Modelos para previsão das taxas de incidência e mortalidade do câncer do colo do útero

Abstract

Objective: analyze and predict cervical cancer incidence and mortality rates, using time series models. Method: ecological study, carried out in the capital Matogrossense. The incidence data were from the Population Based Cancer Registry, from 2000 to 2012, and mortality data, from 1981 to 2018, obtained from the Mortality Information System. And analyzed by the ARIMA and HOLT models. Results: the series of the incidence rate of carcinoma in situ, the appropriate model was the AR (1). Estimated projections (2013-2014) fluctuated between 15,54 and 16,45 per hundred a thousand women. In the invasive cervixl series, the model was Holt’s, and the predictions (2013-2014), with values between 11,28 to 8,9 per 100 thousand women. In the mortality rate, the satisfactory model was ARMA (1,1), whose predictions (2019-2020) fluctuated between 9,46 and 9,62 per 100 thousand women. Conclusion: in the three cases, the predicted rates presented levels close to those observed, indicating that the models allowed estimating and predicting, in an adequate way. It was also possible to observe a slight increase in in situ cancer and a decline in the incidence of invasive cancer, while mortality showed stability in the results and predictions. Descriptors: ARIMA; Cervical Cancer; Incidence; Mortality; Holt's method.