ABSTRACT
An important source of signals in drug safety has been, and will continue to be, the observations of clinicians who report adverse drug reactions (ADRs). Automated methods are routinely used to mine databases of spontaneous reports to detect drug safety signals. Data mining is the discovery of interesting, unexpected, or valuable structures in large datasets. Modern data mining combines statistics with ideas, tools, and methods from computer science, machine learning, database technology, and other classical data analytical technologies (Hand 1998). In the context of signal detection in the pharmaceutical sector the interest is to detect local structures or patterns and to determine if they are real or chance occurrences. Patterns are usually embedded in a mass of irrelevant data. Interesting patterns can arise due to artifacts of the data recording process or genuine discoveries about underlying mechanism. Therefore, deciding whether a pattern is ‘interesting’ should be done using knowledge from experts to understand exactly what is being described. Increasing number of large databases maintained by various regulatory agencies and pharmaceutical companies around the world provide opportunity for some novel exploration in post marketing drug safety.
