Date of Award

January 2026

Document Type

Open Access Thesis

Degree Name

Master of Public Health (MPH)

Department

School of Public Health

First Advisor

Nathan Grubaugh

Abstract

Dengue fever poses a growing public health burden across the Caribbean, yet surveillance infrastructure remains inconsistent and timely outbreak detection in data-poor nations becomes limited. This study investigates whether shared meteorological patterns between Caribbean countries can be leveraged to nowcast dengue activity in low-surveillance settings using climate-matched, well-surveilled nations as epidemiological proxies. Weekly and monthly dengue case data for six countries, the Dominican Republic, Jamaica, Puerto Rico, Trinidad and Tobago, Cuba, and Haiti, were extracted from the OpenDengue database for 2005–2024, and satellite-derived climate data were obtained from the NASA POWER API. Pairwise climate correlations revealed strong temperature synchrony across the northern Caribbean cluster (r = 0.89–0.96), with precipitation showing weaker and more heterogeneous alignment. Cross-correlation and Hilbert transform phase analyses identified the Dominican Republic and Haiti emerged as the strongest cross-surveillance-quality pair, supported by their shared Hispaniola geography, the highest precipitation correlation in the dataset (r = 0.75), and sustained near-zero phase coherence during Haiti's pre-earthquake surveillance period. Among pairs involving Cuba, the Dominican Republic demonstrated the most structured phase relationship, characterized by a slow coherent drift rather than the erratic oscillations seen in Jamaica-Cuba and Puerto Rico-Cuba comparisons, providing independent synchrony-based support for its use as a Cuba proxy. In employing logistic regression for Cuba that incorporated lagged proxy case counts and local climate variables, Puerto Rico emerged as the strongest proxy for Cuba, achieving a perfect AUC of 1.0 and 66.7% sensitivity at a prevalence-calibrated threshold. The Dominican Republic performed comparably (AUC = 0.979), and Jamaica was the weakest proxy, consistent with the inverse phase relationship identified in synchrony analyses. These findings provide proof-of-concept evidence that dengue outbreak probability in data-poor Caribbean nations can be nowcast using epidemiological signals from climatically matched neighbors, with proxy selection guided by synchrony analysis rather than geographic proximity alone.

Comments

This is an Open Access Thesis.

Download
Open Access

This Article is Open Access

Share

COinS