Concepts
Non-clinical sciences/Stats & evidence · Bias & errorlow yield
Lead-time & length-time bias
[object Object]
What it means
Lead-time bias: screening advances the diagnosis date so survival-from-diagnosis looks longer though death is unchanged. Length-time bias: screening preferentially catches slow-growing (better-prognosis) disease, overstating benefit. Overdiagnosis: detecting disease that would never have caused harm.
Worked example
A screened cancer cohort 'survives longer from diagnosis' purely because diagnosis was made earlier — lead-time bias.
In the exam
A screening programme appears to improve survival from diagnosis, but only because the disease is found earlier while the date of death is unchanged.
What settles it
Lead-time = earlier diagnosis inflates survival; length-time = indolent cancers over-represented in screened cases.
Classically confused with
Selection bias
Source: StatPearls