Optimising computerised alerts within electronic medication management systems: A synthesis of four years of research.

Abstract

Most studies evaluating the effect of computerised alerts embedded in electronic medication management systems (eMMS) on prescribing behavior demonstrate positive and often substantial effects. But many studies also report that doctors override computerised alerts, sometimes up to 95% of the time. Alert fatigue, due to excessive numbers of alerts being presented, is the primary reason for alerts being overridden. This paper summarises and sythesises a program of research undertaken to determine whether doctors working in a teaching hospital in Sydney, Australia, were experiencing alert fatigue, and to identify and implement strategies for alleviating alert fatigue. We synthesise several published studies adopting a variety of data collection methods (observation of prescribers as they interact with the eMMS, interviews with users, review of alerts generated in eMMS, and a Delphi technique) to present four key lessons learnt. These are: 1) the fewer alerts the better; 2) context of use matters; 3) people use systems in unexpected ways; and 4) user feedback is invaluable.