Development and validation of a logistic regression-based prediction model and nomogram for vitamin D deficiency: a large-sample study from the high-altitude region of Ningxia, Northwest China

Vitamin D deficiency is a global health concern, but data from high-altitude regions in Northwest China remain limited. This study aimed to analyze the vitamin D status in Haiyuan County, Ningxia Province, and develop a corresponding prediction model. This retrospective study included 9,844 participants, randomly divided into training (70%, n = 6,891) and internal validation (30%, n = 2,953) sets. A multivariable logistic regression model was developed using age, sex, and season as predictors. Model performance was assessed using the area under the curve (AUC), calibration plots, Brier score, and decision curve analysis (DCA). A nomogram was constructed for clinical application. Subgroup analyses and interaction tests were conducted to evaluate model robustness and potential effect modification. The overall prevalence of vitamin D deficiency (< 20 ng/mL) was 70.3%. Female sex, older age, and off-summer seasons were identified as independent risk factors (all OR > 1, P < 0.0001). The model showed moderate discrimination (AUC = 0.715 in the training set), good calibration, and positive net benefit on DCA. These findings were confirmed in the validation set. Subgroup analyses revealed varying model performance across age, sex, and season strata (AUC range: 0.619–0.746). A significant interaction was observed between season and sex (P < 0.001), indicating that the effect of season on vitamin D deficiency risk differs by sex. Vitamin D deficiency is highly prevalent in the Haiyuan area. The developed prediction model and nomogram offer practical, low-cost tools for initial vitamin D deficiency risk stratification in high-altitude regions of Northwest China, supporting targeted preliminary screening in resource-limited primary care settings.