the cancer incidence map (below).
My first thought of an explanation was pollution: that the mountains were cleaner, and thus healthier, while industry had polluted the rivers so badly that people living there were cancer-prone. I don’t think this explanation fits, quite, since I’d expect the Yukon to be pollution free, while the Rio Grande should be among the most polluted. Also, I’d expect cities like Detroit, Cleveland, Chicago, and New York to be pollution-heavy, but they don’t show up for particularly high cancer rates. A related thought was that specific industries are at fault: oil, metals, chemicals, or coal, but this too doesn’t quite fit: Utah has coal, southern California has oil, Colorado has mining, and Cleveland was home to major Chemical production.
Another thought is poverty: that poor people live along the major rivers, while richer, healthier ones live in the mountains. The problem here is that the mountains and deserts are home to some very poor counties with low cancer rates, e.g. in Indian areas of the west and in South Florida and North Michigan. Detroit is a very poor city, with land polluted by coal, steel, and chemical manufacture — all the worst industries, you’d expect. We’re home to the famous black lagoon, and to Zug Island, a place that looks like Hades when seen from the air. The Indian reservation areas of Arizona are, if anything, poorer yet.
Cancer incidence,age adjusted, from statecancerprofiles.cancer.gov
My final thought was that people might go to the river to die, but perhaps don’t get cancer by the river. To check this explanation, I looked at the map of cancer incidence rates. While many counties repress their cancer rate data, the pattern in the remaining ones is similar to that for cancer death: the western mountain and desert counties show less than half the incidence rates of the counties along the southern Mississippi, the Arkansas, and the Ohio rivers. The incidence rates are somewhat elevated in the north-east, and lower on the Yukon, but otherwise it’s the same map as for cancer death. Bottom line: I’m left with an observation of the cancer pattern, but no good explanation or model.
Dr. Robert E. Buxbaum, May 1, 2014. Two other unsolved mysteries I’ve observed: the tornado drought of the last few years, and that dilute toxins and radiation may prevent cancer. To do science, you first observe, and then try to analyze.