Background Medication reconciliation (MR) makes it possible to identify medication errors. Because it is labour-intensive, it is often limited to certain specific hospital units (HU).

Purpose The goal of this study was to evaluate a MR activity targeting patients at the highest risk admitted to the emergency unit.

Material and methods A single-centre prospective study was performed for 6 months in patients hospitalised through the emergency unit. Emergency unit physicians or nurses would fill in a prioritisation grid of MR including 10 clinical and therapeutic factors. This grid, which was based on a bibliographic study and a prior internal study, included a box « don’t know » (DK) for every factor. A pharmacist collected the grids daily and calculated the risk score of each patient: in the case of a score ≥10, a pharmacist performed a MR of the patient in the unit where s/he was hospitalised.

Results A prioritisation grid was filled out for 583 patients. Ten and 36% of the grids included at least one DK box checked by the physicians and the nurses, respectively. Twenty-four per cent of the patients were eligible for MR according to the physicians, 11% according to the nurses, for a total of 130 patients. Fifty-six MR were performed in 15 different HU, which represented 43% of the identified patients, with an average of 1 hour per MR of the pharmacist’s time. The number of unintended medication discrepancies (UMD) was 1.2/patient.

Conclusion This grid seems to be adapted to the prioritisation of MR, because 24% and 11% of the patients had a score ≥10. It identified the need for MR in a large number of HU, which is the originality of our MR activity. All the priority MR could not be performed because of early release/death of patients or lack of time. The low rate of patients at risk and the high rate of DK checked by nurses suggests that nurses under-evaluate this risk. Physicians seem to have a better understanding of the patients and treatment. The MR of patients at risk made it possible to identify a number of UMD similar to that found in other French studies.

No conflict of interest