Diferentes abordagens estatísticas para o estudo de fatores de condenação de carcaças em frangos de corte

Abstract

Carcass condemnation in broiler chickens represents one of the main causes of economic losses in the Brazilian poultry industry, in addition to raising sanitary and ethical concerns related to animal welfare. This study aimed to identify and quantify the main factors associated with partial and total carcass condemnations, as well as the occurrence of pododermatitis, based on a dataset comprising more than 266 million slaughtered birds. Two distinct statistical approaches were applied: fixed-effects analysis of variance (ANOVA) and random-effects models (REML). The results revealed high heterogeneity (I² > 99.84%; p < 0.001) within the dataset, justifying the adoption of the random-effects model for more generalizable inferences. Factors such as housing type, stocking density, sex of the birds, and geographical region were significantly associated with the evaluated outcomes. The REML approach not only allowed the identification of the main risk factors but also enabled the estimation of their relative impact, providing valuable insights for the formulation of more effective mitigation strategies in the context of intensive broiler production. For partial condemnations, the moderators that best explained the variation were slaughter region (R² = 25.58%), litter reuse cycles (R² = 18.47%), and downtime between flocks (R² = 12.32%). For total condemnations, the most relevant factors were average daily gain (R² = 30.68%), slaughter region (R² = 30.68%), and downtime between flocks (R² = 21.64%). Regarding pododermatitis, the main explanatory factors were slaughter region (R² = 66.76%), litter reuse cycles (R² = 40.54%), and slaughter weight (R² = 36.59%). These findings contribute to the improvement of management protocols and the optimization of production processes in the poultry industry.