Background The prescription system known as RELE is the electronic prescribing and dispensing system that has been used by the Balearic Islands Health Service (IB-SALUT, Spain) since April 2006. This repository contains patients’ computerised medication profiles (CMPs). Data from studies conducted in other settings have shown the existence of discrepancies between the medications recorded in patients’ CMPs and the medications actually taken by the patients.

Methods Chronic medications recorded in RELE CMPs were compared with those obtained through a structured clinical interview with patients or the persons responsible for the administration of the patients’ medications. A descriptive analysis of each discrepancy was performed. Potentially associated factors were analysed using Pearson’s χ² test.

Results Of the 215 CMPs studied, 168 (78%) had discrepancies. The rate of discrepancies in the general population inferred from these results ranged from 72.57% to 83.71%, with a 95% confidence level. Polypharmacy was associated with the presence of discrepancies (χ²: 39.998, 1 df, p<0.0001) as well as patient age >65 years (χ²: 18.949, 2df, p<0.0001) and higher comorbidity (χ²: 12.548, 2 df, p=0.002). The association between the types of discrepancies and their causes showed that prescriber-generated dosage errors were the most common errors (n=116; 27%), followed by patient-generated commissions (n=79; 18%), patient-generated dosage errors (n=77; 18%) and prescriber-generated commissions (n=73; 17%).

Conclusion The rate of discrepancies identified in this study shows that clinical interviews with patients during transitions of care continue to be essential.