Bovine respiratory disease (BRD) remains the most significant disease complex in the feedlot industry worldwide. Approximately 16% of the incoming cattle population develops BRD during the feeding period. Commonly, trained feedlot personnel visually observe cattle in the home pen and decide on treatment if necessary. However, inaccurate and late diagnosis contributes to economic losses associated with this important disease. The overall aim of this thesis was to explore methods to increase accuracy and improve timing of BRD detection in feedlots. Research presented in this thesis focused on using automatic recording system for individual feeding behaviour, evaluated 2 alternative automated behaviour monitoring systems, identified the economic value of using automated behaviour recording systems and summarized current methods to improve accuracy of visual detection, timing of identification and prognostic methods for unfavourable BRD outcome. Chapter two demonstrated that continuous monitoring of feeding behaviour with individual feedbunks identified sick steers up to 7 days before visual signs of BRD appeared. Increased average meal-size, meal-time, time between meals and frequency of meals were associated with decreased BRD hazard. Two automatic recording systems for feeding but also other behaviour were evaluated in chapter 3 of this thesis. An ear-motion detector was highly sensitive and moderately specific in feeding time monitoring, but was highly specific with low sensitivity for rumination recording. Conversely, a leg-attached accelerometer accurately measured bunk attendance, bunk visit frequency as well as lying time. Chapter 4 demonstrated that the costs of individual feedbunks as used in chapter 2 did not outweigh the benefits of earlier and more efficient treatment. To be cost effective for BRD detection the per animal costs of the system would have to be < CAD 4, unless the true BRD incidence exceeded 47%. Finally, chapter 5 demonstrated that automated monitoring systems have been implemented successfully to detect BRD. Specifically, feeding behaviour and temperature monitoring could be used for early identification of BRD affected cattle.