Impact of alert rate and alert relevance on computerised alert effectiveness
This project is funded by NHMRC Program Grant 1054146
Project members Macquarie University
|
|
Project Members - External
Associate Professor Melissa Baysari
Associate Professor
Dr Wu Yi Zheng
Postdoctoral Research Fellow
Dr Amina Tariq
Maureen Heywood
Professor Richard Day
Related stream of research
Electronic decision support and human factors in healthcare
Project Status
Current
Centres Related to this Project
Project description and aims
Project main description
Computerized Provider Order Entry (CPOE) systems with computerized alerts are increasingly being adopted by hospitals all over the world as evidence of their effectiveness in reducing prescribing errors is established. Although much research has investigated alert impact on prescribing, no studies have focused on identifying how many alerts are too many. That is, at what point alerts begin to be ignored and overridden by prescribers. In this research, we will investigate this question using a habit development framework. We propose that over time, as prescribers encounter and override more alerts, the override response becomes habitual. Once habitual, prescribers automatically override the alert with little attention given to the alert content.
Aims
To determine whether:
- 1) exposure to a high rate of alerts results in alert override becoming a habit, leading to incorrect responses to alert recommendations;
- 2) exposure to a high rate of irrelevant alerts results in alert override becoming a habit, leading to incorrect responses to alert recommendations; and 3) alert rate and alert relevance influence perceived usability of a CPOE.
Design and method
Participants will complete the study using the training module of a CPOE system.This study will use a 3 x 2 design, where 120 university students will be randomly allocated to one of six experimental groups: High alert rate-High relevance, Medium alert rate-High relevance, Low alert rate-High relevance, High alert rate-Low relevance, Medium alert rate-Low relevance, Low alert rate-Low relevance. Participants will order 80 prescriptions and will also complete a short usability survey. Our primary outcome measure is the proportion of correct responses to triggered alerts.
Content owner: Australian Institute of Health Innovation Last updated: 11 Mar 2024 6:46pm