Survival rate sounds like it means cured. It does not. It means somebody was still alive at the end of a counting window. Usually five years. Follow what that means, because it changes everything.
Starting the clock sooner is not living longer
Imagine a disease that always takes a person at year ten. Nothing about the disease changes. But we get better at spotting it early, and now we find it at year three instead of year eight. Five-year survival goes way up. It looks like enormous progress. Nobody lived one day longer. We just started the clock sooner. This is called lead-time bias, and it is not a fringe idea. It is textbook epidemiology, taught in every program that trains people to read research. There is a second one that stacks on top of it: slow-growing versions of a disease sit in the body longer, so screening catches those far more often than the aggressive fast-moving ones.[2] The screened group ends up loaded with the cases that were going to do better anyway.[3]
The number that is hardest to dress up
I am not telling you screening is bad. Some screening saves real lives and I am glad it exists.[4] I am telling you that "five-year survival went up" and "people lived longer" are two different sentences, and they get used as if they are the same one. The number that is hardest to dress up is overall mortality.[1] Did fewer people die, of anything, in the group that got the thing? You only die once. That number cannot be moved around between columns. When you read about any treatment or any screening program, look for that one. If a study reports every number except overall mortality, that absence is information. Ask better questions. That is the whole game.
Key Takeaways
- "Survival rate" means alive at the end of a counting window (usually five years) -- not cured.
- Lead-time bias: finding a disease earlier raises five-year survival even if no one lives a day longer.
- Screening also over-catches slow-growing cases that were going to do better anyway, inflating the numbers.
- Overall mortality is the number hardest to dress up -- did fewer people die of anything? If a study reports everything except that, the absence is information.
Related
Want the next one of these?
The Signal is one email when there is something genuinely worth your attention -- written the way this post is, in plain language, with the studies behind it. No noise, no selling your address, and nothing to buy.
Spam protection is not configured for this form yet. Please use the email link instead.
The Signal, and nothing else -- no fixed schedule, only when there is something worth sending. We never sell or share your address, and you can unsubscribe in one click any time.