Studying the Role of Patient and Drug Attributes on Adverse Drug Effect Manifestation Using Clustering

摘要

Adverse drug reactions represent the unwanted or undesired effects of drugs. Timely extraction of such effects is highly required so that early warnings can be raised against if any serious beforehand to save patients from any further loss. It also helps in framing alternate treatment plan. There are two different ways for identifying the side effects of drugs. First is premarketing trials, which are conducted before floating drugs into market. But this approach is not as effective as these trials are carried out on a restricted population for restricted time. That is why another approach called as postmarketing surveillance is used. Under this approach, data mining techniques have been applied frequently for finding adverse reactions of drugs. But most of the existing techniques are based on the assumption that all attributes are equally responsible for drug side effects which may not be applicable for all real-life cases. In this paper, we study the role of different patients and drug attributes in manifestation of adverse drug reactions using clustering technique.