Baking Better With Cochrane-Orcutt (part 6)

Assessing the influence of mass viral testing programmes on case positivity: are we experiencing a testdemic? (rev 1.1)

Sometimes I bake a statistical model that smells too good to be true. Back in Baking Better With Cochrane-Orcutt (part 5) such a model popped out of the oven for the prediction of new cases by specimen date using just new virus tests by specimen date alone. Shortly after this I rumbled the UK GOV dashboard game of adding new reinfections to new first episodes in order to supercharge the daily case count, a short account of which may be found here.

This fine July morning let us focus on new first episodes and new virus tests and let us also focus on the period Oct 2021 through to Jun 2022, for this is when lateral flow device use seriously changed the counting game. If I plot these two series on a dual axis chart I get this pretty picture, which should explain why my Cochrane-Orcutt model is able to successfully predict new cases from tests conducted, thus lending weight to the notion of a testdemic: