Nonsteroidal anti-inflammatory drug therapy for neonatal calf diarrhea complex: Effects on calf performance

Todd, C. G.; Millman, S. T.; McKnight, D. R.; Duffield, T. F.; Leslie, K. E.

It should be noted that 4 highly influential observations were included in the final Cox proportional hazards model for time to weaning. Specifically, 2 MEL-treated and 2 PLA-treated calves experienced poor starter ration consumption and did not fulfill the weaning criteria by the conclusion of the 56-d observation period. Therefore, these calves were handled as censored animals in the weaning age analyses. During model validation, these censored calves were identified as having a large impact on the regression coefficients of the final Cox proportional hazards model for time to weaning. Outliers and influential observations can be excluded from analysis, but only if appropriate justification can be provided. If these 4 calves were removed from the data set, the median time to weaning for MEL-and PLA-treated calves would be 40 and 46 d, respectively. Furthermore, treatment would be associated with a hazard ratio of 2.27, indicating that MEL treatment reduced time to weaning by 127%. However, a valid reason for excluding these data from analysis could not be identified. Thus, the 4 very influential calves were included in the final model for time to weaning, even though they may have masked some of the effect of MEL therapy.