A new report on Lancet has looked at the incidence of cancer among Canadian prisoners. Two things that immediatelly come to mind:
- Very large sample size (43k males, 5k females)
- The population of prisoners is in many ways different from the general population.
So, the study found a "significantly" higher incidence of lung cancer, head and neck cancer, and liver cancer. Lower incidence of colorectal and prostate cancer. All this among man. Among women, higher incidence of cervical, lung and liver cancer, lower breast cancer.
Unfortunately this tells us very little. We wonder how many different types of cancers they tested for and if p-values were adjusted in any way. But more than that, can these differences be attributed to the folks being prisoners (like the headline would imply for those not so savvy in stats), or it is something else?
So, I guess the results still offer some food for thought and perhaps development of hypotheses, but much more analyses are needed...