Reliability of the chemotherapy preparation process: double check or gravimetric control?
3 October 2013Laurent Carrez, Lucie Bouchoud, Sandrine Fleury, Christophe Combescure, Farshid Sadeghipour, Pascal Bonnabry Pharmacie des Hôpitaux Universitaires de Genève, Section des sciences pharmaceutiques, Université de Genève, Université de Lausanne, Genève, Suisse
Centralized chemotherapy units have established systematic prevention strategies such as double check and gravimetric control to detect preparation errors.
The aim of our work was to provide scientific answers regarding the effectiveness of the error prevention strategies by the evaluation of two safety levels (visual double check and computer-assisted gravimetric production control) in comparison with a level involving standard operation protocol without control.
A simulation study (using phenylephrine and lidocaine as markers) was carried out in an operational context. For each of the three levels studied, 18 sessions were performed by 11 operators. During each session, 4 preparations with variable doses and volumes, using two different concentrations of stock solutions, were made for each marker (8 preparations in total). Each preparation was assayed by a capillary electrophoresis method with ultraviolet/visible–diode array detection. The results were analyzed according to a specific qualitative (choice of stock solution) and quantitative (accurate: deviation <5% of the target concentration, weakly accurate: deviation 5-15%, inaccurate: deviation 16-30%, error: deviation > 30%) methods.
No final preparation contained the wrong drug. The protocol without control did not detect 1 preparation dose error out of 3 performed (2 detected by the operator) and the double check control 2 errors out of 6 (4 detected by double visual check). The gravimetric method did catch all errors (5 detected). The accuracy of the doses measured by the quantitative analysis of final preparations was equivalent for the 3 levels (p = 0.63 Kruskal-Wallis). The distribution of the preparations was as follows: 58-60% accurate, 36-38% weakly accurate, 4% inaccurate and 0.7% of preparation errors. A high variability in accuracy was observed between operators (odds ratio of 20 for a risk of deviation > 5% of the less accurate operator compared to the most accurate one).
The gravity control was the only method able to prevent all dose errors but without improving the accuracy of the preparations. A dose accuracy <5% cannot always be guaranteed by manual production, even assisted by a gravimetric control and this result raises the question of the role of dose-banding to improve the accuracy of preparations.