Developing indicators of inpatient adverse drug events through nonlinear analysis using administrative data.
This study demonstrated the value of hierarchically optimal classification tree analysis as a promising method for evaluating and providing surveillance of administrative data for adverse drug events.