Background Because pharmacists do not yet systematically analyse prescriptions closely due to lack of time and resources, tools to optimise pharmaceutical validation must be developed.

Purpose To identify pertinent criteria to target high-risk patients or drugs.

Materials and Methods A prospective 3-month study was performed. Daily prescriptions were filtered using an automated computerised system based on selection criteria identified from an analysis of the literature. Pharmaceutical interventions (PIs) that were necessary as a result of the selection criteria and the rest of the prescription were analysed.

Results We analysed 1122 prescriptions, corresponding to 456 patients (mean age 79 years). Prescriptions were grouped together according to several selection criteria: creatininemia ≥125 µmol/l (62.1%), dyskalaemia (23.2% including 55% hyperkalaemia and 45% hypokalaemia), INR ≥ 4 (7.2%) and drug dosages (7.5% including 37% vancomycin, 22% digoxin and 15% gentamicin). Among the prescriptions selected for creatininemia ≥125 µmol/l, 45.7% were associated with severe kidney failure and 26.4% with moderate kidney failure.

The rate of PIs was 16% (n = 179) and varied according to the criteria: 22% and 19.6% respectively for severe and moderate kidney failure, 33.3% for hyperkalaemia, 10.3% for hypokalaemia, 13.8%, for INR, 4.4% for drug dosages.

The PI was related to the selection criteria in 77% of the cases. The rate of acceptance by prescribers was 80.6%.

Conclusions Carefully choosing pertinent criteria to philtre prescriptions seems to be an interesting option to optimise our approach to clinical pharmacy.

Our study shows that criteria associated with kidney failure and hyperkalaemia seem to be pertinent as they are related to numerous PIs. Following up INR > 4 should also be added as a pertinent criterion, as anticoagulants are often the cause of iatrogenic events.

This type of selection could be used to perform a transversal analysis of prescriptions, without discrimination by hospital unit.

No conflict of interest.