SmartCompounders Chemo^® does it well perform? Statistical study and capability analysi

Introduction
To automate part of the production process of injectable anticancer drugs, our hospital acquired the SmartCompounders Chemo^® (SC^®) compounding robot. Preparations are dosed and controlled gravimetrically using internal scales with a precision of 0.05 g. The aims of this study are to evaluate the accuracy of the dispensed volumes and the performance of the internal scale of the robot by comparison with a certified external scale, and also to assess the capability of the SC^® in the preparation of injectable anticancer drugs preparation.

Materials and methods
Bags of 100 mL of 0.9% sodium chloride were weighed on a certified external scale (precision of 0.05 mg) before and after addition by the automated system of three different vehicles. These vehicles, used as placebos for injectable anticancer drugs, included water (density d of 1.00), glucose 50% (d = 1.18) and polysorbate 80 15% v/v (d = 1.01). Five volumes were tested: 3; 15; 30; 45 and 60 mL. For each volume, 21 samples were measured per vehicle. Various statistical tests were performed: Student’s t-test for paired data to 1) assess the accuracy of the internal gravimetric control compared with the reference scale, 2) assess the accuracy of the volume of solution dispensed compared with the theoretical volume and 3) the Kruskal-Wallis test to compare the performance of the automated system with the 3 vehicles. The alpha risk was set at 5%. Additionally, process capability (CP) and process capability index (CPK) indices were calculated.

Results
Regarding the statistical tests: 1) for water, the accuracy of the gravimetric control is demonstrated (p-value = 0.51), which is not the case for the other vehicles. 2) The volume injected by the robot complies with the theoretical volume for water only (p-value = 0.23). 3) The robotic’s performance with all 3 vehicles is similar for volumes ≥ 30 mL (p-value > 0.20). Concerning the capability analysis: for the 3 formulations, the CP and CPK indices exceed 1.33 for volumes ≥ 15 mL with a threshold of +/- 3%. For volumes of 3 mL, the indices reach 1.33 with a threshold of +/- 7.5%.

Conclusion-Discussion
This study highlights the accuracy of the dispensed volumes and the performance of the internal scale for water. Significant differences were observed for the other vehicles. However, the capability analysis demonstrates that the robot is capable of producing results that comply with the specifications when the indices exceed 1.33 at the set thresholds. Thus, this analysis concludes that the SC^®’s performances are satisfactory. Full qualification of the SC^® is curently in progress