Spatial Patterns of Leptospira Seroprevalence

The data about each animal, the characteristics of each farm and laboratory results were entered into an Excel database (Microsoft, Redmond, WA, USA). Prevalence of anti-Leptospira spp. antibodies with the 95% Confidence Interval (CI) was estimated. Also, farms positivity was determined. The association between the outcome seropositivity to Leptospira spp., to Leptospira serogroup Sejroe (adapted serogroup) and to Leptospira serogroup Pomona (incidental serogroup) and the variables under analysis was assessed by a bivariate analysis using a Chi-squared test. Fisher's exact test was used if one or more cells expected value was <5. Odds ratios (OR) and 95% CI were also estimated and calculated for each variable. For quantitative variables, parametric or non-parametric tests were used. The null hypothesis was that there were no differences between groups. All the statistical tests were carried out at α=0.05. Quantitative and qualitative variables at p-value < 0.2 in the bivariate analysis were analyzed by a Generalized Linear Mixed Model (GLMM) with random effect of farm-level risk. All statistical analyses were performed with software R v. 4.0.2 (10 ).
Potential spatial clusters were investigated in the study area with space scan statistics using SaTScan software, v10.0.2. Poisson model for high rates was performed for detecting spatial patterns of the number of MAT positive events in a geographical location, taking each farm as a unit for analysis according to a known population at risk (19 (link)).
Another qualitative variable called “being inside the spatial clusters” was generated, and an animal was considered to be exposed to the variable when it belonged to a farm inside a significant cluster. The animal was not exposed to the variable when it belonged to a farm outside the significant clusters. GLMM was performed with the significant variables detected in the first GLMM with a new variable, “being inside the spatial clusters.”