Next, the effect of every observed variable (Tables 2 and 3) on the
pHLT and the PSE prevalence was examined separately by using a
mixed model, whereby slaughterhouse and sampled group of pigs
(batch), nested within slaughterhouse, were used as random factors.
Only variables classified as significantly (p b 0.05) influencing the
pHLT and the PSE prevalencewere considered in themodel. Correlations
between variables were calculated to check for multicolinearity
problems. If significant correlations were found between covariates
(|r| N 0.6), the most adequate and representative variable was kept in
the model. Insignificant variables and variables with less than 80% of
themeasurementswere left out of the model used to build the checklist
or combination table. Finally, for every significant variable, criteriawere
defined to classify from which value the significant variable has a positive
or negative effect on the pHLT. If every criteria, for each phase, was
performed to ensure a pHLT value, which is considered as indicating no
risk for PSE (pHLT ≥ 6), the phase was qualified as ‘Ok’; if not the phase
was qualified as ‘Not Ok’ (pHLT b 6). To conclude all this information was
merged in one table,whereby combinations of different situationswere
put together and the remaining pHLT and the potential risks to develop
PSE meat (%), for the specific situations, were calculated