SDG 3: Good Health and Well Being

Brief Description:

The Philippines continues to grapple with recurring dengue outbreaks, underscoring the urgent need for timely detection and effective intervention strategies. This study focuses on modeling dengue incidence using advanced Bayesian integer-valued time series models, specifically the Generalized Poisson (GP) and Negative Binomial (NB) transfer function models. These models are designed to address key challenges in epidemiological data, such as overdispersion and the presence of intervention effects, including sudden spikes or gradual shifts in case counts. Utilizing weekly dengue case data from Cagayan de Oro City, the study aims to identify significant outbreak periods and assess potential environmental and public health influences. Through Bayesian inference and model diagnostics, the research demonstrates how these models can be effectively used for outbreak detection and short-term forecasting, providing a valuable tool for improving dengue surveillance and response systems.

Significance/Impact of the Study

This study is significant as it introduces a robust and data-driven approach to forecasting and understanding dengue outbreaks through the application of Bayesian integer-valued time series models, which effectively account for overdispersion and intervention effects commonly observed in disease incidence data. By detecting critical outbreak points and assessing the impact of environmental and public health factors, the study offers valuable insights for proactive epidemic response and planning. Its findings are instrumental for health authorities, particularly the Department of Health, in enhancing dengue surveillance systems, allocating resources efficiently, and implementing timely interventions. Moreover, the study contributes to the broader field of epidemiological modeling by demonstrating the practical utility of advanced statistical methods in addressing real-world public health challenges in dengue-prone regions.

Figure 1: Prediction plot and 95% credible interval based on the Negative Binomial Transfer Function Model in Cagayan de Oro City dengue cases. The black broken line represents the observed cases, while the blue line represents the predicted value of dengue cases in Cagayan de Oro City. The shaded purple region indicates the 95% credible interval of the predicted cases, and the red broken line indicates the intervention detected.

As shown in Figure 1, an intervention was detected in the early weeks of January 2016, characterized by a sudden surge in dengue cases in Cagayan de Oro City. A plausible explanation for this abrupt increase is the significant weather disturbances that occurred in December 2015, brought about by two tropical systems: Typhoon Melor (locally named Nona) and Tropical Depression Onyok. Although Cagayan de Oro was not directly in Melor’s path, its expansive rainbands caused widespread rainfall across Mindanao. This was followed by the landfall of Tropical Depression Onyok over Davao Oriental on December 18, further intensifying precipitation in the region. The combined effects of these weather systems resulted in heavy rainfall and localized flooding in Northern Mindanao, including Cagayan de Oro, creating ideal conditions for mosquito breeding and potentially contributing to the observed spike in dengue incidence.

Furthermore, beginning in 2016, an intervention effect marked by a sharp increase in dengue cases is consistently observed every July in Cagayan de Oro City. This pattern aligns with the peak of the rainy season in the Philippines, during which increased rainfall leads to the accumulation of stagnant water in containers, drains, and natural reservoirs—ideal breeding sites for Aedes mosquitoes, the primary vectors of dengue. Research has shown that precipitation is a key driver of dengue transmission, as mosquito populations thrive in water-rich environments. In addition to rainfall, other contributing factors include favorable temperature and humidity conditions, rapid urbanization, poor waste and water management, localized flooding, and broader climate variability. Together, these elements create a highly conducive environment for mosquito proliferation and, consequently, a higher risk of dengue outbreaks. Importantly, the detection of intervention points through the application of transfer function models serves as a valuable early warning mechanism, enabling health authorities, particularly the Department of Health, to initiate timely and targeted response measures to mitigate the spread of dengue.