Álvaro García
With an annual cost nearly reaching a billion dollars, Bovine Respiratory Disease (BRD) in the US poses a dual challenge in feedlots, impacting both cattle health and the economy of that industry. Upon cattle arrival at the feedlot, decisions regarding antimicrobial preventive health measures are based on evaluating the risk of BRD development within each group. Using data available to predict and classify cattle groups into high- or low-risk could optimize antimicrobial use, potentially improving economic performance. Identified cohort risk factors include average body weight, sex, group size, and pen housing conditions. However, the diverse factors contributing to perceived risk introduce complexities, potentially leading to inaccuracies in classification and subsequent consequences for cattle health and financials.
The economic impact of BRD
In a comprehensive 2011 USDA National Animal Health Monitoring System (NAHMS) survey across 561 feedlots in 21 US states, BRD emerged as the primary cause of morbidity and mortality. Moreover, earlier research shed light on the wide-ranging effects of BRD on marbling scores, quality grade, hot carcass weight, and yield grades. Notably, calves treated for BRD witness diminished returns, with the financial impact intensifying alongside the number of medical treatments. It’s crucial to recognize BRD is not only a viral/bacterial disease but also a condition influenced by various stressors, contributing to its multifaceted nature, and affecting immune function. Accurate diagnosis thus becomes crucial, and traditional methods involve visual appraisal and clinical scoring based on signs such as nasal or ocular discharge, depression, lethargy, emaciated body condition, labored breathing, and elevated rectal temperature. However, these methods are subjective, and their accuracy may be compromised.
In the NAHMS survey BRD affected a significant proportion of cattle, with 87.5% of the animal’s receiving treatment. Injectable antibiotics were the predominant choice, administered in 99% of feedlots, with an average cost of $23.60 per case. Larger feedlots reported a slightly higher cost of $23.90 compared to smaller ones with $23.40. Although there isn’t a more recent study, we can estimate costs for 2024 based on inflation rates, which results in a projected cost per case to be nowadays approximately $32.04. Actual costs may vary due to market complexities, technological advancements, and industry-specific factors.
Research has also shown that successful response rates are observed in 82% of cases after the initial treatment, decreasing to 38% by the third treatment. Achieving positive outcomes through early diagnosis and prompt intervention is crucial. Non-steroidal anti-inflammatory drugs, combined with antimicrobials, lead to a faster reduction in rectal temperature and fewer lung lesions at slaughter. A recent study also suggested that twice as many cattle exhibited lung lesions at slaughter compared to recorded pneumonia treatments, indicating potential overlooked pneumonia cases.
According to the NAHMS study, an estimated 16.2% of cattle in feedlots showed signs of BRD. One could then calculate expenses for treating BRD in a 1,000-head feedlot in 2024 to be approximately $5,190 (162 × $32.04). However, estimating costs for 162 cases of BRD should also factor treatment success rates and potential missed cases mentioned above. If the success rate after the first treatment is 82%, decreasing to 38% by the third treatment, and assuming twice as many cattle have undetected lung lesions at slaughter, the total estimated cases amount to 491. At an average cost of $32.04 per case, the projected expense for these 162 BRD cases is approximately $18,278.68, emphasizing the significance of efficient management practices in mitigating economic challenges associated with BRD in the cattle industry.
A modern approach to diagnose respiratory problems
Cattle respiratory behavior, characterized by abdominal movements, is integral for automatic diagnosis of respiratory-related ailments, like for example heat stress. However, monitoring multiple animals in expansive and dynamic farm environments presents its challenges. Traditional methods, like visual appraisal, are labor-intensive and prone to fatigue by the individual. Cattle contact devices, though used, have drawbacks, including animal stress responses and limited scalability. Noncontact methods, especially those employing computer vision, offer a promising alternative. In livestock farming, computer vision has been instrumental in various applications, leveraging low cost, high efficiency, and information richness.
However, the challenge lies in effectively monitoring the respiratory behavior of multiple animals in a group, delaying the development of automatic monitoring systems in precision farming. While advancements in computer vision and deep learning have led to achievements in livestock respiratory behavior monitoring, methods often focus on individual characteristics related to breathing, like abdominal fluctuations.
An experiment conducted by Wu et al. (2023) aimed to automatically monitor the breathing behavior of multiple dairy cows in a farm setting. Traditional manual observation of cows in groups can be challenging, hindering health assessment. To address this, researchers utilized a sophisticated computer vision model trained to recognize and distinguish individual cows in images called YOLACT, short for “You Only Look At CoefficienTs. The model acts as a smart tool proficient in understanding pictures, efficiently drawing virtual lines around each cow, indicating their positions and whether they are standing or lying down. This segmentation is crucial for understanding the cow’s behavior. The computer model performs these tasks in one go, eliminating the need to analyze pictures twice, making it fast and efficient. After training in numerous cow images, the model became adept at spotting individual cows and determining their positions. In the experiment, the researchers tailored the model to focus on monitoring how cows breathe, providing valuable insights into their health, especially when resting. Once the resting states were identified, detection algorithms were applied to monitor the cows’ respiratory behavior. The researchers tested the system using 60 videos that simulated real-world conditions, including various factors that could interfere with accurate monitoring.
The results were promising, with the proposed method demonstrating high accuracy. In 54 out of 60 videos, the accuracy exceeded 90%, and in 4 videos, it reached 100%. The average accuracy across all tests was 93.56%, showcasing the effectiveness of the algorithm in monitoring respiratory behavior even in challenging situations. This technology has the potential to significantly improve the efficiency of health assessment for dairy cows, allowing for timely intervention and care. It lays the groundwork for the development of automated systems that can contribute to precision livestock farming, where technology helps optimize animal health and welfare.
Implications
To address the economic challenges posed by Bovine Respiratory Disease (BRD) in the American feedlot industry, the research by Wu et al. has proposed a novel approach utilizing computer vision and deep learning. By efficiently monitoring cattle respiratory behavior in feedlots, the method would involve the identification and segmentation of individual cattle, allowing for single-target respiratory behavior monitoring. The method exhibited strong performance, accommodating factors such as occlusions, varied animal resting states, and changes in the environment.
This study lays the groundwork for future advancements in respiratory behavior monitoring and the automated diagnosis of respiratory-related issues in cattle. Further refinements to the monitoring algorithm aim to enhance efficiency and adaptability, making it suitable for deployment on smaller computing platforms like embedded and edge devices. The application of precision livestock farming technologies, including inspection robots with respiratory behavior monitoring capabilities, holds promising potential for reducing manual labor and improving overall cattle management practices in feedlots.
In conclusion, the proposed method not only addresses the challenges associated with manual monitoring but also aligns with the industry’s trajectory toward precision livestock farming, where technological innovations play a pivotal role in optimizing animal health and welfare.
© 2025 Dellait Knowledge Center. All Rights Reserved.
