Computerised decision support (CDS) holds great promise to improve medication safety in hospitals, but there is little guidance available to hospitals on what CDS should be implemented and what constitutes effective CDS.
Observations of doctors in practice, interviews with doctors and medication chart review indicated that doctors working at St Vincent's Hospital (SVH) were being presented with too many alerts while prescribing. A consequence of this over-alerting was that doctors were ignoring many alerts, even those that presented safety critical information.
Based on this research, several significant changes were made to the electronic prescribing system at SVH as well as to hospital policy and work practices. These included the removal of an entire alert type (pregnancy alerts), the redesign of many other alerts and changes to the electronic-prescribing system training which all doctors attend in order to prescribe medications in the hospital. SVH is actively evaluating and responding to research evidence to improve the safety and effectiveness of CDS. More widely, this research is being applied by other Australian hospitals as they seek to design and assess CDS in their hospital systems.
Dr Baysari, leading this research, was recently invited to present the research at NSW Health's Electronic Medication Management Clinical Leadership Forum, a regular forum for NSW hospitals to learn about effective design and approaches to system implementation. She was also invited to contribute to a workshop being run by the Australian Commission on Safety and Quality in Health Care at this year's Health Informatics Conference (HIC2015) titled 'Implementation of electronic medication management systems: 101 for Hospitals'. Dr Baysari's talk, titled 'How much is too much?' provided workshop participants insights into the design of effective CDS for prescribers.
AIHI research on CDS for antibiotic prescribing has shown that implementation of decision support is unlikely to result in improved antibiotic prescribing if introduced in a setting where ambiguity surrounds antibiotic policies and practices. Following an audit and feedback trial, several practical problems with the policy and antibiotic approval process were identified at a hospital site. These issues were communicated to the antimicrobial stewardship committee and this resulted in a number of significant changes to the policy (both paper format and embedded in CDS). More broadly, this research provides several general lessons for hospitals as they move from paper to electronic prescribing.