Análise exploratória de dados de mortalidade por câncer colorretal no Estado de Mato Grosso e no Brasil

Abstract

Cancer is considered a public health problem that is the main cause of death in the world, it is recognized by the pressure on the body, it is financed by individuals, families, communities and health systems. Colorectal Cancer (CRC) is the 4th most common type in the world and the 5th most lethal; in Brazil, in both sexes, CRC was the 2nd most common. Objective: To identify the trend in mortality due to CRC in Mato Grosso (MT), use the Location Quotient (LQ) as a strategy to Spatial analysis mortality rates and validate the relationship between the LQ and the aging index (AI). Methods: This is an ecological study with data from the Sistema de Informação de Mortalidade (SIM) of cases of death whose basic cause is: colon cancer (C18), cancer of the rectosigmoid junction (C19) and cancer of the recurrence (C20) between 2000 and 2019 in the state of Mato Grosso. It was used exploratory data analysis (EDA) to detect trends and growing hypotheses, the AI to support trends in demographic dynamics, descriptive statistics, and the LQ. We used the joinpoint regression model. We calculated the Moran’ statistics for spatial correlation. Spearman and Pearson correlation tests were used to analyse the correlation between AI and LQ in our research. Results: There were 2,327 cases in total, a large part of the population is between 50 and 75 years old, mainly males (52%). The results of the joinpoint analysis do not present inflexibility points for the analysed trends. The MT microregions have very different AI and the LQ varies greatly, with the highest concentration being in Cuiabá (LQ=1.53). The Moran’ statistics has a current estimate of 0.08. In MT, the Spearman correlation was 0.19 (p-value = 0.39) and the Pearson correlation was 0.15 (p-value = 0.52). Conclusions: The LQ is a useful and simple strategy for spatial analysis of mortality cases. Although there is a significant correlation in the Brazilian data between the QL and the IE, as was intuitively expected, such a relationship does not appear in the MT data, perhaps due to microregions with a small sample size.