A mixed-methods approach for vector-borne disease surveillance in Colombia
Analytics
167 views ◎141 downloads ⇓
Abstract
Vector-borne diseases (VBDs) affect more than 1 billion people a year worldwide, cause over 1 million deaths, and result in hundreds of billions of dollars in societal costs. Dengue fever (DENF), chikungunya (CHIK), and Zika are three emerging VBDs that are transmitted by the Aedes mosquito. A combination of increased urbanization, globalization, climate change, and decreases in vector control have resulted in global increases in VBD epidemics, especially in previously unaffected regions. In Colombia, the co-circulation of DENF, CHIK, and Zika have resulted in severe epidemics where hundreds of thousands of people have been infected during the past decade. DENF has been endemic in Colombia for decades, and CHIK and Zika first appeared in 2013. It is critical to implement surveillance strategies that can improve the understanding of VBD transmission. Integrating mixed-method approaches in VBD surveillance are important because solely using quantitative approaches will not capture the experiences and behaviors of individuals who are susceptible to disease; and those responsible for policy-making and public health interventions. This dissertation combines spatial and space-time statistical models, surveys, and semi-structured interviews to understand the socioeconomic, environmental, political, and institutional factors that influence the transmission of DENF, CHIK, and Zika in Colombia – one national level study; and three studies in the city of Cali. First, I detect and visualize space-time clusters of both DENF and CHIK at the national level between 2015 and 2016; and compute relative risk for each municipality that belongs to a cluster. Second, space-time conditional autoregressive (ST-CAR) models are developed to identify significant predictors of DENF, CHIK, and Zika at the neighborhood level in Cali, Colombia; and the models also include meteorological variables that are temporally lagged to predict VBD outbreaks (early warning system). Third, I administer 327 Knowledge, Attitude, and Practice (KAP) surveys to individuals in healthcare centers and select neighborhoods in Cali, Colombia in June 2019. KAP surveys are used to shed light on at-risk communities’ understanding of the vector, the pathogen, prevention and treatment strategies. I utilize Generalized Linear Models (GLMs) to identify significant predictors of KAP regarding DENF, CHIK, and Zika. The findings suggest that knowledge is related to community characteristics, while attitudes and practices are more related to individual-level factors. Access to healthcare also forms significant predictor of residents participating in preventative practices. Finally, I conduct six semi-structured interviews with high-ranking public health officials about their experiences regarding DENF, CHIK, and Zika prevention, treatment, and surveillance. Overall, the results can be leveraged to inform public health officials and communities to motivate at-risk neighborhoods to take an active role in vector surveillance and control, and improving educational and surveillance resources in Cali, Colombia.