Impact of sociodemographic factors and screening, diagnosis, and treatment strategies on colorectal cancer mortality in Brazil: A 20-year ecological study

Abstract

Colorectal cancer (CRC) caused 261,060 deaths in Brazil over a 20-year period, with a tendency to increase over time. This study aimed to verify the sociodemographic factors predicting higher mortality caused by CRC and survival rates. Moreover, we aimed to verify whether the performance of screening, diagnostic and treatment procedures had an impact on mortality. Ecological observational study of mortality due to CRC was conducted in Brazil from 2000-2019. The adjustment variable was age, which was used to calculate the age-standardized mortality rate (ASMR). The exposure variables were number of deaths and ASMR. Outcome variables were age-period-cohort, race classification, marital status, geographic region, and screening, diagnostic, and treatment procedures. Age-period-cohort analysis was performed. ANOVA and Kruskal-Wallis test with post hoc tests were used to assess differences in race classification, marital status, and geographic region. Multinomial logistic regression was used to test for interaction among sociodemographic factors. Survival analysis included Kaplan-Meier plot and Cox regression analysis were performed. Multivariate linear regression was used to test prediction using screening, diagnosis, and treatment procedures. In Brazil, mortality from CRC increased after age 45 years. The highest adjusted mortality rates were found among white individuals and in the South of the country (p < 0.05). Single, married, and widowed northern and northeastern persons had a higher risk of death than legally separated southern persons (p < 0.05). Lower survival rates were observed in brown and legally separated individuals and residents from the North (p < 0.05). An increase in first-line chemotherapy and a decrease in second-line chemotherapy were associated with high mortality in the north (p<0.05). In the south, second-line chemotherapy and abdominoperineal rectal resection were associated with high mortality (p < 0.05). Regional differences in sociodemographic factors and clinical procedures can serve as guidelines for adjusting public health policies.

Fig 1

Age-period-cohort analysis using the Wald test (A), with the analysis of all age deviations (A) and the mortality rates (B) by age. In Fig 1A and 1B, the lines represent the oscillation of the Y-axis value in relation to the X-axis; the circle represents the moment on the X-axis; the shadows above the lines represent the 95% CIs.

Fig 2

Age-period-cohort analysis using the Wald test, with the analysis of net drift (A), temporal trends (B), period rate ratios (C), the cross-sectional age curve (D), the age incidence pattern for every period (E), and all period deviations (F). In Fig 2B-2D and 2F the lines represent the oscillation of the Y-axis value in relation to the X-axis; the circle or square represents the moment on the X-axis; the shadows above the lines represent the 95% CIs.

Fig 3

Age-period-cohort analysis using the Wald test (A), with the cohort RR analysis (B), the longitudinal age curve (C), cohort deviations (D), and age incidence patterns for every cohort (E). In Fig 3B-3D, the lines represent the oscillation of the Y-axis value in relation to the X-axis; the circle represents the moment on the X-axis; the shadows above the lines represent the 95% CIs.

Fig 4

Colorectal cancer mortality rates according to geographic region (A), race (B), and marital status (C). ^{a, b, c} p < 0.05, post-hoc Bonferroni. *p < 0.05, post-hoc Bonferroni in comparison with White. #p < 0.05, post-hoc Bonferroni in comparison Black.°p < 0.05, post-hoc Bonferroni in comparison with Yellow.

Fig 5

Survival curves among the different racial classification groups (A), marital status (B) and geographic regions (C) using Kaplan-Meier plots.