Asynchronous Pharmacy Decision Support Within a Multi-Institutional Telemedicine Infrastructure

Our health system does not yet have physician order entry to facilitate synchronous pharmacy decision support. To fill a needed gap to screen drug orders, we leveraged an enterprise clinical data repository to develop and implement two asynchronous, rule-based expert systems for screening drug dosage regimens and drug-drug interactions (DoseChecker and PharmADE, respectively). Originally deployed in batch mode operating once every 24 hours, these systems now operate in near real time and use a variety of notification technologies for routing clinical alerts. When a new prescription satisfies an alerting rule, an alert is generated and then routed to a responsible pharmacist. Database triggers are also used to generate alerts for existing orders on the basis of changing patient laboratory data. Both expert systems have had a demonstrable effect on prescribing errors, even though the intervention is downstream from the ordering process and is essentially transparent to the ordering physician. Asynchronous decision support is an effective means of correcting prescribing errors and improving the drug-ordering process.