I assume if we have a history of data about a person(1 or 2-day sleep data) then we could detect early signs of infection by respiration and heart rate changes, or from more disrupted sleep.

If data could be measured in non-clinical settings, then we may more closely approximate active infections?

What do you think?