Data-based and welfare-oriented reproductive performance assessment for application in commercial pig farms

Abstract

The objective of the present study was to simplify the previously developed milk score (MS) to improve its on-farm practicability in the assessment of sows’ reproductive performance. To enhance the applicability of the MS, a linear regression model was constructed, streamlining the farmer's workload by eliminating the need for second and individual piglet weighing. The model incorporates the sow's parity number as a parameter to account for parity-specific performance differences. The simplified model demonstrates a predictive accuracy of 81.66% in assessing sow performance. In contrast to previously developed sow replacement models, the adjusted MS prioritizes smaller litter sizes, lower piglet mortality, and higher piglet weights. This adjustment aids in mitigating adverse animal welfare effects associated with larger litter sizes. The recommendation is made that older parity sows should not be replaced solely based on fewer piglets born alive, as they can still achieve acceptable reproductive performance. The higher piglet birth weights observed in older parity sows (≥10) contribute to reducing the negative animal welfare effects associated with larger litter sizes. For sow replacement decisions, the farmer should replace the sows with the lowest model-based performance assessment, which are also expected to perform low again in the subsequent parity. To this end, a dataset comprising 574 litters from sows with at least two consecutive parities was utilized, revealing that 61.50% of sows show a similar performance in the subsequent parity. Furthermore, the loss of body weight in lactating sows emerges as an additional parameter for assessing performance, with high-performing sows exhibiting the most significant relative weight loss during lactation (P < 0.05). The adjusted MS places emphasis on a data-driven and welfare-oriented approach to performance assessment, presenting a novel perspective. The simplified model holds promise for further development into a management tool applicable in practical pig farms.