Trend in cervical cancer mortality in Chile: application of joinpoint regression models

Objective. To identify and model changes in the cervical cancer mortality trend in Chile between 1990 and 2009 in order to contribute to the design of health policies on cervical cancer. Methods. The study analyzed deaths from cervical cancer in Chile between 1990 and 2009. The cases were identified by the codes ICD-9 180 and ICD-10 C-53. The population at risk was calculated on the basis of projections by the National Statistics Institute of Chile. Mortality was modeled using Poisson regression analysis, including the variables of age in decades starting at age 30 and time interval in calendar periods and the term of interaction between the two variables. A joinpoint regression model, stratified by age group, was used to assess changes in the trend. In both models, the annual percentage of change (APC) was used as a summary measure. Results. During the period studied, the APC in Chile was -4%. Women aged 40 to 49 experienced the greatest reduction in mortality (incidence rate ratio = 0.98; P < 0.05) (reference group: women aged 30 to 39). Those with the least reduction in mortality were women over 80 (APC -1.18%). The joinpoint model identified six significant changes in the time trend for the women studied: the greatest reductions occurred in 1999-2002 and 2003-2009, with APCs of -6.9 and -4%, respectively. Conclusions. Mortality from cervical cancer exhibited a downward trend during the period studied, and the trend was more pronounced among women in the youngest age group.