StatsMiniBlog: Lead time bias

http://sci-med-cartoonery.tumblr.com/post/120663951238/whats-so-good-about-early-anyway

This is a cracking cartoon that highlights everything that’s misleading about lead time bias. And will also lead to the terrible situation of the parents of the bird having to try to keep it away until proper bedtime to stop this happening tomorrow too.

Lead time bias is the apparent improvement in duration of a condition (often length of survival or time to relapse) without actually changing the age at death/relapse etc.

A great example is in neuroblastoma, which could (theoretically) be screened for in asymptomatic infants with urinary catecholamine testing. Doing so certainly seems to detect more children with neuroblastoma, but the age of death from neuroblastoma did not effectively alter both during and after screening was introduced and stopped in Japan.

Once again, we have to look very carefully about what outcome measures actually mean.

– Archi

(Visited 217 times, 1 visits today)