Table 3. ORs for livestock-associated MRSA in pigs and in humans with increasing use of antimicrobial drugs, the Netherlands, 2011–2013*.

Characteristic	ORs for a 2-fold increase in DDDA/Y
	Pooled pig samples					Farmers and family members
	No.†	OR‡ (95% CI)	p value	−2 log RSPL§		No.¶	OR# (95%CI)	p value	−2 log RSPL§
All farms	1,421	1.16 (1.02–1.33)	0.03**	6,937.5		626	1.22 (1.01–1.48)	0.04	3,196.9
Supply of gilts††
Open	867	1.11 (0.97–1.27)	0.12**	3,828.9		365	1.08 (0.85–1.38)	0.53	1,806.9
Closed	554	0.86 (0.69–1.33)	0.79	3,132.2		261	1.31 (0.94–1.81)	0.11	1,424.3
Production type
Farrowing	476	1.38 (1.03–1.86)	0.03**	2,399.2		158	1.28 (0.85–1.94)	0.24	784.3
Farrow-to-finish	954	1.11 (0.95–1.30)	0.18	4,621.4		468	1.19 (0.95–1.50)	0.13	2,439.8

*Farm antimicrobial use was defined as 1 unit increase in the log₂ DDDA/Y. Results from the random intercept generalized linear mixed models accounting for the repeated measurements design and adjusting for confounders. DDDA/Y indicates the number of days of antimicrobial use per year for an average animal on the farm. It was determined by dividing the total number of kilograms treatable with a single mass unit of the antimicrobial drug concerned, according to the package insert information, by the average number of animal kilograms on the farm. The denominator comprised sows and fatteners. DDDA/Y, defined daily dosages animal per year; MRSA, methicillin-resistant Staphylococcus aureus; OR, odd ratio; RSPL, residual pseudo-likelihood. Bold type indicates significance (p<0.05).
†Number of observations at all sampling times together (10 pooled pig samples per farm on 36 farms in 4 sampling times). Values are missing for 19 observations.
‡For analysis in pigs, a farm random intercept was included in the mixed models and adjustment of ORs was made for sampling time and age group of pigs in the pool.
§RSPL from the generalized linear mixed models. Models per stratum of external supply or type of production are not nested and -2 log RSPL cannot be used for comparison.
¶Number of observations in all sampling times together (158 persons, 4 sampling times). Values are missing for 6 observations.
#For analysis in humans, a farm and a person random intercept were included in the mixed models, and number of hours worked on the farm and sampling time were used for adjustment of ORs.
**These models additionally showed significant antimicrobial use–time interaction indicating parallel change in antimicrobial use and livestock-associated MRSA prevalence over the study period (see extended explanation in text).
††Farms were defined as open when they received external supplies of gilts ≥1 time per year from at least 1 supplier and as closed when they received no external supply of gilts.