The objective of this study was to evaluate four different automated activity monitoring (AAM) systems to identify anovulatory cows in early lactation. A total of 852 lactating Holstein cows (221 primiparous and 631 multiparous cows) from four commercial dairy herds were enrolled. On each farm, cows were equipped with a respective AAM system (SB: Smartbow; HT: Heatime; DP: Delpro; and CM: CowManager). Each cow was sampled three times within the voluntary waiting period (VWP) in a two-week interval to detect the blood progesterone (P4) concentration. Cows were classified based on the concentration of P4 as follows: (1) none of the three blood P4 concentrations exceeded 1.0 ng/mL (anovulatory); (2) at least one of the three blood P4 concentrations was above 1.0 ng/mL (ovulatory). Cows were classified based on estrus alerts as follows: (1) no estrus alert was detected by an AAM system from 7 to 60 DIM (anestrus); (2) at least one estrus alert was detected by an AAM system from 7 to 60 DIM (estrus). Accuracy, sensitivity, specificity, positive predictive value, and negative predictive value were calculated for each AAM system for anovulatory cows [(SB: 77.6%; 26.8%; 89.3%; 36.7%; and 84.1%); (HT: 79.2%; 63.6%; 83.6%; 52.8%; and 88.9%); (DP: 47.2%; 78.8%; 41.5%; 19.5%; and 91.6%); (CM: 80.5%; 23.7%; 92.7%; 39.1%; and 85.3%)].
