A Computational Re-evaluation of Spatial Trials for Zoonotic Tuberculosis Control: Model Misspecification, Diagnostic Miss-classification, and the Illusion of Wildlife Culling Efficacy

This scholarly examination scrutinizes the historical Randomised Badger Culling Trial (RBCT), a pivotal empirical study shaping zoonotic tuberculosis (Mycobacterium bovis) control strategies. The researchers challenge the RBCT's trade-off hypothesis, which suggests that lower cattle infections are balanced by a disruption effect due to altered host dispersal patterns.

By employing generalized linear mixed models with a generalized Poisson error distribution to account for historical data overdispersion, the study compares standard parametric methods against cluster-constrained permutation tests. The findings indicate that earlier reported treatment and perturbation results are merely statistical quirks.

Inside the culling zones, the parametric significance cannot be validated through exact permutation verification due to extreme data influence in localized cluster blocks. A critical revelation is that when diagnostic misclassification biases are removed by examining the total reactor datasets, all visible culling effects vanish, and information criteria overwhelmingly support nested null models.

The study suggests that unconfirmed reactors might represent genuine biological infections overlooked by low-sensitivity post-mortem macro-necropsy, implying that host removal influences observed noise rather than authentic zoonotic transmission.

The empirical scaling in the research uncovers a novel mathematical saturation effect due to unconsidered disease recurrence, rather than an ecological reality. The study concludes that existing zoonotic tuberculosis intervention frameworks are based on fundamentally flawed statistical models, leading to misguided, large-scale veterinary actions that cause significant ecological and economic damage without delivering genuine public health, animal health, or disease control benefits.