Spss program for windows version 19

The statistical analysis was performed using SPSS program for Windows version 19.0 (SPSS Inc, Chicago, IL, USA). Statistical associations between ante mortem rejections and causes of post mortem carcass condemnations with the factors season were performed. Interactions between year and season were also examined. Regression models were compared using contrast Chi-square test and the Akaike information criterion. Finally, negative binomial model was used to evaluate the effect of season on the rate of antemortem rejections and post-mortem condemnations. In this analysis only causes of condemnations with more of twenty cases were included. The exposure was the total of pigs entering the abattoir each month and the outcome was the number of animals condemned for each reason of the condemnation. Winter was the season of reference because there are a lower number of postmortem condemnations, which allows a better comparison of the values of the variables. Regression coefficient β (and confidence interval) was transformed into risk ratio (RR) taking the exponent of the coefficient as follows: RR = exp (β). RR is statistically significant when the 95% confidence interval (CI) does not include the unit. A RR>1 means that the incidence rate increases. A RR<1 means that incidence rate is reduced. A p-value <0.05 level was considered as statistically significant.

An initial power analysis based on the pilot results determined that 15 subjects yielded a power of 0.8 at a significance level of 0.05. Statistical analyses were performed using the SPSS program for windows version 19.0 (SPSS Inc., Chicago, IL, USA). In addition to standard descriptive statistical calculations (means and standard deviations), ANOVA was used to determine intra- and inter-group statistical differences. When ANOVA yielded results indicating significant difference between the groups, a Tukey’s test was performed. The mean values were followed by 95% confidence intervals, and all the data were expressed as the means±standard deviations. The statistically significant levels were pre-determined at P<0.05. The post-hoc power analysis was performed, and the power was estimated to be>0.95.