OBJECTIVES: Human papillomavirus (HPV) has been recognized as a major factor for cervical cancer causation. Other factors, relating to reproduction, are also important. This study aims to disentangle the roles of baseline HPV infection, high vaginal parity (defined as having ≥4 vaginal deliveries), and the interaction between the two in predicting cervical cancer risk. METHODS: The authors apply a newly developed causal-pie modeling technique to analyze a cohort of more than 10,000 women conducted in Taiwan with more than 10 years of follow-up. The rate ratios adjusted by age and menopausal status were further modeled by an additive Poisson regression with non-negative parameters. The index of causal-pie weight (CPW) was calculated to indicate the proportion of cervical cancer cases attributable to a particular class of causal pies. RESULTS: It was found that the CPWs are 36.3 % for baseline HPV infection, 35.6 % for baseline HPV infection and high vaginal parity, and 28.1 % for other factors. CONCLUSIONS: A causal-pie modeling based on a women cohort in Taiwan successfully disentangles the roles of virus factors and reproductive factors at study entry, independently or interactively, on subsequent cervical cancer risk.
OBJECTIVES:Human papillomavirus (HPV) has been recognized as a major factor for cervical cancer causation. Other factors, relating to reproduction, are also important. This study aims to disentangle the roles of baseline HPV infection, high vaginal parity (defined as having ≥4 vaginal deliveries), and the interaction between the two in predicting cervical cancer risk. METHODS: The authors apply a newly developed causal-pie modeling technique to analyze a cohort of more than 10,000 women conducted in Taiwan with more than 10 years of follow-up. The rate ratios adjusted by age and menopausal status were further modeled by an additive Poisson regression with non-negative parameters. The index of causal-pie weight (CPW) was calculated to indicate the proportion of cervical cancer cases attributable to a particular class of causal pies. RESULTS: It was found that the CPWs are 36.3 % for baseline HPV infection, 35.6 % for baseline HPV infection and high vaginal parity, and 28.1 % for other factors. CONCLUSIONS: A causal-pie modeling based on a women cohort in Taiwan successfully disentangles the roles of virus factors and reproductive factors at study entry, independently or interactively, on subsequent cervical cancer risk.