OBJECTIVES: More than half of medication errors are reported during administration. “Smart” Infusion Safety Systems (SISS) have been developed to avoid administration errors and provide data for continuous quality improvement (CQI). Few data produced by SISS have been analysed. Given the lack of clinical trials, a model was developed in order to assess the economic benefits of SISS from the perspective of a French public hospital. METHODS: Comprehensive data from 6 intensive care units (ICU) were analysed. Two types of alerts were defined: hard (absolute) and soft limits. An avoided error was defined as the detection of a scheduled infusion over a hard limit or a scheduled infusion over a soft limit that was later overridden by staff. The severity of consequences was estimated on the HARM INDEX score, which is based on the pharmacologic risk, overdose detectability, overdosing range and the type of hospital unit. According to this score, errors were categorised as minor, significant, and serious. The economic value was estimated based on the hypothetical resulting length of stay had the error not been avoided, which increases according to error severity, and the financial investment for SISS. RESULTS: Overall, 207,025 infusions were analysed from the database of 6 ICUs and 8,503 of them were associated with a safety alert (4.1%), including 987 errors (0.48%). Applied to one public ICU equipped with 5 SISS, the model estimates that SISS would prevent 78 errors per year (54 minor, 12 significant, 12 serious) and avoid 173 days of hospitalisation per year. This corresponds to an annual savings of €139,491 for the hospital (39% and 61% from significant and serious errors, respectively). CONCLUSIONS: This model demonstrates the high economic burden of medication errors for French public hospitals and the need for CQI. To date, no equivalent evaluation has been conducted in France.