[slingnow]

Think of what you're saying. If it was so bimodal that the screened people "gained years" while the ones who didn't "gained 0", then how could the average possibly be 4 months?

If we take your theory at face value, on average the screened people would have an 8 month increase.

[francisofascii]

A very simplistic hypothetical scenario could be, people who get colon cancer die an average of 6-7 years earlier than they would have. And if 5% of the population get colon cancer, then the average gain would be 4 months with screening.