Integrated mycotoxin risk management strategies for grain handling and feed manufacturing industries
Date
2023-08
Authors
Branstad-Spates, Emily Hannele
Major Professor
Advisor
Mosher, Gretchen A
Bowers, Erin L
Hurburgh, Charles R
Dixon, Philip M
Appuhamy, Ranga J
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Abstract
Major advancements have been made regarding mycotoxin management in food and feed products; this can largely be attributed to improvements in knowledge of how mycotoxins develop from fungi under certain environmental conditions, incorporating advanced technology, and the great importance of utilizing integrated mycotoxin management strategies as a holistic approach. Mycotoxin management is dire, as many associated challenges include economic losses, reduced crop yields and value, detrimental effects on livestock production parameters, and human health impacts. This research explored integrated mycotoxin risk management strategies in grain handling and feed manufacturing industries in the United States (US) with a particular focus on aflatoxins (AFL). No matter how adequate the risk management strategy is for AFL, it must be implemented by stakeholders in the integrated supply chain. Therefore, stakeholders’ perceptions of mycotoxin risk influence the degree to which they may properly interpret and implement a management strategy. Improved predictive models can improve stakeholder confidence in their investments in mycotoxin management and their decision-making. The first study in this research aimed to identify risk perceptions associated with mycotoxin management in grain handling and feed manufacturing facilities in the Midwest region of the US among three stakeholder groups. Qualitative, semi-structured interviews and a thematic analysis identified six major themes: 1). challenges throughout the integrated grain and feed supply chain, 2). communication, 3). future research, 4). mitigation strategies, 5). mycotoxin occurrences, and 6). quality and food safety management. Furthermore, differences were observed among stakeholder groups with regard to their perception of mycotoxins as a hazard on a risk assessment scale and mycotoxin predictions in the future. These findings contribute to a broader effort to understand relevant supply chain stakeholder perceptions of mycotoxins and to provide information on how to better inform risk management strategies, tying into the next objective of understanding influential factors for AFL contamination in Iowa corn. In 2012, hot drought conditions in the US Corn Belt raised concerns about widespread AFL contamination. The Iowa Department of Agriculture and Land Stewardship (IDALS) sampled Iowa corn to assess the state's incidence and severity of AFL contamination. Three hundred ninety-six samples were analyzed for AFL; the statewide average mean for all tested samples was 5.57 ppb. Compared with the rest of the state, AFL levels were significantly higher in the Southwest (SW; mean 15.13 ppb) and South Central (SC; mean 10.86 ppb) crop reporting districts (CRD). Using the case study as a guideline, the next objective aimed to evaluate the performance of AFL prediction in Iowa corn utilizing gradient boosting machine (GBM) learning and feature engineering. Historical monthly meteorological and soil property data from four years, combined with historical Iowa AFL data collected in the same four years (2010, 2011, 2012, and 2021), were used in the GBM model for two AFL risk thresholds for high contamination events: 20-ppb and 5-ppb. An overall accuracy of 96.77% was achieved for AFL prediction with the GBM model, with a balanced accuracy of 50.00% for a 20-ppb risk threshold, whereas an overall accuracy of 90.32% with a balanced accuracy of 64.88% was developed for a 5-ppb threshold. Mycotoxin prediction models are practical and implementable in commodity grain handling environments and can aid in achieving the goal of being preventative versus reactive to outbreaks. Therefore, the last study aimed to 1). Understand the probabilities of pivotal events with AFL in corn at FSMA-regulated entities in food and feed safety systems through an event tree analysis (ETA), and 2). Propose recommendations based on factors identified through the ETA and Iowa-centric model for AFL risk management. The ETA was used to systematically evaluate pivotal events and hypothetical safety-decision-making scenarios for corn users based on the AFL risk each year calculated by the Iowa-centric model. The results showed four single-point failures (SPF) within the food and feed safety systems for AFL control, with an overall failure of 49.86% and 50.14% success rate, respectively. This research concluded that risk-informed decisions are needed for effective AFL monitoring and prediction in Iowa corn with timely intervention strategies to minimize the overall effects on end-users.
Series Number
Journal Issue
Is Version Of
Versions
Series
Academic or Administrative Unit
Type
dissertation