Skip to main content

Table 3 Variables associated with PWD in over 60% of all training-test data splits using elastic-net (n = 22/69)

From: Risk factors associated with post-weaning diarrhoea in Austrian piglet-producing farms

Variable

5% Percentile

Median

95% Percentile

% Non zero

Reduced chance of PWD

Treatment

–1.40

–1.07

–0.71

100

Yes

Neonatal diarrhoea

0.15

0.43

0.73

100

No

All in-all out in nursery units

–0.82

–0.51

–0.19

99

Yes

Probiotics

–0.99

–0.62

–0.18

99

Yes

PDS

0.05

0.36

0.68

98

No

Horseradish

–1.03

–0.66

–0.24

97

Yes

Number of feeding phases

–0.38

–0.20

–0.03

96

More phases

Solid floors

0.02

0.35

0.65

95

No

Commercial peat

–0.94

–0.54

0.00

94

Yes

Rodent burden

0.00

0.22

0.51

90

Low

Nurse sows

0.00

0.28

0.53

90

No

Plastic or metal floors

–0.60

–0.30

0.00

89

Yes

Dairy products

–0.74

–0.30

0.00

86

Yes

Looped water distribution system

–0.64

–0.23

0.00

82

Yes

Rock salt

–0.67

–0.21

0.00

82

Yes

Coke / phosphoric acid

0.00

0.28

0.61

82

No

Cross fostering

0.00

0.28

0.75

80

No

Regular cleaning and disinfection of nursery units

–0.44

–0.14

0.00

76

Yes

Automatic feeding

–0.55

–0.20

0.00

75

Yes

Yeast

–0.48

–0.16

0.00

73

Yes

Fibre in creep feed

–0.34

–0.07

0.00

69

Yes

Electrolytes, glucose

–0.63

–0.07

0.00

60

Yes

  1. % Non zero: percentage of splits with log odds (ratios) unequal zero; ZnO: zinc oxide; PDS: postpartum dysgalactia syndrome; Variables for which the 5 and 95 percentile of the logit distribution across 100 elastic-net splits did not include zero are marked in bold