The document discusses a Bayesian confidence propagation neural network (BCPNN) technique that has been used since 1998 for data mining adverse drug reactions (ADRs) reported to the WHO. The BCPNN provides an objective initial assessment of all drug-ADR combinations to transparently select combinations for clinical review. It calculates the strength of dependencies between drugs and ADRs using a measure of disproportionality and Bayesian statistics. The BCPNN enhances early signal detection from the WHO ADR database without replacing traditional methods.