The authors develop a coding theory approach to error control in residue number system product codes. Based on this coding theory framework, computationally efficient algorithms are derived for correcting single errors, double errors, and multiple errors, and simultaneously detecting multiple errors and additive overflow. These algorithms have lower computational complexity than previously known algorithms by at least an order of magnitude. In addition, it is noted that all the literature published thus far deals almost exclusively with single error correction.>