4 ms·

This is really well presented, but wouldn't it be more accurate and useful with actual cancer case data instead of estimated cancer risk?

• s1artibartfast 5y ago
  too many confounding factors.
• LorenPechtel 5y ago
  Too hard to figure out. I'm thinking of where I grew up. A local anti-establishment rag noted a cancer cluster in one part of town and wouldn't let go even when shown the truth. The end result of the mess was it went from a cheap but decent area to somewhere I wouldn't want to venture even by day. The only toxic stuff in the neighborhood was benzene from all the old cars about--the "cluster" was because it was the cheapest decent area in town, people who got sick and had medical bills and couldn't work ended up moving there. The rate of *diagnosis* of cancer there was below average, the cluster was purely due to immigration.
  • readams 5y ago
    "Cancer clusters" by themselves are not really surprising, since statistically if you look at enough small subsets it's certain you'll see such seemingly-anomalous "clusters." It's interesting only as the first step in an investigation, and not at all interesting as evidence.
  • refurb 5y ago
    But the problem with estimated cancer risk is you end with the whole breast implant issue - it was “proven” silicone breast implants caused cancer until data proved they didn’t.