This problem comes from concepts and practice of mathematical finance by Joshi Chapter 8 problem 9.
Develop a pricing formula for an American digital put option
Joshi's solution - He states that we simply need to compute $$e^{-rT}\mathbb{P}\left(m_{T}^{s} \geq k \right)$$
where $m_{T}^{s}$ is denoted as the minimum up to time $T$. I really do not understand where we arrives at this conclusion at all or how to solve the problem in any other way than just using Monte Carlo simulation since we are dealing with a dynamic stopping problem.