Objective:To evaluate the association between the use of emergency medical services (EMS) and the severity of disease among patients admitted to the emergency room, to analyze the characteristics of the patients, and to build prediction model providing evidence-based use of EMS resources.Methods:The data of patients admitted to the Emergency Room of the First Affiliated Hospital of University of Science and Technology of China from January 2020 to July 2021 were extracted from the Chinese Emergency Triage Assessment and Treatment (CETAT) database. Patients were divided into the EMS use group (AB+ group) and self-seeing group (AB-group) according to whether they used EMS. The patients’ general condition, vital signs and laboratory tests results were recorded. The severity of patients’ condition was judged based on whether the patient was admitted to the department of critical medicine, specialized care unit, emergency operation and/or emergency percutaneous intervention. A 9-variable model that did not require laboratory inspection and 22-variable model that required laboratory inspection were established to correct the propensity score to analyze the correlation between the severity of disease and the EMS use. In the subgroup analysis, the correlation between the EMS use and severity of the patients was analyzed according to the reason of the patient’s visit.Results:During the study period, 16 489 patients were admitted to the emergency room, and 6975 patients were finally enrolled in this study. There were 2768 patients (39.7%) in the AB+ group and 4207 patients (60.3%) in the AB-group. In the AB+ group 522 patients (18.9%) were in high risk, and in the AB-group 563 patients (13.4%) were in high risk. Compared with the AB-group, patients in the AB+ group were older and had a higher proportion of coma, a faster autonomic heart rate, and a lower diastolic blood pressure and peripheral oxygen saturation (SpO 2). In the 9-variable model, sex, consciousness, temperature, heart rate and diastolic blood pressure were associated with the EMS use. In the 22-variable model, consciousness, SpO 2, neutrophils, and albumin were the relevant factors for patients using EMS. Before the correction of propensity score, the EMS use was an independent risk factor for critically ill patients ( OR=1.5, 95% CI 1.32-1.72, P<0.001). After adjusted using 9-variable propensity score, the EMS use ratio decreased significantly compared with that without correction ( OR=1.24,95% CI 1.08-1.42, P<0.001). Interestingly, after adjusted with propensity score match with 22-variable model, there was no association between the severity of disease and t the EMS use ( OR=1.10,95% CI 0.95-1.28, P=0.195). In subgroup analysis, patients’ chief complaint of central nervous system, cardiovascular system, and trauma were the top three reasons at admission. Before the propensity score correction, the EMS calling patients with chief complaint of central nervous system, digestive system, and trauma were related to the severity of the patients. After adjusted with 9-variable model the EMS use was associated with the severity of the disease only in trauma patients, and after adjusted with 22-variable model there was no statistical difference considering the severity of the disease in all subgroups. Conclusions:The EMS use is common. However, the association of the EMS use with the severity of disease is decreased with variable models using propensity score. These findings indicate that the EMS use should be based on multivariable models, which may be important in detecting critically ill patients, optimizing the EMS use, and avoiding unnecessary call in the future.