Spatio-temporal correlation networks of dengue in the state of Bahia

• Published in: BMC public health Vol. 14; no. 1; p. 1085
• Main Authors: Saba, Hugo, Vale, Vera C, Moret, Marcelo A, Miranda, José Garcia V
• Format: Journal Article
• Language: English
• Published: England BioMed Central Ltd 18.10.2014; BioMed Central
• Subjects: Brain research; Brazil - epidemiology; Complexity; Dengue - epidemiology; Dengue - prevention & control; Dengue fever; Disease; Disease prevention; Disease transmission; Epidemiology; Female; Health Promotion - methods; Health Promotion - statistics & numerical data; Health Services Needs and Demand - statistics & numerical data; Humans; Incidence; Male; Municipalities; Public health; Public Policy; Randomization; Seasons; Socioeconomic Factors; Space-Time Clustering; Time series; Urban Population - statistics & numerical data; Vector-borne diseases; Viral diseases; Brazil
• ISSN: 1471-2458; 1471-2458
• DOI: 10.1186/1471-2458-14-1085

Dengue is a public health problem that presents complexity in its dissemination. The physical means of spreading and the dynamics of the spread between municipalities need to be analyzed to guide effective public policies to combat this problem. This study uses timing varying graph methods (TVG) to construct a correlation network between occurrences of reported cases of dengue between cities in the state of Bahia-Brazil. The topological network indices of all cities were correlated with dengue incidence using Spearman correlation. A randomization test was used to estimate the significance value of the correlation. The correlation network presented a complex behavior with a heavy-tail distribution of the network edges weight. The randomization test exhibit a significant correlation (P < 0.0001) between the degree of each municipality in the network and the incidence of dengue in each municipality. The hypothesis of the existence of a correlation between the occurrences of reported cases of dengue between different municipalities in the state of Bahia was validated. The significant correlation between the node degree and incidence, indicates that municipalities with high incidence are also responsible for the spread of the disease in the state. The method proposed suggests a new tool in epidemiological control strategy.