Premix versus On-Demand Workflow Models in a Veterans Affairs Hematology/Oncology Pharmacy: Impact on Estimated Waste Costs and Patient Wait Time

• Published in: Journal of pharmacy practice 2024-10, Vol.37 (5), p.1175-1182
• Main Authors: Pan, Samantha, Gritte, Anndee, Malamakal, John, Horowitz, Amy, Harris, Alexa, Vig, Sierra J.
• Format: Article
• Language: English
• ISSN: 0897-1900; 1531-1937; 1531-1937
• DOI: 10.1177/08971900241245465

Background: The balance between reducing patient wait time and mitigating waste of parenteral products has not been well described in literature. Objective: Evaluate the patient wait times and cost-effectiveness of employing a premix versus an on-demand workflow model for compounding parenteral admixtures in a hematology/oncology infusion setting. Methods: This single center, retrospective cost analysis compiled manually documented monthly waste reports and estimated drug pricing for the institution to calculate the cost of waste during both premix and on-demand compounding workflows. Time to administration was audited for one week with both models. Results: Over a period of 28.5 months following the premix model, 564 products were documented as wasted ($1,196,014.01 in estimated drug purchasing cost). Over a period of 3 months following the on-demand model, 12 products were wasted ($34,823.98 in estimated drug purchasing cost). Switching models reduced the monthly average number of wasted products from 20 to 4 per month; the average cost of waste was reduced from $41,965.40 to $11,607.99 per month (P < .0001). Overall patient wait time from clearance until administration, excluding any recommended wait times after premedication administration (if applicable), was similar in both models: an average of 38.26 minutes in the premix model and 40.97 minutes in the on-demand model. Conclusion: Premixing parenteral admixtures was not cost effective at our institution. After resuming an on-demand compounding model, the monthly cost of waste (based on drug pricing alone) was reduced by over 70%. The wait time from clearance to treatment administration was similar in both models.