A predictive model between dietary vitamin intake and glaucoma was established using the nhanes database

Jiankang Liu^*bingli Wangqiuming li^*

• First Affiliated Hospital of Zhengzhou University, Zhengzhou, China

Objective: To establish and validate a predictive model of the relationship between vitamin intake in daily diet and glaucoma using data from the NHANES database. Methods: Data from the NHANES database for the 2005-2006 and 2007-2008 cycles were collected, and data with missing key information were excluded. For the filtered data from the 2005-2006 cycle, eight different modeling methods were used, and cross-validation grid search was adopted to obtain the optimal model parameters. The optimal model parameters of the different models obtained were validated and compared in the 2007-2008 cycle to determine the predictive model with the best performance among the eight models. Results: After cross-validation grid search calculations, it was found that the elastic net modeling method yielded the best model performance in the validation set (AUC = 0.69). In this model, it was discovered that age and being of Non-Hispanic Black ethnicity were associated with a higher risk of glaucoma, while folic acid intake in daily diet was a protective factor for the occurrence of glaucoma. Conclusion: This study established a predictive model of the relationship between nutrient intake and the occurrence of glaucoma and validated the model, concluding that folic acid intake is a protective factor for the development of glaucoma.