Title: Identification of the Most Influential Risk Factors Associated with Cervical Cancer among Women in Kenya
Authors: JF Murere, S. Wangila, J. Koech
Volume: 10
Issue: 8
Pages: 229-234
Publication Date: 2026/08/28
Abstract:
Cervical cancer remains one of the leading causes of cancer related deaths among women in Kenya as it accounts for approximately 3,200 deaths recorded yearly. This is largely attributed to low screening uptake (16%) and late diagnosis. Identifying the factors that are related with cervical cancer is important for guiding targeted screening. This is particularly important in resource constrained region like Western Kenya. A cross-section study design was adopted by this study to establish the most influential risk factors that are associated with cervical cancer screening among the 968 women who were involved the study from health facilities in Western Kenya. Five machine learning models: Logistic Regression, Random Forest, Decision Tree, Support Vector Machine (SVM), and Artificial Neural Network (ANN) were used for the analysis of feature importance. Pap smear results and hormonal contraceptive use emerged as the two most important predictors of cervical cancer. Hormonal contraceptive ranked highly in four models while Pap smear ranked top three for four out of the five models that were fitted. Notable predictors that were also flagged as contributors to cervical cancer included HIV status, age, age at first sexual intercourse, and number of sexual partners. This finding showed that there was at least one influential risk factor for cervical cancer in this population. It also underscored the importance of integrating clinical and demographic information into machine learning based cervical cancer disease screening approaches. The study showed that targeted intervention that focus on women with abnormal pap smear results, HIV positive women and those using hormonal contraceptives could significantly improve early cervical cancer detection and also reduce mortality.