This study employs a multimodal model trained on chest X-ray images and clinical data to create an osteoporosis screening program. Chest X-ray images and clinical data, including age, sex, BMI, blood composition, and chest X-ray descriptive information, were collected for patients as well as Dual-energy X-ray Absorptiometry (DXA) data. An image model and a clinical parameter model, trained on chest X-ray images and clinical data, respectively, were then integrated into a multimodal model. Overall, the model performed relatively well, requiring only a small amount of training data and providing quick processing capability. Area under the curve (AUC) values were reported as 0.975, a significant improvement over models developed solely on X-ray imaging. Despite the findings, the project still lacks generalizability and requires further training across different populations, especially since patients were selected from a limited age group and demographic profile. In addition, chest X-ray imaging inherently limits lumbar spine and humeral head imaging, especially since those two sites are typically analyzed for osteoporosis diagnosis.