Influence of sample size on the assessment of lung lesions in slaughter pigs.

摘要

The study is aimed at assessing the influence of sample size on the estimation of pneumonia and pleuritis in pig batches at the slaughterhouse. Data from lung examinations of 700 pigs were randomly selected from 10 herds. The frequency of pleuritis, pneumonia and extended pneumonia, and the mean and median pneumonia scores of the initial sample of 70 pigs, were calculated for each herd. These five parameters were calculated for 12 sub-samples of 10 to 65 pigs randomly selected from the initial sample of 70 pigs. Random samplings of these sub-samples were simulated 500 times and all five parameters were calculated for each simulation. The mean percentage relative discrepancy between each parameter of the initial sample and the estimated parameter of the sub-samples was calculated. The mean level of discrepancy increased as the sample size decreased, whatever the parameter. A sample size of 30 pigs should be sufficient to obtain a reliable and accurate estimate of the frequency and severity of pneumonia from the mean or median scores. A sample size of at least 60 pigs would be required to reliably estimate the prevalence of pleuritis and of pigs with high pneumonia scores.