Table 5.

Final multiple logistic regression model for factors associated with lameness.

			Lame
Predictor	Category	Parameter estimate	Odds ratio	Confidence interval (95%)	P-value
Intercept		−1.82	0.16***	0.06–0.43	< 0.001
BCS	Optimal	Reference	–	–	–
	Overconditioned	−0.14	0.87	0.48–1.59	0.656
	Underconditioned	0.46	1.59*	1.10–2.30	0.014
Parity	1	Reference	–	–	–
	2	0.0021	1.00	0.62–1.61	0.993
	≥ 3	1.00	2.71***	1.83–4.01	< 0.001
Milk yield	Continuousa	−0.02	0.98*	0.96–1.00	0.05
Leg cleanliness	Little/no manure	Reference	–	–	–
	Minor splashing	0.03	1.03	0.71–1.50	0868
	Distinct plaques of manure	0.47	1.61*	1.00–2.61	0.05
	Solid plaque of manure	0.94	1.30	0.66–2.57	0.443
Hocks	No observable skin change	Reference	–	–	–
	Hairless patch	0.26	1.30	0.76–2.20	0.338
	Swelling and/or wound	0.94	2.56**	1.43–4.61	0.002
Rib swelling	No	Reference	–	–	–
	yes	1.03	2.81***	1.70–4.64	< 0.001
Farming type	Conventional farming	Reference	–	–	–
	Organic farming	−0.46	0.63	0.35–1.14	0.125
Length of stalls	Long	Reference	–	–	–
	Medium	0.56	1.76*	1.07–2.87	0.025
	Short	1.15	3.17***	1.93–5.19	< 0.001
Farm size	Small Medium Large	Reference0.290.55	– 1.34 1.72	–0.87–2.081.15–2.58	– 0.189 0.008
Year	2016 2017 2018 2019	Reference−0.89−0.85−0.84	– 0.41 0.43 0.43	–0.20–0.820.22–0.840.19–0.96	– 0.012 0.014 0.040

Out of the initial 22 predictors, 10 factors associated with housing conditions and the individual animal were maintained within the final model. The model incorporated data from 1,006 dairy cows on 56 farms.

^a1 unit increase.

^*p < 0.05,

^**p < 0.01, and

^***p < 0.001.