Problem: A medicine/oncology unit in a 250-bed community hospital identified an increase in rapid response calls over a 90-day period as a problem. During this time, 20% of calls came from a medicine/oncology unit and contributed to 20% of unplanned ICU transfers. Failure to detect early clinical deterioration results in poor patient outcomes. Purpose: The purpose of this quality improvement (QI) project was to reduce unplanned ICU transfers among adult inpatients in a medicine/oncology unit by implementing and measuring adherence to the Modified Early Warning System (MEWS) algorithm. Methods: The MEWS algorithm uses a color-coded system with corresponding numerical scores to indicate the risk of patient deterioration. Nurses assessed the MEWS score every eight hours, documented provider notification, and appropriate interventions performed in the Electronic Health Record (EHR). Flyers with the color-coded algorithm were placed on computers for easy reference. This project was expected to benefit 450 hospitalized adult patients in the medical/oncology unit. Results: Over 15 weeks, 140 yellow and 10 red MEWS scores were recorded. Of these patients, 33% received appropriate interventions. Thirty-four rapid responses and 34 ICU transfers occurred. Conclusions: Findings correlate with literature that the MEWS tool increases rapid responses, however a decrease in ICU transfers was not observed. The low uptake of the tool indicates that the tool alone may not be sufficient to reduce ICU transfers. Keywords: Modified Early Warning System, ICU transfers, rapid responses, clinical deterioration