Automated Dispensing Cabinet Stocking Schedule and Inventory Management Optimization

Prior practice in a 523-bed community hospital required automated dispensing cabinets (ADCs) to be replenished daily by early afternoon to ensure accurate drug ordering. Because of the tight time window, technicians frequently refilled ADCs during peak medication administration times. After a positive report from a smaller facility in the health system, the study facility implemented a quality improvement project to optimize the ADCs. This involved new processes of every-other-day ADC replenishment staggered via shift-specific workflows, and a dedicated technician was assigned to optimize ADC inventories.

Monthly reports of ADC refilling and dispensing transactions were compiled from inpatient, nonprocedural units. Data on stock-out transactions, when a nurse attempts to remove a medication from an empty ADC pocket, were obtained from facility reports from the same machines. Based on hospital policy for standard medication administration times, peak times were identified. For these time windows, an overlap ratio was calculated by dividing the relative frequencies of refill and dispense transactions. The average overlap ratios were calculated for the pre- and post-implementation periods. Stock-out transactions were tracked monthly, and a 95% confidence interval was constructed for all months before implementation.

Post-implementation, monthly stock-out counts significantly decreased by 38% within 6 months. While there was no change in medication dispensing patterns, there was a major shift in refilling activity patterns. Overlap during morning peak times decreased, but overlap increased during the evening peak times due to the rearrangement of technician shift times and workflows.

After implementation of the new process and ADC inventory optimization, refill habits significantly changed, leading to reduced technician-nurse overlap during the morning medication pass, likely due to the spread out replenishment schedule. Dispensing habits did not change during the study period, and while there was a significant decrease in stock-outs post-implementation, there was insufficient data to assess effects of the new refilling processes.