I'm guessing the big innovation here would be developing useful diagnostics that skew towards high specificity at the expense of lower sensitivity. Additionally, If you're collecting a bunch of datapoints much more frequently than current screening/testing regimes, You would need some smarter averaging to find trends and weed out many, if not all false positives.

I think Reliable, cheap, user friendly medical diagnostics would be a game changer, but of course the data analytics to make good use would be different than how current, one-off, medical testing is treated.