Lead Time: survival went up, nobody lived longer

Twelve patients. Same disease, same deaths, in the same years. Start screening earlier and watch the headline survival statistics climb — then check the one number that matters. This is the bias that makes screening programmes look like cures.

life (undetected disease) counted as “survival after diagnosis” death screening starts
--
average survival after diagnosis
--
“5-year survival rate” (the screening headline)
--
when the 12 patients actually die

The wait, what?

Drag the screening start earlier. The survival numbers explode — and not a single death date moved. Every patient still gets sick in the same year and dies in the same year. The “extra survival” is time spent knowing about the disease, not time spent alive.

This is lead-time bias, and it is why “5-year survival” is the most flatterable number in medicine: diagnose earlier and you automatically “survive longer,” even if the treatment does nothing. The statistic that would actually prove a screening programme works is mortality — are fewer people in the population dying of the disease? — and that one does not move here.

Same lesson as everywhere else in this collection: a statistic measures where you started counting. Before celebrating any number, ask what it counts, from when, and who chose the starting line.