RESUMEN
OBJECTIVE: To evaluate a predictive model's ability to determine cattle mortality following first and second treatment for bovine respiratory disease and to understand the differences in net returns comparing predictive models to the status quo. METHODS: 2 boosted decision tree models were constructed, 1 using data known at first treatment and 1 with data known at second treatment. Then, the economic impact of each outcome (true positive, true negative, false positive, and false negative) was estimated using various market values to determine the net return per head of using the predictive model to determine which animals should be culled at treatment. This was compared to the status quo to determine the difference in net return. RESULTS: The models constructed for the prediction of mortality performed with moderate accuracy (areas under the curve > 0.7). The economic analysis found that the models at a high specificity (> 90%) could generate a positive net return in comparison to status quo. CONCLUSIONS: This study showed that predictive models may be a useful tool to make culling decisions and could result in positive net returns. CLINICAL RELEVANCE: Bovine respiratory disease is the costliest health condition experienced by cattle on feed. Feedyard record-keeping systems generate vast amounts of data that could be used in predictive models to make management decisions. It is essential to understand the accuracy of predictions made via machine learning. However, the economic impact of implementing predictive models in a feedyard will influence adoption.
RESUMEN
OBJECTIVE: To evaluate predictive model ability to determine whether an animal finished the feeding period using data known at first treatment for bovine respiratory disease (BRD). Additional comparisons evaluated the potential benefits of predictions by adding weather data, utilizing balancing techniques, and creating models for individual feedyards. ANIMALS: This retrospective study included animal, pen, and feedyard data from 12 US feedyards from 2016 to 2021. The final dataset consisted of 96,382 BRD cases of which 14.2% did not finish the feeding phase. PROCEDURES: Five predictive models were trained and underwent threshold probability adjustment to maximize F1 score. Model performance was evaluated using accuracy, sensitivity specificity, positive and negative predictive values, and area under the receiver operating characteristics curve (AUC). RESULTS: Overall, model performance was low with a median AUC value of 0.675. The addition of weather data had little effect on AUC but resulted in more variation in sensitivity and specificity. Resampling the dataset had a limited effect on performance. Individual feedlot models had higher AUC values than others with the decision tree typically performing best in most feedyards. CLINICAL RELEVANCE: Results indicated some utility of predictive models evaluating BRD cases to predict cattle that did not finish the feeding phase. These models could be valuable in assisting health providers making decisions on individual cases.