BACKGROUND: Although tuberculosis (TB) is an endemic infectious disease in Brazil, the Federal District has a different epidemiology from other Brazilian locations. Investigation of disease seasonality, which needs to be further explored in tropical regions, may indicate risk factors specific to peak incidence seasons that could be controlled if better understood. To investigate the influence of seasonality (SA) in the occurrence of tuberculosis in the Federal District (FD) of Brazil.
METHODS: This is an ecological time series (TS) study based on secondary data. The unit of analysis was the month of TB diagnosis from 2001 to 2018. We used the X-13ARIMA-SEATS (X-13) seasonal adjustment software in automatic mode with additive decomposition. Seasonality was analyzed using the Friedman test (0.001) and Kruskal-Wallis test (0.01) for stable seasonality. Moving seasonality was verified by Friedman test (0.05), and the identifiable seasonality was constructed from the combination of the three tests. The annual average seasonal amplitude was calculated from the isolated seasonal factors.
RESULTS: 6,161 TB cases were diagnosed in the Federal District, 28 cases per month on average (± 6.6). We observed peaks in April and August and decreased detection in November and December, with an average annual seasonal amplitude of 34.0%. The F test and the Kruskal-Wallis test revealed evidence of SA at 0.001 and 0.01, respectively. Moving SA was not identified at a significance level of 0.5 neither in the combination of the three tests.
CONCLUSION: the TS tends to suggest a limited seasonality of TB in the FD. Some incidence peaks were observed in some specific months, showing a relevance of seasonality at the local level, which may reflect an influence of access to health service patterns from the onset of symptoms to the diagnostic elucidation of the disease.