Modelling the interactive behaviour of users with a medication safety dashboard in a primary care setting

摘要: Abstract Objective To characterise the use of an electronic medication safety dashboard by exploring and contrasting interactions from primary users (i.e. pharmacists) who were leading intervention secondary non-pharmacist staff) used to engage in safe prescribing practices. Materials methods We conducted a 10-month observational study which 35 health professionals instrumented for audit feedback purposes clinical practice as part wider study. modelled user interaction computing features representing exploration dwell time through interface events that logged on remote database. applied supervised learning algorithms classify against users. Results observed values accuracy above 0.8, indicating 80% we able distinguish user. In particular, Multilayer Perceptron (MLP) yielded highest precision (0.88), recall (0.86) F-measure (0.86). The behaviour was distinctive they spent less between mouse clicks (lower time) screens showing overview trends. Secondary exhibited higher more visual search activity (higher exploration) displaying patients at risk visualisations. Discussion conclusion interactive care using timestamped events. Primary competent population monitoring activities, while struggled activities involving detailed breakdown patients. Informed these findings, propose workflows group adaptive nudges increase engagement.