Erreurs de prescription des traitements anticancéreux : état des lieux dans un hôpital universitaire belge et estimation bicentrique de l'impact du logiciel Chimio

Details

Supervisors

Dalleur, Olivia ; Leclercq, Gisèle

Faculty

Faculté de médecine et médecine dentaire

Degree label

Mémoire de master en sciences pharmaceutiques

Abstract

Background: Chemotherapy circuit is a high-risk process. Aims of this observational study were to evaluate the most usual prescribing error (PE) types and their severity before computerized physician order entry system's (Chimio®) implementation. Then, we evaluated Chimio®'s ability to prevent these error types. This study has been conducted in 2 hospital pharmacies; one where Chimio® was recently implemented and the other where it was employed for 10 years. Method: In the first hospital, specific taxonomies for parenteral antineoplastic prescribing error types and severity were created. Prescriptions with PE were retrospectively selected over a one-year period before Chimio®'s implementation and analyzed according to these taxonomies. A batch of prescriptions was selected and a simulation was performed with Chimio®. Furthermore, we observed error types during 6 months after Chimio®'s implementation. In the second hospital, where Chimio® has been implemented for years, we collected errors during 15 days with our propriety tool. The purpose was to detect the software's limits. Results: On the first site, 231 prescriptions with 991 PE were eligible. The 3 most frequent error types were incomplete prescriptions, body measurement and identity errors. More than 50% of error types had potential severe or fatal consequences. Simulation with Chimio® demonstrated that45,1 % of error types becarne impossible and 43.7% unlikely. However, it doesn't prevent every error type (11.2%) and some new ones appeared, such as wrong settings.On the second site, 19 prescriptions with 43 mistakes were eligible. The 3 most frequent error types were unit mistakes, double prescription, and gap between clinical trial handwritten and computerized prescriptions. Conclusion: With electronic prescribing, we can expect a decrease of at least 50 % of PE. However, because of error types not blocked by the system and new types emerging, the pharmacists have to maintain their vigilance during pharmaceutical validation and parameterization.