Deriving cut-off values through meta-analysis of individual studies