Background Patient safety is the most important concern for healthcare professionals, patients and healthcare systems. Adverse drug events (ADEs) are a common cause of hospitalisation and occur with increasing frequency in hospital as patients age.

Purpose Determine the ADEs at admission and during the stay in a psychogeriatric unit.

Material and methods One-year prospective cross-sectional study (July 2015–2016), in a psychogeriatric ward (21 beds). Included patients with dementia, admitted presenting neuropsychiatric/behavioural and psychological symptoms in dementia (BPSD). Excluded patients with previous psychiatric illness or palliative conditions.

Prescription information: Aegerus and Catalonia clinical record HC3. Variables: demographics, diagnoses, Global Deterioration Scale (GDS-R), functional assessment (Barthel Index), fall risk (Downton JH and Tinetti). Anticholinergic risk level (Drug Burden Index (DBI). ADE assessment: causality by Naranjo algorithm, severity by classification system of the Institute for Healthcare Improvement and preventability by the Schumock–Thornton algorithm. Drugs classification by the Anatomical Therapeutic Chemical Code (ATC).

Results Sixty patients (60% females), age: 84.8 years (68–96). Dementia: unidentified (43%), Alzheimer’s (31%), Lewy bodies (8%), vascular (8%), mixed (5%) and others (6%). GDS-R 4.5 (±1.2), moderate cognitive impairment. Barthel Index 43.8 (±23.9), moderate dependence. Patients comorbidities 4.8 (±1.6). Drugs/patient 9.03 (±3.12). DBI high risk in 69% of the patients. High risk of falls, Tinetti (15.5±8) and Downton test (4.5±1.3).

Sixty-eight ADEs (53 patients, 81.5%, 22.6% more than one). 73.5% not related to falls. From this 80% were related to the ATC nervous system (46% (23 ADE) psycholeptics). Naranjo algorithm one definite (2%), probable 45 (90%) and possible four (8%). Severity Category E: temporary harm to the patient and required intervention in 34 (68%) and F: temporary harm to the patient and required initial or prolonged hospitalisation in 16 (32%). Schumork–Thornton test 58% (29) of the ADE were preventable, possibly preventable 6% (three) and unavoidable 36% (18). Main ADE drowsiness/somnolence 27.7%, 12.8% weakness and hypoactivity and 10.7% hypotension. DBI significant differences related to fall as ADE result (fall group 1.288±0.79, non-fall group 0.95±0.71 p=0.05).

Conclusion The balance between effective treatments and safety is complex in these patients. Mostly ADE are related to the pharmacological treatment of this BPSD.

Anticholinergic load is a determinant for a specific ADE and were related to falls, the worst consequence in this patient, clinical and economic impact.

References and/or acknowledgements No conflict of interest.