Objective: This study assessed the therapeutic benefits and modeled the cost-effectiveness of oral nutritional supplements (ONS) in China. Methods: Data were collected from 27 152 adult inpatients between January 1, 2018, and December 31, 2020. Propensity score matching was used for balancing the baseline characteristics between the ONS group and non-ONS group. A decision-tree model was developed to assess the cost-effectiveness of ONS for patients with nutritional risk, and the incremental cost-effectiveness ratio was the metric to determine the most cost-effective strategy. One-way sensitivity and probabilistic sensitivity analyses were conducted to assess the model's stability. In addition, subgroup analysis was conducted based on clinical characteristics. Differences in clinical outcomes between the groups were compared using Student's t test, Mann-Whitney U test, or chi-square test. Results: The ONS group displayed significantly lower levels of prealbumin, albumin, hemoglobin, and BMI than the non-ONS group at admission. The incidence of malignant tumors, intestinal obstruction, and inflammatory bowel disease was significantly higher in the ONS group than the non-ONS group. The ONS group had a significantly higher effective rate than the non-ONS group (51.7% versus 50.3%, P < 0.05). Analysis of the decision-tree model revealed that the ONS group experienced an increase in cost of 19 850.96 yuan but achieved an additional 1.3406 effectiveness rate, resulting in an incremental cost-effectiveness ratio of 14 807.51, which fell below China's 2020 per capita gross domestic product of 71 965 yuan. Sensitivity analysis further confirmed the robustness of the model. Conclusions: ONS are demonstrated a high rate of efficacy, although patients currently using ONS are typically in a severe disease state. In addition, ONS is cost-effective. We suggest that the reimbursement coverage of ONS be expanded to include in-hospital patients who are at high nutritional risk. (c) 2024 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license