EVALUATION OF PRODUCER-RECORDED CAUSES OF PREWEANING MORTALITY IN DANISH SOW HERDS

The validity of producer-recorded preweaning mortality causes was eval uated in 31 Danish sow herds. For 1206 piglets producer-recorded prewe aning mortality cause was compared with post-mortem diagnosis. The pre dominant preweaning mortality causes at post-mortem were injury (traum a, lain on, and savaged), miscellaneous (unknown, starvation, illthrif t, euthanasia, and low viability), and diarrhoea, with 391, 219, and 1 19 recordings, respectively. The sensitivity of producer-recorded prew eaning mortality causes was generally low, and further analyses showed that the probability of being born alive and the probability of being given 'a producer diagnosis other than unknown' were associated with the herd and that a correct producer-recorded mortality cause relative to the post-mortem diagnosis depended on the mortality cause. Rare mo rtality causes with primarily internal signs tended to be misclassifie d. The producers often gave 'a diagnosis other than unknown' to pigs w ith a weight at death under 0.75 kg and it was often correct. To pigs with a weight at death over 1.75 kg, they rarely gave 'a diagnosis oth er than unknown', but when they did it was often correct. (C) 1997 Els evier Science B.V.