Implementation and Evaluation of Robotics Intravenous Compounding Technology in a Tertiary Hospital Oncology Pharmacy

Purpose: The objective of this study was to compare the compounding efficacy, accuracy, and pharmacy technician repetitive strain injury risks between manual pharmacy technician compounding and robotic compounding processes. Method: This quasi-experimental study at Houston Methodist Hospital Oncology Pharmacy analyzed three weeks of manual compounding data and three weeks of robotic compounding data. The primary endpoint was the medication turnaround time for both manual and robotic methods, calculated from the start to the completion of dose preparation. Secondary endpoints included medication rejection rates, number of doses compounded per technician hours, medication preparation time, and the risk of repetitive injuries for both compounding methods. Statistical analysis used included t-tests, segmented regression, and chi-square tests to evaluate preparation times, failure rates, and ergonomic risks. Results: A total of 447 medications were prepared manually and 200 doses were prepared via robotic compounding during the study period. The mean turnaround time for manual preparation was 26.6 minutes, compared to 29.0 minutes for robot preparation (95% CI, -5.1-0.1; P=0.053). A total of 2.3% of doses were rejected in manual preparation and 1.5% of doses were rejected in robot preparation (P=0.118). The mean manual preparation time was 15.7 minutes versus the robot's mean preparation time of 6.9 minutes (95% CI, 8.0-9.6; P<0.0001). The compounding efficiency was 0.78 doses per technician hour, indicating a 14% increase in compounding efficiency. Regarding employee safety, for manual compounding, two observations fell in the medium risk range, and two observations were in the high-risk range. For robot compounding, all four observations were in the low-risk range. The Chi-Square Test of Independence was performed to examine the relationship between the method of compounding (robot vs. manual) and the risk of repetitive strain injury (low, medium, high). The result was X^2=8 with a significance level of P=0.018. Conclusion: Robotic compounding shows comparable results in medication turnaround time. It also offers significant advantages in reducing the risk of repetitive strain injuries associated with manual compounding for pharmacy staff. Moreover, robotic compounding has a similar dose rejection rate compared with manual compounding. However, it faces technological limitations and requires future advancements in technology and workflow design to fully optimize automation in oncology pharmacy operations.