We present a case study concerning the use of electronic medical records (EMRs) acquired in an intensive care Unit (ICU). In particular, we focus on the prob- lem of exploiting this emerging new type of data for predicting Acute Kidney Injury (AKI), a frequent complication in hospitalized patients during patient stay using data collected in the Pediatric Cardiac Intensive Care Unit of Bambino Gesu` Childen’s Hospital. We discuss the methodological issues related to pre-processing the available EMR data, analyze the possible alternative ways of defining the outcome and use different tools for making predictions.