Audit-Based Deviation Patterns in Trichinella Testing Establishments: Evidence from Southern Spain

Deviation patterns identified through quality management system (QMS) audits in Trichinella spp. meat inspection may vary across species processed, activity idiosyncrasies and audit scenarios. This study retrospectively analysed 3023 category-level audit records (2012–2018) from 18 supervised Satellite Laboratories (SLs) in southern Spain. Graphical canonical discriminant analysis (GCDA) explored differences in audit-related characteristics across groups defined by species/breed, activity and audit scenario. Following multicollinearity screening, final models retained two variables: auditor combination and audit-derived risk description. Statistically significant differences were identified across groups, although relatively low canonical correlations and record-level cross-validation indicated limited discrimination. Species/breed classification was particularly challenging, with poor identification of Iberian pig records and confusion between wild boar and White pig records. Complementary cluster-aware analyses examined deviation counts, classifications, requirement scopes, deviation types and audit scores, accounting for laboratory–year clustering. These analyses provided additional insights into deviation-related characteristics beyond those retained in the final GCDA. Overall, findings demonstrate the value of routine audit data for characterising QMS implementation across operational contexts and identifying opportunities to strengthen audit practices and support continuous improvement. Although diagnostic effectiveness, detection reliability and meat safety were not directly assessed, these results highlight the complementary role of audit information in QMS monitoring.