Our objective was to explore change in activity variables relative to clinical identification of bovine respiratory disease (BRD) in cattle maintained in a commercial feedlot. Four blocks (n=364) of high-risk, crossbred beef calves were affixed with an accelerometer device upon feedlot arrival that continuously logged activity variables (standing time, step count, lying bouts, motion index) for a 56-day period. Overall, 51.5% of calves were diagnosed at least once with BRD, while 15.2 and 4.5% had a second and a third BRD event, respectively. The level of activity in cases was compared with that of control cohorts relative to the time of BRD diagnosis in cases. Reductions in activity variables in cases started at least 6 days prior to BRD diagnosis, and were more pronounced the day before disease identification. All activity variables were reduced (P?0.01) at day -1 for cases vs controls. Average standing time on the day prior to diagnosis (day -1) was 559 minutes for cases compared to 613 minutes in controls. Step count on day -1 for cases and controls were 843 and 1,472 steps, respectively. The number of lying bouts for cases and controls was 11.4 and 14.5, respectively on day -1. Activity information provided by accelerometers, used as an objective method for identification of BRD in cattle, may assist in management and early detection of sick cattle.