Objectives To construct a standardized Socioeconomic Position (SEP) index by integrating multiple socioeconomic status (SES) dimension indicators and evaluate its predictive power for Pulmonary Tuberculosis (PTB) incidence across China. Methods This longitudinal study leveraged a total of 20 SES indicators and PTB annual incidence rates across 31 provinces of China from 2012 to 2023 from large-scale national databases. We applied principal component analysis (PCA) and generalized estimating equations (GEE) to identify potential SES indicators. Subsequently, those indicators were standardized to construct SEP Indexes and prediction models for PTB incidence. we used relative error (RE), root mean square error (RMSE), quasi-likelihood independence criterion (QIC), and residual diagnostics to evaluate the predictive accuracy, goodness-of-fit and residual distribution of models. Results Four GEE models were constructed based on seven SES indicators through PCA contribution weights and univariate screening. Of those, SEP-1 and SEP-2 models, demonstrated good predictive performance with seven significant indicators in univariate analysis and four significant indicators in multivariate analysis, respectively. Specifically, SEP-1 model achieved a strong balance between simplicity and predictive accuracy (RE = 0.463; RMSE = 28.63; QIC = 381.50); SEP-2 model showed slightly better model fit (QIC = 310.76) but higher prediction error (RE =0.619; RMSE = 30.73). Conclusions The SEP index offers a practical and interpretable approach for rapid PTB incidence prediction. The corresponding SEP-1 and SEP-2 model, balancing simplicity and accuracy, is well-suited for scalable public health applications. Public Health Implications This study proposes a novel method quantifying SES to construct a standardized SEP index, offering a promising approach for the rapid prediction of TB incidence, supporting region-specific resource allocation and disease control strategies in diverse socioeconomic settings.

Pulmonary Tuberculosis, Socioeconomic Position Index, Generalized Estimating Equations, Prediction