2.4. Statistical analysis
The data were stored in EpiInfo 6.04 software (CDC, Atlanta,
GA, USA) database. A case of prior Leptospira infection was
defined as having a positive IgM ELISA reaction. Univariate
analysis was performed to evaluate risk factors for acquiring
a prior Leptospira infection. The 2 and Fisher’s exact tests
were used to determine significance between proportions.
ANOVA and Wilcoxon’s tests were used to assess significant
differences for continuous data. The influence of rainfall
on number of cases of leptospirosis was analysed in a linear
regression model that included the year, the yearly rainfall
value and the number of cases notified. The net effect
of rainfall was measured by the semi-partial correlation
between the last two variables.