Ketosis is a common metabolic disorder affecting dairy cattle that results in a greater risk for the development of disease. The Dairy Herd Improvement Association (DHIA; CanWest DHI, Guelph, ON, Canada) have begun to utilize an infrared test (MilkoScan FT600, Foss Analytical A/S, Hillerød, Denmark) to detect milk β-hydroxybutyrate (BHB) as a herd surveillance test. However, the test has not been compared with the gold standard serum BHB as determined by a referent laboratory. The objective of this cross-sectional diagnostic accuracy study was to validate the DHIA milk BHB test to identify cows with hyperketonemia as determined by quantification of BHB in serum. A total of 320 cows from 17 dairy herds in southwestern Ontario with milk collected at routine DHIA collection and blood sampled within 24 h of the milk test. The BHB concentration in milk was determined using the DHIA milk BHB test and serum was sent to the Animal Health Laboratory (AHL). A nonparametric receiver operating characteristic (ROC) curve was generated to compare DHIA milk BHB concentrations to serum BHB concentrations determined at a laboratory. A total of 33 cows (11%) had a level of serum BHB >1.2 mmol/L. A linear regression model demonstrated that the concentration of DHIA milk BHB was moderately correlated with the concentration of serum BHB yielding an R² value of 0.61. The optimal cut point to determine hyperketonemia (>1.2 mmol/L) on the DHIA milk BHB test was ≥0.135 mmol/L, yielding a sensitivity of 79% and specificity of 92%. The performance of the DHIA milk BHB test varied depending on the DIM of the cows tested and the herd-level prevalence of hyperketonemia. This study demonstrates that the DHIA milk BHB test is a reliable measure to evaluate hyperketonemia when tested at routine DHIA collection and could be used as a monitoring tool for ketosis.