The objective of this study was to assess how producers categorized alerts from a daily generated alert list designed to identify sick or injured animals. Data from 4 commercial farms in Kentucky were collected from October 2015 to 2016. Each cow was equipped with a CowWatch leg tag (Alta Genetics Inc., Watertown, WI) measuring steps (steps/d) and lying time (h/d), and a neck tag measuring eating time (h/d). Dairy producers evaluated a technology-generated daily herd health report where alerts were generated based on a cow level threshold of ≥30% decrease in steps, lying time, or eating time compared with each cow’s 10-d moving average. Producers evaluated alerts within overall categories: alert accepted to be true and cow checked (A), alert accepted to be true and cow not checked (B), and alert not accepted to be true (C). A total of 25,027 cow alerts occurred, with producers evaluating 15,644 cow alerts (62%). The FREQ procedure of SAS 9.4 (SAS Institute Inc., Cary, NC) with a Chi-squared analysis was used to assess category distribution. Overall, more alerts were categorized as A (5,129) and B (8,424) compared with C (2,091; P < 0.01). Producer categorization by farm (Table 1) indicated most of evaluated alerts were accepted to be true (75 to 99%, total A and B). However, 27 to 88% of alerts were categorized as B, indicating although behavior changes were real, the producer did not check the cow. Reasons for not checking cows included: changes from normal not enough for producer to worry, no time for producers to check alerts, changes in ambient temperature, and hoof trimming or veterinary checks. The frequency of A and B alerts indicated that producers believed the behaviors changes identified were real. The frequency of cows in category B indicated alert generation should be refined to identify only sick or injured cows. Table 1.

Item	No. (% of total)
	Farm 1	Farm 2	Farm 3	Farm 4
Category A	105 (2%)	429 (7%)	1,599 (28%)	2,996 (44%)
Category B	2,802 (42%)	2,708 (45%)	1,778 (32%)	1,136 (17%)
Category C	292 (4%)	609 (10%)	1,151 (20%)	39 (1%)
Missing	3,427 (52%)	2,282 (38%)	1,100 (20%)	2,574 (38%)