A fluid milk spoilage simulation framework reveals the need for spoilage intervention strategies that account for frequency of bacterial postpasteurization contamination

Abstract

Postpasteurization contamination (PPC) with gram-negative bacteria and the growth of spore-forming bacteria are major causes of fluid milk spoilage, typically leading to sensory defects when bacterial concentrations exceed 6 log10 cfu/mL. Existing models focus on individual spoilage pathways, limiting their ability to capture the complexity of milk spoilage. To address this, we developed a simulation framework that simultaneously models the growth of both types of bacteria in high-temperature short-time pasteurized milk along a supply chain. Dairy processing plants were categorized into 3 groups based on shelf life of historical fluid milk: long, medium, and short shelf-life plants. We assumed varying initial PPC frequencies for each category, with ranges of 0% to 33% for long, 34% to 66% for medium, and 67% to 100% for short shelf-life plants. Shelf life, defined as when 25% of milk containers exceeded 6 log10 cfu/mL, was predicted as 25, 12, and 8 d for long, medium, and short shelf-life plants, respectively. Our predictions aligned with observed bacterial growth in commercial milk stored at 6°C, with the percentage of milk containers exceeding 6 log10 cfu/mL on d 14 falling within the 5th to 95th percentiles of simulated values. Sensitivity analysis identified key parameters influencing bacterial concentrations at shelf-life d 7, 14, and 21 for long, medium, and short shelf-life plants, respectively, guiding intervention strategies. What-if scenario analysis revealed that effective intervention strategies to extend shelf life vary by plant categories. While interventions targeting spore-forming bacteria, such as microfiltration, bactofugation, and improved home storage conditions, extended the shelf life for long shelf-life milk by 3 to 5 d, PPC reduction extended shelf life by 5 and 4 d in medium and short shelf-life plants, respectively. This simulation framework provides a comprehensive spoilage prediction tool to support data-driven decision making for fluid milk processors.