Identifying older inpatients at high risk of unintentional medication discrepancies: a classification tree analysis

Unintentional medication discrepancies at admission are differences between the best possible medication history and the prescribed treatment at admission, and are associated with adverse outcomes, particularly in older people. This study aimed to identify the clinical profiles of geriatric inpatients with unintentional medication discrepancies at hospital admission. A classification tree Chi-square Automatic Interaction Detector (CHAID) analysis was conducted to assess those patients' profiles and characteristics that were associated with a higher risk of unintentional medication discrepancies. One-hundred and thirty consecutive older patients admitted to acute care (87 ± 5 years old; 61.8% women) were assessed. The CHAID analysis retrieved 5 clinical profiles of older inpatients with a risk of up to 94.4% for unintentional medication discrepancies. These profiles were determined based on combinations of three characteristics: use of eye drops, frequent falls (≥ 1/year), and admission due to urgent hospitalization. These easily measurable clinical characteristics may be helpful as a supportive measure to improve pharmacological care.