OBJECTIVE: To quantify the amount of potential prescription errors for anticancer drugs in order to improve the quality of care. SETTING: The cytotoxic reconstitution unit at the Laennec Hospital in Paris, France during 6 months in the year 2000. METHOD: Pharmacist carried out a systematic analysis (e.g. dose, protocol, physicochemical properties) of the prescriptions for anticancer drugs (ACDs) before the compounding and handling at the cytotoxic unit of the pharmacy took place. The detected errors and potential errors were documented and analysed. MAIN OUTCOME MEASURE: Numbers and kinds of medication errors. RESULTS: Documenting data was difficult because of the inadequacy of existing classifications that usually do not include potential errors in prescriptions for with ACDs. Despite the presence of a formulary on the clinical wards, 349 errors were detected and documented, mainly concerning pharmaceutical aspects. Physicians were not aware of this type of errors due to lack of studies about their impact. Teaching aids such as a computer program for prescribing and production of ACDs could help in minimising these errors. CONCLUSION: The 349 detected errors involved mainly the physicochemical properties of preparations. A computerised prescription network could possibly reduce the number of such errors. Furthermore, a redefinition of the classifications of errors for cytotoxic preparations seems desirable, and such classifications should include typical pharmaceutical problems, and potential errors that do not reach the patient.
OBJECTIVE: To quantify the amount of potential prescription errors for anticancer drugs in order to improve the quality of care. SETTING: The cytotoxic reconstitution unit at the Laennec Hospital in Paris, France during 6 months in the year 2000. METHOD: Pharmacist carried out a systematic analysis (e.g. dose, protocol, physicochemical properties) of the prescriptions for anticancer drugs (ACDs) before the compounding and handling at the cytotoxic unit of the pharmacy took place. The detected errors and potential errors were documented and analysed. MAIN OUTCOME MEASURE: Numbers and kinds of medication errors. RESULTS: Documenting data was difficult because of the inadequacy of existing classifications that usually do not include potential errors in prescriptions for with ACDs. Despite the presence of a formulary on the clinical wards, 349 errors were detected and documented, mainly concerning pharmaceutical aspects. Physicians were not aware of this type of errors due to lack of studies about their impact. Teaching aids such as a computer program for prescribing and production of ACDs could help in minimising these errors. CONCLUSION: The 349 detected errors involved mainly the physicochemical properties of preparations. A computerised prescription network could possibly reduce the number of such errors. Furthermore, a redefinition of the classifications of errors for cytotoxic preparations seems desirable, and such classifications should include typical pharmaceutical problems, and potential errors that do not reach the patient.