| Popis: |
A novel quality assurance approach was tested for its applicability to assess data validity and meat inspection performance by means of modeling and training of official meat inspectors (OMIs). General linear mixed models (GLMM) were used to estimate the variance in 20 selected lesions assessed by 12 official meat inspectors for 247.507.pigs. The pigs originated from 72 conventional farms and were slaughtered at one abattoir in Austria in the period January 2008 to June 2011. The heterogeneity in the health state of the pigs, variation due to the farms of origin, piglet producer, quarter and medication was considered in the statistical model. Based on the model meat inspection showed hardly any variation for post-mortem findings such as pericarditis, arthritis and milkspots, whereas meat inspection of blood aspiration, scalding water lungs, skin lesions and hepatitis can be deemed as not sufficiently standardized. Training of OMIs resulted in better detection rates of blood aspiration and pleurisy, but not in the detection of skin lesions related to slaughter technology and animal welfare. Grading of pneumonia has to be further improved in future training sessions. An improved data record system was established based on the outcome of the model and training of OMIs. In future research the benefit of the new code system and training effect to standardize meat inspection will be assessed in form of statistical re-evaluations. |
