Development of a productivity model at a pharmacy consolidated service center.

Background: Health systems have implemented pharmacy consolidated service centers (PCSCs) to address increased patient volume, elevated drug costs, and decreased reimbursements. Assessing pharmacy productivity remains a challenge given that metrics have historically been determined by calculations of variables that do not capture the actual work. Several investigators have demonstrated improved labor outcomes in health-system pharmacy with the use of novel productivity models. However, the utility of a novel productivity model at a PCSC has not been assessed.

Objective: This study aimed to develop a productivity model with validation by comparison to past time periods to represent work at a PCSC.

Methods: The amount of time needed to complete work was determined by performing time studies. A modified Delphi process was used to ensure an appropriate perception of workload. Time standards for each category were averaged to determine the specific relative value units, which were then multiplied by total biweekly orders and combined with fixed activities to determine the unit of service. Actual hours worked were obtained for 6 prior pay periods to compare tool productivity with actual productivity.

Results: Time studies were performed over a 3-month period. The total average hours per pay period calculated by the tool for repackaging was 167.4 or 2.1 full-time equivalents (FTEs) and for warehousing was 176.8 or 2.2 FTEs. Although tool productivity followed the same trends as historical calendar day productivity, it was consistently higher per pay period over the 12-week comparison.

Conclusion: By performing time studies, a productivity model was developed for a PCSC that generated productivity data that correlated with 12 weeks of data using a historical model. This study provides the ability to assess trends over time with a more precise evaluation of work leading to the discussion that this tool is superior to historical productivity models.