Background: Overutilization of emergency room services by oncology patients is a known problem associated with increased admission rates and health care expenditure. A review of our oncology patients’ emergency room (ER) visits from January to May 2015 demonstrated that 48% of ER visits happened during office hours. Consequently a rapid cycle quality improvement project was developed with an aim to decrease ER visits by 30% by September 2015. Methods: A multidisciplinary team completed an action plan, starting with a project charter and definition of aim statement. A process map for patient scheduling/triage was created. A cause and effect diagram helped identify potential causes patient utilization of the ER. Diagnostic data were obtained querying our EMR (EPIC) for ER visits from January to May 2015. A Pareto chart identified Breast, Hematology and GI malignancies as main diagnosis utilizing the ER. Plan-do-study-act (PDSA) #1 began with development of a protocol to guide the handling of patients' calls that could previously resulted in an ER visit. Staff from the patient access center (PAC), a telephone operator service, and physicians’ offices were trained on its application. PDSA #2 focused on patient education to the importance of contacting the PAC for any concern or symptom related to active chemotherapy treatment. Results: The implementation of a triage system at our PAC resulted in a 60% decrease in the number of patients utilizing the ER, which met our goal. Patients’ calls to the PAC have increased. Two new materials were developed: a telephone triage form categorizing the patient’s complaint and the resulting action by our PAC center staff, and a patient Clinical Intervention Triage Tracking Log which allowed for the tracking of all patients triaged, their data, and the responsible team member. Conclusions: This study suggests that the development of a tool to properly identify and address emergent chemotherapy symptoms without utilizing the ER during working hours resulted in an intervention that positively affected the pre-specified endpoint.