Skip to main content

More effective, acceptable prescribing alerts

Dr Gregory Scott MSc MBBS, Dr Priya Shah MBBS, Professor Jeremy Wyatt DM FRCP,

Dr Boikanyo Makubate PhD, Mr Frank Cross MS FRCS

Email: gregory dot scott99 at imperial dot ac dot uk




Prescribing alerts are an important tool to improve the safety of medicines management and are commonly implemented in GP and hospital e-prescribing systems. However, their design is rarely user friendly, so this function is often turned off. The result is that such decision support, which systematic reviews have shown to be effective at improving prescribing, is seldom available to the junior hospital doctor.

Project aims


The aim is to study factors that might make prescribing alerts more effective and acceptable to clinicians, and use these insights to design and deliver more useful alerts that interrupt the clinician less often but when they do, are highly effective.



The first aspect of alerts that has been studied is whether or not the alert interrupts the prescribing process. We developed a simulated e-prescribing system and 30 prescribing scenarios typical of those facing junior doctors. We then wrote prescribing alerts that would prevent the doctor committing a prescribing error associated with each scenario. In a randomised study using 24 junior doctors, we measured the impact on prescribing errors and user acceptability of alerts that did not interrupt the prescription process and alerts that did, requiring the user to click on an OK button (interruptive alerts).



The interruptive alerts reduced prescribing errors by a factor of 10, while alerts with identical wording but which did not interrupt the prescribing process reduced errors by only 3 times, i.e. interruptive alerts were over three times as effective as non-interruptive alerts. Both kinds of alerts were acceptable, but participants suggested that the interruptive alerts could be reserved for serious errors. We are now exploring the use of symbols to increase the salience of non-interruptive alerts.

NHS relevancePrescribing Alerts


This project fits with QIPP goals in the following ways:

Quality and safety will be improved through more appropriate prescribing alerts and reduced prescribing errors

Innovation will be demonstrated through new methods of alert delivery

Productivity will be improved by fewer interruptions of the prescribing workflow

Prevention: this project will prevent adverse drug events by improved medicines management