Document resource
Background and Importance Automated unit dose dispensing systems aim to improve efficiency, traceability, and medication safety. A custom hybrid version of ACCED v3 (Deenova) was implemented, representing the only deployment of this configuration. Previous reports focus mainly on expected performance, but few describe real-life use. Assessing the gap between expected and observed performance was necessary to address organisational and technical issues.Aim and Objectives To evaluate the implementation of ACCED v3 Hybrid (Deenova) in a hospital pharmacy, focusing on technical performance, organisational impact, and medication safety in a long-term care unit.Material and Methods A retrospective observational study was conducted from 4 February to 4 September 2025 in a 52-bed Long-Term Care Unit (USLD). Three hybrid systems produced weekly nominative booklets (about 3100 sachets/week). Two hybrid workflows (v1 and v2) were used during the study. Outcomes included incidents, non-conformities, productivity indicators, and IT issues; outcomes focused on technical performance, organisational impact, and medication safety.Results Expected picking productivity (300 sachets/h) was not reached: v1 achieved 270 sachets/h with manual corrections for missing sachets; v2 reached 170 sachets/h without missing sachets. Filling productivity was 480 doses/h versus 500 expected. From June to August, 184 incidents were recorded: 51.6% technical filling, 3.8% IT filling, 28.3% technical picking, and 16.3% IT picking. Hotline response time exceeded 2 hours. Thirty-seven bug types were identified; 12 persisted. Data extraction was incomplete due to software limitations. Non-conformities decreased from 1% (33% of patients) to 0.3% (8%). Nurse-detected errors were 0.03%. Non-conformity types fell from 11 to 3. Preparation time increased sixfold, and nurses’ rounds doubled. The supplier estimated 0.5 full-time equivalent (FTE) was needed, but 2 FTE were required. Consumable costs averaged €470/month; ~3100 sachets/week were discarded.Conclusion and Relevance These observations are consistent with limited published data on hybrid automated dispensing systems, highlighting similar operational constraints in real-life settings. Implementation revealed gaps between expected and observed performance. Scaling to 1200 patients was not achieved. Software limitations, increased workload, and incident persistence were observed. Three different systems have been used since deployment. Hybrid models presented maintenance and monitoring constraints; appropriate version selection, clinical team involvement, and on-site technical support were required. Shifting to a non-hybrid version is under consideration.Conflict of Interest No conflict of interest