Examining Spatial Patterns of Rabies in Arkansas, 1993–2021

Abstract

In this study, Moran's index of spatial autocorrelation (global) paired with the local Moran’s I (LISA) statistic was used to identify spatial patterns of rabies occurrence in Arkansas through time. Correlation and standard multiple regression were used to explore independent variables and their relationships to rabies. In addition, mean center and standard deviational ellipse calculations performed for total rabies cases separated by time period demonstrated a spatial progression of rabies in the state. The observed spatial pattern of rabies cases was significantly clustered when using global statistics at the scale of the entire study area in all experiments. Statistically significant clusters and outliers were present within the datasets in all LISA experiments and the locations of these clusters and outliers were shown to change through time. Correlation and multiple regression results revealed that mean population density (p/sq km) had a statistically significant direct relationship with the percentage of recorded rabies cases in all experiments, likely confirming bias due to sampling through passive surveillance. Percent farmland had a statistically significant inverse influence on the percentage of total recorded rabies cases, but not with percentages of recorded rabies cases separated by host organism which were based on smaller sample sizes. Mean elevation (m) had a statistically significant direct influence on the percentage of total recorded rabies cases and on the percentage of “skunk-unknown” rabies cases. Limitations apply. Explanatory variables may be contributing but not causal influencers. None of the regression coefficient values were high, so the influence of statistically significant independent variables was not large. Analyses of the rabies pathogen outside the context of a specific host may mask patterns unique to virus variants. Future research is encouraged.