• Chart of the day: Lead time bias

    I grabbed this from a nice post on prostate cancer screening by Kenny Lin, which I recommend reading in full. (It’s short.) Aaron has made this lead time bias point many times. See his many survival rate tagged posts.

    lead time bias

    This is very important. If you don’t get it, ask in the comments, though maybe give Kenny’s and/or one of Aaron’s posts a read first (links above). H/t: Brad Flansbaum.


    • A question, not a comment.

      Is colorectal cancer screening going to turn out to be a similar story, or is it going to turn out to be the right thing? How many cancers do all that polyp snipping actually prevent? We’ve been at it a while, it seems, so is the colorectal cancer mortality rate per 100,000 population changing?

      From a brief google, I got the impression that there aren’t many naysayers here.

    • The window in which one can undertake effective clinical interventions varies dramatically from one type of cancer to another. There are enormous variations in the magnitude of the “therapeutic window” within specific sub-types of cancer that affect the same organs/tissues.

      There is a subset of cancers which are not terribly aggressive, and within which the lag between “screen” detection and clinical detection doesn’t have a significant impact on how long the patient ultimately survives, at least in the aggregate.

      Most forms of prostate cancer fall into that category. There are many others that do not, for which early detection is the single most meaningful determinant of survival. E.g early detection has a real effect on if, and for how long the victim survives.