Development and internal validation of a vaginal microecology-based multivariable prediction model for persistent high-risk human papillomavirus infection: a retrospective study

1

A multivariable prediction model for persistent high-risk HPV infection was developed using vaginal microecological and clinical characteristics.
2

The study included 1,186 women with baseline HR-HPV positivity and 12-month follow-up testing at a tertiary hospital.
3

Seven predictors were identified: age, smoking status, HPV16/18 infection, vaginal pH, non-Lactobacillus microbiota, bacterial vaginosis, and inflammation.
4

The final model demonstrated good discrimination in the training cohort (AUC 0.892) and acceptable discrimination in the validation cohort (AUC 0.821).
5

The model showed potential clinical usefulness but requires external validation before broader implementation in clinical practice.

1