Unplanned readmissions may be avoided by accurate risk prediction and appropriate resources could be allocated to high risk patients. The Length of stay, Acuity of admission, Charlson comorbidity index, Emergency department visits in past six months (LACE) index was developed to predict hospital readmissions in Canada. In this study, we assessed the performance of the LACE index in a Singaporean cohort by identifying elderly patients at high risk of 30-day readmissions. We further investigated the use of additional risk factors in improving readmission prediction performance.Data were extracted from the hospital's electronic health records (EHR) for all elderly patients ≥ 65 years, with alive-discharge episodes from Singapore General Hospital in 2014. In addition to LACE, we also collected patients' data during the index admission, including demographics, medical history, laboratory results, and previous medical utilization.Among the 17,006 patients analyzed, 2051 or 12.1% of them were observed 30-day readmissions. The final predictive model was better than the LACE index in terms of discriminative ability; c-statistic of LACE index and final logistic regression model was 0.595 and 0.628, respectively.The LACE index had poor discriminative ability in identifying elderly patients at high risk of 30-day readmission, even if it was augmented with additional risk factors. Further studies should be conducted to discover additional factors that may enable more accurate and timely identification of patients at elevated risk of readmissions, so that necessary preventive actions can be taken.