Table 2.

(A) Unadjusted and (B) Multivariable logistic regression analysis (odds ratio OR (95% CI) P value) of environmental factors contributing at 5 % significance level to the presence of B. pseudomallei in soil at the four field sites in the dry and wet season. Italics in unadjusted models indicate non-significant odds ratios. The following environmental factors were included in the initial model: presence of introduced grass, presence of native spear grass such as Sorghum spp., soil pH, soil water status, soil colour, sampling round. Soil texture was consistent both within and between sites and was not included in the analysis (see text and Supplement Table 1). Road drainage was included as a confounder covariate. Controlling for road drainage, the Mantel-Haenszel estimate of the odds ratio of presence of B. pseudomallei associated with exotic grasses in the dry season was 4.7 (1.8–12.3) P<0.001. The regression analysis was clustered for sites and the final multivariable models with significant predictor variables were specified correctly as tested by a linktest and the fit of models was declared good (Pearson goodness-of-fit, P>0.36).

	Dry Season(n=213)		Wet Season(n=166)
	Unadjusted Model	Multivariable Model	Unadjusted Model	Multivariable Model
Exotic Grass	10.2 (2.8–37.5) P=0.001	11.7 (3.0–44.7) P<0.001	3.5 (1.5–8.0) P=0.004	1.8 (1.1–2.8) P=0.013
Moist to Wet Soil	5.8 (1.2–28.1) P=0.028	7.0 (2.2–22.6) P=0.001	1.2 (0.3–4.3) P=0.779
Road Drainage	8.3 (3.8–18.1) P<0.001		4.7 (1.6–13.8) P=0.005	3.3 (1.2–8.8) P=0.019
Soil pH	1.3 (0.9–2.0) P=0.128	1.5 (1.2–1.9) P=0.002	1.0 (0.7–1.4) P=0.838
Soil Color Reddish Grey	0.5 (0.1–3.6) P=0.531		1.8 (0.7–4.6) P=0.242	2.6 (1.1–5.8) P=0.025
Soil Color Reddish Yellow	2.3 (0.6–8.7) P=0.237		2.6 (1.3–5.0) P=0.005	1.3 (1.1–1.4) P<0.001