so, if you want to understand how to transform the canadian epidemiology into the american, and back, the formula should be something like this:

u(f(t)) = 10*c(t)

where,

u(t) = american deaths at time t
c(t) = canadian deaths at time t
f(t) is an equation that relates the domains of the two functions.

if we measure t in weeks and allow it to be linear, even though we know it isn't, and f(t) is roughly t-6 then,

u(t - 6) = 10*c(t),

so,

c(t) = 0.1*u(t-6)

to be a little more complicated, you'd need to derive a fourth order differential equation for t and solve it. you'd want it to be fourth order so that the derivative of the acceleration is still a non-scalar function of t. that is, the acceleration is not constant. and, this is an empirical question, so you could fit this, if you sat down and did it...