Evaluation of a machine learning tool to screen for hypoadrenocorticism in dogs presenting to a teaching hospital

Reagan, K. L., Pires, J., Quach, N., & Gilor, C. (2022). Evaluation of a machine learning tool to screen for hypoadrenocorticism in dogs presenting to a teaching hospital. Journal of Veterinary Internal Medicine. https://doi.org/10.1111/jvim.16566

This study evaluated a machine learning model (MLM) for screening hypoadrenocorticism (HA) in dogs presenting to a veterinary hospital. Out of 1025 dogs screened, 12 were MLM-positive for HA, with 5 confirmed true positive cases. Three MLM-positive cases were classified as "undetermined" due to incomplete diagnostics, and 4 were false positives. The MLM demonstrated high accuracy in predicting HA with acceptable false positive rates, especially in a population with low HA prevalence.