Background The constantly growing incidence of cancer and long term treatment are leading to an increasing number of cytotoxic preparations in hospital pharmacies. Quality standards for cytotoxic preparations are essential to assure treatment efficiency and limit iatrogenic toxicity.

Purpose To establish a quality control that ensures traceability and safety in the preparation of cytostatic drugs as well as ensure consistency between prescription and product made to minimise errors such as administration of defective chemotherapies.

Material and methods Gravimetric method for qualitative and quantitative control of cytostatic drugs was computer aided in all stages. The method consists of three weighings: just before injection of cytotoxic drugs, weigh the dose of cytotoxic and weigh the bags containing solutions and drugs just after injection of the cytotoxic drug. This weight depends on the volume injected and the density of the cytotoxic solution. The volume depends on the prescribed dose of the cytotoxic drug and its concentration. For each active ingredient, the density value was collected from the supplier beforehand.

It allows comparison between the exact amount of drug added to the mixture and the amount of drug prescribed, qualitative control by uniquely identifying products used by data matrix codes and traceability of the batch used, and finally the control of all of the processes.

Descriptive retrospective observational study between October 2014 and August 2015. We calculated the following indicators: degree of coverage (%) of technological qualitative control and rate of defective preparations (DP) intercepted (DP×1000 preparations).

Results During this period 6420 preparations were prepared. Quantitative control coverage was 82.3% (5347 preparations) and qualitative control coverage was 83.4% (5352 preparations). 347 errors were detected: 61(0.9%) by gravimetry and 286 (4.5%) by qualitative control. Global error rates intercepted were 11.4 DP×1000 preparations by gravimetry and 53.4 DP×1000 preparations by data matrix reading.

Conclusion This method improved quality and safety because it allowed errors in preparation of antineoplastics to be corrected in real time and so were prevented from reaching the patient, and avoided us having to repeat or discard defective preparations with economic losses. It is necessary to learn this system because it allows full traceability and real assess to the intercepted errors.

No conflict of interest.