Discrete time homogeneous backward semi-Markov reward processes and an application to disability insurance