Author(s)

Jiarui Chang, Jiayi Chen, Xiaoqing Chen

Corresponding Author

Jiarui Chang

ABSTRACT

Cervical cancer is one of the most preventable cancers, and human papillomavirus (HPV) infection is one of the major factors that can cause it. A cost-effective model that helps low-income countries diagnose cervical cancer needs to be found as soon as possible. Data of 858 patients in 2017 was used to analyze possible behaviors that could cause HPV infection. The logistic regression model was built to find the most significant variables that can help estimate HPV infection. A confusion matrix was used to evaluate the reliability of the model. Logistic regression analysis showed that the final explanatory variables were age, number of sexual partners, number of pregnancies, and years of using hormonal contraceptives (p < 0.05). The optimized new model has an accuracy of 0.5877. Our study suggests that the behaviors commonly viewed as risk factors of HPV infection are insignificant when combined to analyze. The findings have played a great role in enlightening other studies on the infection of HPV and the pathogenic factors of cervical cancer and opened a new path.

KEYWORDS

Cervical cancer, HPV infection, Sexual behaviors