During EAAP Congress 2023, the Rezoolution team presented the Bayesian model used for the development of an APP to detect sows at risk for stillbirth and to improve farrowing monitoring in order to reduce the stillborn rate.

Summary of the presentation: On average, more than 60% of sows give birth to stillborn piglets in French farms. This is an important cause of piglet mortality, which continues to increase with the constant improvement of sow prolificacy. The objective of this study was to build a predictive model of stillborn rate.

This study was performed in two farrow-to-finish farms and one farrowing farm located in Brittany. In each farm, the number of total born (TB), born alive (BA), stillborn piglets (S), and the same data at the previous farrowing (TBn-1, BAn-1, and Sn-1), backfat thickness (BFT) just before farrowing and at weaning, and parity rank were added to our dataset. In total, 3,686 farrowing data points were recorded.

Bayesian networks, as an integrated modeling approach, were used for analyzing stillborn rates at farrowing using BayesiaLab® software. Our results propose a hybrid model to predict the stillborn percentage during farrowing. Three significant main risk factors were retained by the model: the parity rank (percentage of total mutual information: MI=64%), Sn-1 (MI=25%), and TBn-1 (MI=11%). Additionally, a fourth factor (BFT just before farrowing) was also retained for sows of parity five or more (MI=0.4%).

Practically, for example, in the best conditions (i.e., low litter rank, less than 8% of stillborn piglets, and a prolificacy lower than 14 piglets at the previous farrowing), our model predicted a stillborn rate almost halved, from 6.5% (mean risk of our dataset) to 3.5% for a sow at the next farrowing. Conversely, in worse conditions (i.e., older sows with a BFT <15 mm, more than 15% of stillborn piglets, and a prolificacy higher than 18 piglets at the previous farrowing), the risk would be multiplied by 2.5, from 6.5% to 15.7%.

To our knowledge, this is the first study updating the Blackwell grid published in 1987. Our results highlight the impact of previous prolificacy and stillborn rate on the probability of stillborn piglets. Moreover, the importance of backfat thickness, especially in older sows, is to be considered. These promising results would allow farmers to classify sows according to their risk of giving birth to stillborn piglets and to manage them accordingly.→ Download the presentation