This was an 8 year longitudinal study of those in pain. Pain is commonplace and usually meaningless but what if it was indicative of risk of cancer?
Evidence from:
GJ Macfarlane et al. BMJ. (2001) #7314 p 662.
The report includes estimates of the prevalence of pain and the statistical association between pain and cancer outcomes. It would appear that the pain being described by people before they died had no direct link with the cancer that killed them.
The Radar report is available to subscribers: