Background Adverse drug events (ADEs) pose a heavy burden on health systems worldwide. The assessment of risk tool (ART) was developed to help prioritise hospitalised patients for interventions such as medication reconciliation (MR) and review, and to reduce potential ADEs from unintentional medication discrepancies and prescribing errors.

Aim To validate ART, and to determine which of the selected 25 flags were significantly associated with risk of unintentional medication discrepancies and prescribing errors.

Method This was a prospective observational study of 247 admissions to two medical teams in a New Zealand hospital. Study pharmacists undertook MR and review for all eligible admissions over 5 months. A Kruskal–Wallis test was used to determine differences in numbers of unintentional medication discrepancies and prescribing errors, followed by generalised linear modelling, to find the estimated ratio of the mean number of unintentional medication discrepancies between the three ART groups. In part 2 of the study, using multivariable logistic regression, combinations of flags significantly associated with risk were identified.

Results Significant differences in the number of unintentional medication discrepancieswere found between the ART risk groups (p<0.0001, Kruskal–Wallis test). Patients in the high-risk group had a significantly greater number of unintentional medication discrepancies than those in the medium- and low-risk groups. There were no significant differences in prescribing errors between the ART risk groups (p=0.08, Kruskal–Wallis test). Flags significantly associated with risk, included ‘more than eight admission medications’, OR=3.7 (95% CI 2.2 to 6.4) and ‘readmission within 30 days’, OR=6.8 (95% CI 3.0 to 15.2), p<0.0001.

Conclusion ART was effective for prioritising patients for interventions such as MR.