Análise espaço-temporal das internações hospitalares por câncer de boca no Brasil e sua correlação com a expansão da rede de atenção à saúde

Abstract

Introduction: Oral cancer (BC) is a chronic degenerative disease with significant incidence, morbidity and mortality. Its high incidence has been reflected in the increase in the number of outpatient treatments, rates of hospital admissions (IH) and the public resources demanded to pay for the treatments. In Brazil, the Health Care Network (RAS) is responsible for comprehensive care for cancer patients. However, there is little evidence to assess whether the expansion of RAS has any effect on IH by CB that have evolved to death in the country. The aim of this study was to analyze IH and deaths from BC in Brazil and its correlation with the expansion of RAS. Method: Ecological, longitudinal and analytical study of IH by CB registered in the Hospital Information System available at DATASUS / MS between 2007 and 2019. The outcome of interest was the proportion of IH by CB that evolved to death. For the analysis of the RAS, the following variables were considered: doctor per inhabitant, dentist (CD) per inhabitant, coverage of the family health strategy (ESF) coverage of ESB in the ESF, number of Dental Specialty Centers (CEO), number of beds oncological. Descriptive, spatial analyzes and negative binomial regression were performed, estimating crude and adjusted regression coefficients and respective 95% confidence intervals (95% CI), considering  = 5%. Results: There were 262,728 IH by CB in Brazil between 2007 to 2019. The number of IH who died was 10.92%. The average hospital stay was 6.3 days. Tongue cancer was the most frequent anatomical site. All Federative Units (UF) had higher FHS coverage in 2018 than in 2007. In 2007, the Moran index (I) revealed a negative spatial correlation between the proportion of hospital deaths by BC and the FHS coverage (I = - 0.234), ESB (I = -0.228) and CEO (I = -0.100); and a positive spatial correlation with the ratio of CD / inhabitant (I = 0.241), doctors / inhabitant (I = 0.365) and of cancer beds / inhabitant (I = 0.238). In 2019, the spatial autocorrelation indexes were inversely proportional for all the predictors used. In the adjusted analysis, the proportion of HI with deaths from CB was higher in UF in the North of the country; and it was inversely associated with the ratio of doctors ̸habitant (β = -0.014; p = 0.040), DC / inhabitant (β = -0.720; p = 0.045), number of CEO (β = -0.004; p <0.000), with the per capita amount paid by IH (β = -10.350; p <0.000) and the number of biopsies performed (β <0.001; p = 0.010), number; and positively associated with the average number of days of IH (β <0.001; p = 0.002). Conclusion: The increase in the coverage of the RAS decreased the severe IH by CB in Brazil.