However, this is a bit misleading in a real world environment. Screening is intended to pick up abnormalities, not diagnose cancer. A false positive will not automatically result in undergoing chemo, even if it does result in other procedures which are ultimately unnecessary.

Furthermore, in the real world, the numbers are very messy. One person's interpretation of a pap smear can differ from another's, which results in apparently very different numbers in different testing sites, and consequently different (and misleading) apparent outcomes. Not to mention the budget consequences.

The conservative course is generally to treat a borderline case as positive, which is not necessarily a sound approach.

It is possible to improve the screening process by various methods such as raman spectroscopy. This pretty much takes the human judgement out of the equation and gets a computer to do it. And for that very reason, it scares the living bejesus out of insurers - wrongly.