Developing a Cloud-Enabled AI System for Pharmacovigilance: Automated Detection of Adverse Drug Reactions from Patient Reports

Abstract

An essential part of pharmacovigilance is the identification of adverse drug reactions (ADRs) associated with medicines. Traditionally, ADRs have been reported by health professionals, but there is growing recognition that consumers should likewise be included. Many countries now have online systems that enable consumers to self-report. Nevertheless, the vast majority of patient reports remain unused because there are too many for manual analysis. To exploit these valuable self-report data, a cloud-enabled AI system has been developed capable of automatically detecting ADRs in patient reports. The viability of the paradigm has been demonstrated by having the system identify ADRs in a large number of patient reports, and the system architecture has been designed to facilitate easy redeployment on various cloud platforms.

 Cloud computing is now a common paradigm in a range of application domains, having emerged largely because it provides an effective means of supporting high-level business functions. Many industries, including healthcare, have embraced the idea that consuming computing services as needed — and paying for only those resources that are used — makes good business sense. Implementing pharmacovigilance with the assistance of cloud computing and AI is an innovative way of exploiting the best that current technology has to offer.