Análise espacial e modelos de predição da leishmaniose visceral humana no Estado do Maranhão

Abstract

Introduction: Human visceral leishmaniasis is an infectious disease, prevalent in tropical countries. Prediction of the occurrence of infectious diseases through epidemiological modeling, the use of geoprocessing and sophisticated statistical methods incorporated into the analysis of spatial data are health tools that aim to understand their occurrence dynamics, guiding the conducts regarding the control of this disease. Objective: Analyze the spatial dynamics and develop a prediction model for the occurrence of LVH for the state of Maranhão, 2001 to 2018. Methods: Monthly data on LVH cases were collected through the Notifiable Disease Information System corresponding to the period 2001 to 2018. For the prediction model, the Box-Jenkins method was applied to adjust a SARIMA prediction model for general incidence and by sex (male and female) of LVH for the period from January 2019 to December 2023. For the analysis of the spatial pattern, the Moran Global Index and the Local Indicators of Spatial Association (LISA) were calculated. Results: During the 216-month period of this time series, 10,431 cases of VHL were registered in Maranhão, with an average of 579 cases per year. In relation to age group, there was a greater number of records in the pediatric population (0 to 14 years old). There was a predominance of males, with 6437 cases (61.7%). The Box-Pierce test values for general incidence, male and female, reinforced by the results of the Ljung-Box test suggest that the autocorrelations of residues present a white noise behavior. For general monthly incidence and by male and female, the SARIMA models (2.0.0) (2.0.0), (0.1.1) (0.1.1) and (0.1.1) (2, 0, 0) were the ones that best fit the data, respectively. The behavior of the time series, according to the SARIMA model, in general, for the total incidence, a decreasing trend was observed. However, in women, there was a trend towards an increase in incidence for the forecast period. The global spatial autocorrelation analyzes showed that the Moran Global Index of LVH incidence in Maranhão varied significantly, indicating the presence of spatial clusters during the study period. Univariate LISA analysis identified clusters of transmission of LVH predominant in the east and west portions of the state. LVH in Maranhão throughout this historical series had an important spread of its occurrence, with the emergence of new clusters of cases. In the period considered, the disease was registered in 206 of the 217 municipalities. Conclusion: The SARIMA model and the Global Moran Index and the local Moran index proved to be adequate tools for forecasting and trending the incidence of LVH in Maranhão and analyzing the spatial dynamics, revealing that the disease will persist as a public health problem in the coming years, reinforcing the need for prevention and control measures. The determination of temporal variation and its prediction are crucial in guiding health intervention measures.