According to one method for using machine learning in detecting drug diversion, the method comprises receiving, as input, an observed drug dispensation amount associated with a drug dispensation event related to a drug provider along with other drug dispensation event data, wherein the drug dispensation event data comprises drug provider information, procedure information, and patient information; generating, using a trained drug diversion detection algorithm and the drug dispensation event data, an expected drug dispensation amount associated with the drug dispensation event and determining, using the observed drug dispensation amount and the expected drug dispensation amount, whether the observed drug dispensation amount is aberrant, wherein the drug diversion detection algorithm includes at least one machine learning algorithm and is trained using one or more data sets associated with related drug dispensation events; and outputting, by the drug diversion detection algorithm, information indicating that the observed drug dispensation amount is aberrant.