[nonbel]

The people who don't know what they are talking about (or are maliciously feigning ignorance) are those who look at 5 yr survival without considering the effect of increased surveillance/screening. https://en.m.wikipedia.org/wiki/Lead_time_bias

[stewbrew]

You have to distinguish between screening detected cancer and symptomatic cancer. 5yr srvl rates say little in the former case but they are relevant for the latter. You'd really prefer having skin cancer today than ten years ago.

Your short exchange of arguments is a good example of what goes wrong in the public discussion.

[nonbel]

I don't understand what you mean by "detected cancer" and "symptomatic cancer". Eg if a cancer is already detected then by definition it is impossible to screen for it.

[mbreese]

(Screening detected) Found during a routine/annual colonoscopy/mammogram/etc...

vs

(Symptomatic detected) Hey, what is this lump here? Why am I so tired? etc...

[nonbel]

I see, it should have been written "screening-detected" to be as clear as possible.

Yes, the latter is more likely to be comparable over time but it will still have issues with changes in how likely someone is to report it to their doctor, etc.

[mrosett]

Look, just because a bias exists doesn't mean that it explains every single bit of improvement. We have randomized controlled trials showing that these drugs work.

[HeadsUpHigh]

Immunotherapies have a significant increase in survival rates of metastatic cancers i.e. cancers that were for the most part not detected early.