Environmental forces have been associated with dengue fever occurrences in endemic areas

Environmental forces have been associated with dengue fever occurrences in endemic areas, as well as fecal indicator bacteria variability in recreational waters (Chowell and Sanchez 2006, Pednekar et al. 2005). These are important to model and understand to protect public health. Nevertheless, these interactions are complex and by just modelling them with linear models we might be missing important data (Chebud et al. 2012, He and He 2008). This research provides a better understanding of how environmental factors are related to dengue fever and culturable enterococci in a tropical setting, applying linear and nonlinear models with satellite-derived data and long-term epidemiological data.
Chapter Two showed that dengue incidence rates generally increased in July (wet season) and decreased in November (dry season) in Yucatan, Mexico. Changes in previous dengue fever cases explained the most variability and were positively correlated with current cases. Precipitation, minimum air temperature, humidity, and SST were selected as the best variables to explain dengue fever incidence. These results showed that increases in SST precede increased dengue incidence rates by eight weeks and that dengue incidence rates were positively correlated to SST changes. It is concluded, then, that dengue fever incidence rates can be modelled using environmental variables alone, and that by including satellite-derived regional-scale SST the modelling was improved. Nevertheless, it is important to note that even though seroprevalence studies are expensive, the inclusion of human immune background can allows to have more robust models.
Chapter Three showed that precipitation, mean sea level (MSL), direct normal irradiance (DNI), SST, and turbidity explained some of the observed variation. These parameters preceded changes in culturable enterococci concentrations with lags spanning from 24 h up to 11 days. The highest influence on culturable enterococci concentration was between 480 mm – 900 mm of 4-day cumulative precipitation. Higher culturable enterococci were observed during higher turbidity anomalies, warmer SST anomalies, and lower MSL anomalies. A significant decrease in culturable enterococci concentrations was observed during increased solar irradiance. Better monitoring of recreational water quality can be achieved by understanding the influence of environmental factors on culturable enterococci concentrations and how marine waters influence culturable enterococci decay rates (Anderson et al. 2005). It is concluded, then, that culturable enterococci concentration variability can be explained by looking at the combined effects of precipitation, SST, MSL, and turbidity.
In Chapter Four, a predictive model was applied to predict dengue fever outbreak occurrences in San Juan, PR and Yucatan, MX. These models were modified to predict dengue fever outbreak occurrences for the population at highest risk of infection (i.e., < 24 years old) and highest vulnerability of infection (i.e., 65 years old; Mendez-Lazaro et al. 2014). These groups were based on previous studies (Laureano-Rosario et al. 2017, Mendez-Lazaro et al. 2014) and data provided by the Department of Health of Mexico and Puerto Rico. Based on these predictions, the most influential variables to predict dengue fever outbreak occurrences in both Puerto Rico and Mexico were previous dengue incidence rates, minimum/maximum air temperatures, date, and population size. These models showed an accuracy of ~50%, with an overall power greater than 70%. Nonetheless, these results showed that the most influential variables to predict dengue fever occurrences are those related to demographics, followed by environmental factors such as temperatures (i.e., sea temperature, air temperature) for both Puerto Rico and Mexico. Therefore, it is concluded that, while demographic factors are important for prediction and mitigation, environmental factors should always be taken into account, and that these relationships are location-specific.
The predictive model was also applied in Chapter Five to predict culturable enterococci concentration exceedance at Escambron Beach surface waters. The model showed the following as the most influential factors: 48 h cumulative precipitation, turbidity anomalies, DNI, MSL anomalies, and SST anomalies. These predictions had an accuracy greater than 70%, higher than the predictive capability of only using a simple linear regression model. Thus, modelling culturable enterococci concentration exceedance at Escambron Beach was achieved by the predictive nonlinear model, where it identified the combined effects of these environmental factors influencing culturable enterococci concentrations.
The results of this dissertation can be integrated into future models to better understand the burden of water-related pathogens correlated with fecal indicators and vector-borne diseases in specific locations. The World Health Organization (WHO) estimates about 720,000 deaths per year related to 12 vector-borne diseases, where 80% of the world’s population is at risk and those younger than 5 years old are considered more susceptible (WHO 2018). Understanding the relationship and seasonality of these vectors, as shown in Chapter Two and Four with dengue fever, can help achieve better predictions and further develop disease surveillance and prevention strategies. In terms of water, sanitation, and hygiene (WASH), WHO reports about 840,000 deaths per year with 361,000 of those being children younger than 5 years old, and where 58% of these deaths could be averted through better sanitation practices (WHO 2018). While these statistics include both freshwater (i.e., drinking water) and marine waters, the results of this dissertation can help better understand patterns of specific indicators and how those are related to human activities and climate. Consequently, this dissertation supports and expands on efforts to understand diseases occurrence on specific population segments and seasonal variability of vector-borne diseases and water indicators related to poor recreational water quality.
This study demonstrated that the combined effects of environmental factors can improve our understanding of the ecology and epidemiology of diseases and microbial indicators over time, which would have been missed by just looking at just one environmental variable. Combining environmental and oceanographic variables improved modelling of dengue fever in Mexico and recreational water quality in Puerto Rico. Thus, this research contributes to the understanding of the influence of environmental factors on public health issues through the comparison of linear and nonlinear modelling as well as predictive models targeting specific population segments and geographic locations.