Background Elderly patients suffering from many diseases and disorders are more likely to have multiple prescriptions. These multiple prescriptions could result in drug toxicity, reduce compliance and cost a lot.

Purpose Following the French health agency recommendations, we evaluated professional practise on drug prescriptions for very old inpatients of a university hospital.

Materials and Methods All prescriptions of 60 consecutive patients aged over 80 years admitted between November and December 2010 to the geriatric unit (35 beds) were evaluated following a grid. This grid contained 18 items divided into 4 themes:

• organisation of the prescriptions, drug schedule and dosage

• patients’ weight

• number and type of drugs (psychotropic drugs for instance), presence of contra-indications

• biological adaptations

Results Median age was about 88 years (range: 80–96) and 70% of patients were women. The median number of drugs on the 60 prescriptions was 8 (range = 1–12). All prescriptions presented data on dosage and the drug schedule. Drugs were classified by therapeutic category on only one prescription. Half of the prescriptions specified the patients’ weight. More than 80% of prescriptions had more than 5 drugs without redundancy concerning psychotropic drugs and non-steroidal anti-inflammatory agents. Four contra-indications were noticed among all prescriptions evaluated. All biological adaptations were followed.

Conclusions In conclusion, we notice a good level of quality concerning prescriptions in this geriatric unit where software-assisted prescribing with pharmaceutical analysis has been effective since 2009. This software does not allow physicians to organise prescriptions by disease area. Concerning the patients’ weight, senior clinicians will inform junior clinicians of its importance in the patients’ file and prescription. Another evaluation will be scheduled to analyse the link between the number of drugs and the number of diseases. The final aim is to reduce the number of drugs in order to avoid drug-related adverse events.

No conflict of interest.