Deriving low-risk gambling limits from longitudinal data collected in two independent Canadian studies