Let :
$P$ and $Q$ two equivalent probability measures and $Z$ the Radon Nikodym derivative : $Z = \frac{dQ}{dP}$ .
$Z(t)$ is the expectation of the Radon Nikodym derivative $Z(t) = E^P \left[Z|F(t) \right]$, and $Y$ is an $F(t)-$ measurable process.
Following this definition we know that $E^Q(Y) = E^P(YZ)$. And we can prove that $E^Q(Y) = E^P(YZ_t)$.
My question : In the equality (Equality from Shreve - Stochastic Calculus for Finance II) below :
$$E_Q\left[1_A\frac{1}{Z(s)}E^P\left[ Y Z(t) | F(s) \right]\right] = E_P\left[1_AE^P\left[ Y Z(t) | F(s) \right]\right]$$
We apply $E^Q(Y) = E^P(YZ_s)$. Why choosing $Z_s$ and not $Z_t$ or others ? I mean how do we know the indice of $Z$ when doing a change of measure as we just did? $s$? $t$? ..
here's the attached context of my question :
