The objective of this study was to identify characteristics of milk production systems from farms in Minas Gerais, Brazil, with major expenses on animal disease prevention. Furthermore, the study aimed at investigating the existence of the negative relationship between prevention and treatment expenditures. The data were obtained from 861 questionnaires from the "Diagnóstico da Pecuária Leiteira do Estado de Minas Gerais em 2005" publication. Thus, factor analysis was utilized to generate variables that are linear combinations of chosen expenditures. These factors were used in a Three-stage Least Squares regression model, which allows considering the expected endogenity between treatment and prevention expenditures The results indicated that treatment costs increase with increasing the percentage of milking cows and that expenses on prevention increase with production intensification. Farms that use predominantly family labor have lower treatment costs and higher expenses on prevention. The prevention equation has shown that systems with higher expenses with pasture, sugar cane plantations and labor costs spend more on prevention. In the treatment equation, the significant factor indicated that more intensive systems have greater probability of disease transmission and thus, greater treatment costs. Properties that received more than six visits of a technician per year have higher treatment costs but lower prevention costs. The association between treatment and prevention spending was contrary to the expected, showing that while spending on prevention increases, treatment spending also increases. Thus, that producers have a reactive rather than proactive attitude regarding cattle health.
animal health management; factor analysis; prevention expenditures; three-stage least squares; treatment expenditures
