Swedish Deep Bedded Sow Pool System
January 1, 2007
John J. McGlone, PhD
Professor
The Swedish Deep-Bedded Sow Pool System is a system that has sows on deep bedding, in a non-heated, open-air barn, with individual feeding stalls, and with sows transported when pregnant to a contract farrowing farm. After weaning, sows are transported back to the main farm to be re-bred.
I observed a system like this in 2004 in Sweden. My assessment of the situation was that the sows experienced very poor welfare. Sows would have been much better off (in my view) if they were in crates. This is not to say that crates are best for sows, just that this system was not good in terms of sow welfare.
My observations caused me to write a commentary to the JAVMA. Below is a collection of supporting materials.
Photos of Swedish Deep Bedded Sow Pool System (2004)
A good picture as might be used to describe the system
The wounds
ranged from mild to serious
Sows at the same farm showing a version of oral-nasal-facial behaviors called Stereotypies (believed by some to be a sign of poor welfare).
Sample Size Calculations for Auditing
|
In a herd of 1,000 sows, the sample size needed to detect a "problem" given the rate of that problem. Based on statistical sampling principles (Sawyer et al., 1996) |
|
| Rate of 'problem" |
Sample size needed for 95% confidence in detection |
| 0.1 | 950 |
| 0.2 | 800 |
| 0.3 | 650 |
| 0.4 | 550 |
| 0.5 | 500 |
| 1.0 | 275 |
| 1.5 | 200 |
| 2.0 | 150 |
| 2.5 | 125 |
| 3.0 | 100 |
| 4.0 | 75 |
| 5.0 | 60 |
| 7.5 | 40 |
| 10 | 30 |
| 20 | 15 |
| 30 | 10 |
| 50 | 5 |
Note that as the rate of the problem increases, the sample size needed to estimate the mean "problem" declines. Very few animals are needed to estimate the rate of sow wounds (for example) if the rate of wounding is high. At 20% of the sows wounded, the auditor needs to examine only 15/1,000 sows. In practice, 100 sows is the minimum, just to be sure the sample is accurate.