We consider a multistage variant of the classical stochastic capacitated facility location problem under facility disruption uncertainty. Two solution algorithms for this problem class are presented: (1) stochastic dual dynamic integer programming (SDDiP), the state-of-the-art algorithm for solving multistage stochastic integer programs, and (2) shadow price approximation (SPA), an algorithm utilizing trained parameters of the linear value function approximation to minimize an upper bound on the optimal objective value. Numerical investigations demonstrate SPA consistently outperforming SDDiP across all instances from an additional dataset adapted from renowned library ORLib.