Background The ‘Study on patient safety in primary health care’ (APEAS), published in 2008 by the Spanish Health Ministry declared that 48% of adverse events (AEs) detected in these patients were due to medicines errors (MEs). The Institute for Safe Medication Practices (ISMP) promotes the development of internal systems to report medicines-related incidents in hospitals in order to achieve effective preventative measures.

Purpose To analyse total errors in an intravenous mixing unit and establish checkpoints to prevent them.

Materials and Methods Prospective observational study (August–December 2011) which included outpatients who might be exposed to an error with intravenous medicines. The variables were: Wrong drug, original prescription service, prescription type (manual or printed), who detected the error and process error (prescription, validation, preparation or administration). Errors were classified according to severity category and error type based on the taxonomies listed in ISMP Spain. The errors observed and reported by the staff involved with the process were recorded by the pharmacist. The differences between frequencies were checked with the Chi-Square statistical test.

Results The total error frequency (EF) was 1.27%. The drugs most frequently involved were natalizumab (2.43%), infliximab (1.23%) and intravenous immunoglobulin (1.23%). No statistically significant differences between EF of each drug and the mean frequency were detected (P = 0.94, 0.76 and 0.94). The services involved were: Gastrointestinal (2.98%), Neurology (1.57%), Rheumatology (1%), Haematology (0.15%) and Oncology (0.035%). Only in the Haematology and Oncology services were differences from the average found (P = 0.038, p = 0.001). Most failed orders were manual (67%). All incidents occurred in the prescribing process and were detected by the pharmacist during validation. No errors reached the patient (category B). In the classification by error type: 67% were incorrect date (periodicity in the cycle), 22% dosage (50% excess) and 11% in the rate of administration.

Conclusions After reviewing the results we can assume that the main checkpoints where our activities should focus on are the following: incorrect date, dosage and rate of administration.

A possible methodological bias can be considered because the data were collected in the pharmacy unit and all errors were prescription errors – no pharmacy or process errors.

No conflict of interest.