Latent dimensional structure and typology of pig farm biosecurity using Biocheck.UGent™: a multivariate analytical study

<title>Abstract</title> <p>Understanding the multidimensional structure and classification of farm biosecurity is essential for designing targeted and effective disease prevention strategies. This study aimed to characterize biosecurity performance among commercial pig farms using the Biocheck.UGent™ scoring system and to evaluate its latent dimensional structure, internal consistency, and typological classification. A cross-sectional study was conducted among 76 commercial pig farms in Ogun State, Nigeria. External, internal, and overall biosecurity scores were obtained using the Biocheck.UGent™ tool. Descriptive and reliability analyses were performed, followed by principal component analysis (PCA) and Hierarchical cluster analysis (HCA), which was validated using linear discriminant analysis, multinomial logistic regression, and random forest modeling. Structural equation modeling (SEM) was used to evaluate latent biosecurity constructs and their relationship with overall biosecurity performance. Mean external, internal, and overall biosecurity scores were 48.37 ± 14.39, 65.92 ± 13.20, and 57.34 ± 11.26, respectively, indicating substantially weaker external biosecurity implementation. External biosecurity domains demonstrated moderate internal consistency (Cronbach’s α = 0.65). PCA identified two latent dimensions explaining 58.8% of the total variance. Hierarchical clustering identified four distinct biosecurity typologies, which were validated with high classification accuracy using linear discriminant analysis (94.7%) and multinomial regression (100%). Random forest modeling confirmed transportation, animal introduction, and visitor biosecurity as key determinants of typology. SEM demonstrated that external biosecurity represented a significant latent construct and was a strong predictor of overall biosecurity performance (standardized effect = 1.201, p &lt; 0.001). Biosecurity implementation in commercial pig farms is a multidimensional construct characterized by distinct and reproducible typological patterns. External biosecurity, particularly animal introduction and transportation practices, represents the primary determinant of overall biosecurity performance. These findings highlight the importance of targeted, domain-specific interventions and demonstrate the value of multivariate structural modeling for improving biosecurity surveillance and disease prevention strategies in pig production systems.</p>