Malaria intensity in Colombia by regions and populations

Abstract

Determining the distribution of disease prevalence among heterogeneous populations at the national scale is fundamental for epidemiology and public health. Here, we use a combination of methods (spatial scan statistic, topological data analysis and epidemic profile) to study measurable differences in malaria intensity by regions and populations of Colombia. This study explores three main questions: What are the regions of Colombia where malaria is epidemic? What are the regions and populations in Colombia where malaria is endemic? What associations exist between epidemic outbreaks between regions in Colombia? Plasmodium falciparum is most prevalent in the Pacific Coast, some regions of the Amazon Basin, and some regions of the Magdalena Basin. Plasmodium vivax is the most prevalent parasite in Colombia, particularly in the Northern Amazon Basin, the Caribbean, and municipalities of Sucre, Antioquia and Cordoba. We find an acute peak of malarial infection at 25 years of age. Indigenous and Afrocolombian populations experience endemic malaria (with household transmission). We find that Plasmodium vivax decreased in the most important hotspots, often with moderate urbanization rate, and was re-introduced to locations with moderate but sustained deforestation. Infection by Plasmodium falciparum, on the other hand, steadily increased in incidence in locations where it was introduced in the 2009-2010 generalized epidemic. Our findings suggest that Colombia is entering an unstable transmission state, where rapid decreases in one location of the country are interconnected with rapid increases in other parts of the country.

Fig 1

We start with a given data set (image A), for this example the points correspond to a sample of the unitary circle with a small amount of noise. For convenience we will use the euclidean distance to calculate the distance between each pair of points. In the next step, we select the projection onto the Y coordinate as our filter function and apply it to the data set (image B).

Fig 2. We now divide the image of the data set (under the filter function) into evenly distributed overlapping intervals (image C) and compute the corresponding points in their pre-image (image D).

Notice how each pair of overlapping intervals, defines two different subsets of data that can have elements in common.

Fig 3

Inside every defined subset of data, we execute a clustering algorithm to detect isolated groups of points (image E). Each of the resulting groups will correspond to a node on the output graph (image F). Notice how nodes are joined together by edges when their corresponding groups have points of the data set in common. Also, the size of the node in the cluster corresponds to the amount of points in its corresponding group.

Fig 4. Significant outbreaks of malaria in Colombia from 2007-2015, calculated using the scan statistic developed by [25] based on a likelihood ratio.

The significance threshold parameter was calculated using a Bernoulli model where cases were simulated for each municipality, and taking the maximum value. The process was iterated many times and the distribution of the maximum values was calculated to determine the 95% confidence interval.

Fig 5. Graph constructed using TDA and t-SNE component plot using the epidemic occurrence vectors, where selected groups have been highlighted.

These groups were selected by high overall disease intensity and high epidemic rate. Each cluster can be interpreted as a group of municipalities with Plasmodium falciparum incidence that have similar temporal behavior. Notice how the colored dots in the component plot are somewhat grouped together and since these represent municipalities with high epidemic rate, we have highlighted locations with several positive entries in the occurrence vector distributing differently across time.

Fig 6. Similar to Fig 5, highlighted groups where selected by high overall disease intensity and high epidemic rate for Plasmodium vivax.

Fig 7. Selected municipalities by TDA over the Colombian territory for both parasites.

As expected, the clusters follow some geographic pattern, since the time series where constructed using a Kulldorf procedure that detects clusters geographically. For Plasmodium falciparum all clusters are concentrated near the pacific coast and northern Antioquia. Unexpected results happen in cluster 5 for Plasmodium vivax, where the grouped municipalities belong to two different geographic regions of the country. This means that the municipalities in this cluster from Chocó and Amazonas have similar time pattern, regardless of their geographical distance.

Fig 8. Malaria by parasite species, age, sex, ethnicity and cluster groups of human cases in Colombia.

For all parasites, the indigenous ethnic group shows a pattern of endemicity, with most cases being reported for the youngest ages, while people with no ethnic denomination and the Afrocolombian population show a pattern consistent with occupational hazard risk. For Plasmodium falciparum, the indigenous and Afrocolombian populations in clusters 1, 2, 3 and 4 suggest that these populations experience intense exposure to malarial infection, with the Afrocolombian population showing occupational hazard transmission. Histograms for the population with no ethnic denomination in all clusters except 4 suggest malarial infection is associated with occupational hazard. And for Plasmodium vivax, the indigenous population in clusters 3, 4, and 5 suggest intense exposure to malarial infection among these populations. The population with no ethnic denomination experiences malarial infection as an occupational hazard in all clusters, except 4.

Fig 9. Total cases of Plasmodium falciparum by weeks, between the years 2007 and 2015.

Fig 10. Total cases of Plasmodium vivax by weeks, between the years 2007 and 2015.

Fig 11. Anthropogenic change in Colombia, 1999-2013, using the nighttime lights dataset NOAA-DMSP-OLS.

A mean for 5-year periods was computed for each pixel, and then map algebra was used to calculate the difference between the two periods. Very Rapid anthropogenic change was observed in the region of Bogotá and the Eastern Plains. Rapid change was observed in proximity of the main urban areas along the Andes (Bogotá, Cali, Medellín, and the Coffee Region, the urban areas of the Caribbean, and the Eastern Plains. Moderate and Medium anthropogenic change was observed throughout the Andes and the Caribbean, and the lower Cauca Basin.

Fig 12. Deforestation alerts in Colombia for years 2013-14, as published by SIAC [83].

Moderate and medium rates of deforestation were observed during the study period along the Pacific Coast, and throughout other parts of the country but more scattered. Rapid deforestation rates were observed in the lower Cauca Basin. Very Rapid deforestation rates were observed in Caquetá, the North Eastern Region of the Amazon Basin.

Fig 13. Density for evidence of alluvial gold exploitation in Colombia in 2016, as published by [84].

Three categories are included: Low (less 1 habitant per square kilometer), Medium (between 1.1 and 5 habitants per square kilometer) and High (more than 5 habitants per square kilometer). Evidence of intense mining activities was observed in the lower Cauca and Magdalena Basins, and the the Central Pacific region. Scattered mining activities were observed in the Southern Pacific, some parts of the Eastern Plains.

Fig 14. Malarial incidence for both species in Colombia from 2007-2015.

Intervals where constructed using the Jenks procedure [88].