The 2nd IS-HEART session of Feb. 2015 will be scheduled in week 4 instead of week 3.
Title: Physician Order Alert Assessment Support using Data Mining Techniques.
When: 27/02/2015, 12:30-13:30
Also see the online agenda item.
Abstract: During the last four decades there has been a major increase in the number of automatic reporting systems for adverse drug events (ADEs)in the healthcare setting. Computerized prescription alerts can assist in avoiding ADEs but research has indicated that one important example of an ADE is the situation where a medical specialist fails to adequately assess computerized prescription alerts. These computerized prescription alerts can be shown to a physician when for example there is a problem with the dosage of a prescribed medicine or an interaction with other medicines is found. Hospital Pharmacist Drs. P.A.G. De Klaver indicates that Maxima Medical Center’s Pharmacy has to deal with 500-800 prescription alerts on an average day, which is considered to be too high and time consuming. The alert system that is currently in use at the hospital is emphasizing the breadth of coverage instead of clinical relevance or severity of the event, which leads to low clinician acceptance of drug alerts. Many alerts can cause alert fatigue; a state in which the user becomes less responsive to alerts in general.
The aim of this research project is to find data mining techniques to create models that can support processing of prescription alerts at the Maxima Medical Center in Veldhoven. This research will not focus on all ADEs but only on drug-drug interaction effects and their alerts because those alerts are considered to be the most time consuming for pharmacy personnel.
In this presentation, an overview of recent work on improving healthcare alerts systems is given and the overall research proposal is clarified.