Evaluating the cost implications of a radio frequency identification feeding system for early detection of bovine respiratory disease in feedlot cattle.

New technologies to identify diseased feedlot cattle in early stages of illness have been developed to reduce costs and welfare impacts associated with bovine respiratory disease (BRD). However, the economic value of early BRD detection has never been assessed. The objective was to simulate cost differences between two BRD detection methods during the first 61 d on feed (DOF) applied in moderate- to large-sized feedlots using an automated recording system (ARS) for feeding behavior and the current industry standard, pen-checking (visual appraisal confirmed by rectal temperature). Economic impact was assessed with a cost analysis in a simple decision model. Scenarios for Canadian and US feedlots with high- and low-risk cattle were modeled, and uncertainty was estimated using extensive sensitivity analyses. Input costs and probabilities were mainly extracted from publicly accessible market observations and a large-scale US feedlot study. In the baseline scenario, we modeled high-risk cattle with a treatment rate of 20% within the first 61 DOF in a feedlot of >8000 cattle in Canada. Early BRD detection was estimated to result in a relative risk of 0.60 in retreatment and 0.66 in mortality compared to pen-checking (based on previously published estimates). The additional cost of monitoring health with ARS in Canadian dollar (CAD) was 13.68 per steer. Scenario analysis for similar sized US feedlots and low-risk cattle with a treatment rate of 8% were included to account for variability in costs and probabilities in various cattle populations. Considering the cost of monitoring, all relevant treatment costs and sale price, ARS was more costly than visual appraisal during the first 61 DOF by CAD 9.61 and CAD 9.69 per steer in Canada and the US, respectively. This cost difference increased in low-risk cattle in Canada to CAD 12.45. Early BRD detection with ARS became less expensive if the costs for the system decreased to less than CAD 4.06/steer, or if the underlying true BRD incidence (not treatment rate) within the first 61 DOF exceeded 47%. The model was robust to variability in the remaining input variables. Some of the assumptions in the baseline analyses were conservative and may have underestimated the real value of early BRD detection. Systems such as ARS may reduce treatment costs in some scenarios, but the investment costs are currently too high to be cost-effective when used solely for BRD detection compared to pen-checking.