Abstract: Objective To optimize signal detection model to extract more valuable signals in adverse drug reaction monitoring system in china. Methods By data quality controlling, optimizing signal detection methods, building signal screening strategy, the signal detection model was optimized. Results The variable matching method was built to find and delete duplicate cases. Signal detection methods have been optimized by Shrinkage. Signal screening strategy was built to rank signals, which had veaducedced the false positive signals. Conclusion Optimized signal detection model can extract more valuable signals from ADR cases database.

Key words: adverse drug reaction, signal detection, optimizing