Interdependence of welfare outcome measures and potential confounding factors on finishing pig farms

Abstract

The factors influencing the observed prevalence of some welfare outcome measures were investigated as part of a larger project examining the feasibility and benefits of incorporating on-farm welfare outcome assessments into UK Pig Farm Assurance Schemes. All finishing pigs (>40 kg) on six farms that had pens containing 25 pigs or less were assessed for dirtiness, body lesions, tail lesions, bursitis, lameness, oral behaviour and pigs requiring hospitalisation. The overall small number of significant correlations between the measures across all farms (29 out of a possible 120) and the lack of a significant correlation between any pair of measures on all six or even five farms suggests that no measure can be reliably replaced by another. On the three farms with the highest farm prevalence of dirtiness there were significant negative correlations between pen prevalences of dirtiness and recorded tail and/or body lesions ( p < 0.02). The prevalence of pigs requiring hospitalisation was positively correlated with lameness on four farms ( p < 0.001) and tail lesions on three farms ( p < 0.005). Using a univariate general linear model the time of observation was only found to correlate with one measure, the dirtiness of the pigs ( p = 0.004), with the prevalence of dirtiness in pens on the three ‘dirtiest’ farms increasing as the day progressed. There was no effect of the stocking density, group size, time since feeding and time since straw provision on the prevalence of any of the outcome measures. It is recommended that to prevent bias in the recording of tail and body lesions farm assurance assessments are conducted on pigs with a prevalence for dirtiness of less than 17%. Lower prevalences of dirtiness are more likely to occur on some farms first thing in the morning.