Table 2.

Tabulation of a select set of key predictors by bovine infection status.

	2007			2010			2016
	n positive	Total tested	% Positive	n positive	Total tested	% positive	n positive	Total tested	% positive
Total	63	473	13.32	34	197	17.26	14	71	19.72
Bovine age b
<2	3	43	6.98	6	35	17.14	NDa	NDa
2 to 4	26	220	11.82	9	59	15.25	NDa	NDa
5+	28	137	20.44	18	93	19.35	NDa	NDa
Missing	6	73	8.22	1	10	10.00
Prevalence of human infection in village
0–2.53%	14	147	9.52	0	76	0.00	4	27	14.81
2.54– 11.11%	13	158	8.23	11	57	19.30	8	29	27.59
≥11.12%	36	168	21.43	23	64	35.94	2	15	13.33
Number of infected human household members
0	38	366	10.38	19	154	12.34	12	61	19.67
1+	19	84	22.62	13	29	44.83	2	10	20.00
Missing	6	23	26.09	2	14	14.29	0	0	--
Household has improved sanitation
No	47	365	12.88	27	129	20.93	8	34	23.53
Yes	16	108	14.81	7	68	10.29	6	37	16.22
Village mean asset score (0–9)
<2	39	225	17.33	0	0	0.00	0	0	0.00
2–3	24	248	9.68	26	143	18.18	2	23	8.70
≥4	0	0	0.00	8	54	14.81	12	48	25.00
Household pig ownership c
0	NDa	NDa		27	116	23.28	14	64c	21.88
1+	NDa	NDa		7	81	8.64	0	7	0.00
Mean number of bovines per household
< 0.5	5	116	4.31	4	59	6.78	10	57	17.54
0.5 – 1	40	300	13.33	19	99	19.19	4	14	28.57
≥ 1	18	57	31.58	11	39	28.21	0	0	0.00
Household bovine ownership
0	49	334	14.67	20	126	15.87	8	38	21.05
1+	14	139	10.07	14	71	19.72	6	33	18.18
Mean area of rice planted per household (mu)
<0.75	14	143	9.79	9	90	10.00	7	18	38.89
0.75–1.5	19	183	10.38	12	48	25.00	5	41	12.20
≥1.5	30	147	20.41	13	59	22.03	2	12	16.67
Area of rice planted by household (mu)
0	16	146	10.96	5	56	8.93	5	13	38.46
<2	24	195	12.31	16	84	19.05	4	31	12.90
2+	23	132	17.42	13	57	22.81	5	27	18.52
Mean area of dry summer crops planted per household (mu)
<1	11	184	5.98	1	14	7.14	0	3	0.00
1 –2.5	44	195	22.56	23	110	20.91	2	26	7.69
≥2.5	8	94	8.51	10	73	13.70	12	42	28.57
Area of dry summer crops planted by household (mu)
0	8	108	7.41	0	4	0.00	0	0	--
<3	39	242	16.12	18	99	18.18	2	26	7.69
3+	16	123	13.01	16	94	17.02	12	45	26.67
4+	16	108	14.81	7	48	14.58	0	13	0.00
Mean number of buckets of night soil used on rice crops per household
<1	3	58	5.17	29	146	19.86	8	41	19.51
1–9.9	21	153	13.73	4	38	10.53	6	27	22.22
≥10	39	262	14.89	1	13	7.69	0	3	0.00
Number of buckets of night soil used on rice crops by household
0	40	304	13.16	29	168	17.26	8	53	15.09
1+	22	168	13.10	5	29	17.24	1	7	14.29
Missing	1	1	100.0	0	0	--	5	11	45.45
Mean buckets of night soil used on dry summer crops per household
<10	13	190	6.84	24	99	24.24	1	15	6.67
10–30	18	109	16.51	7	47	14.89	13	47	27.66
≥30	32	174	18.39	3	51	5.88	0	9	0.00
Number of buckets of night soil used on dry summer crops by household
0	31	248	12.50	24	105	22.86	6	36	16.67
1–25	6	53	11.32	5	32	15.63	5	22	22.73
>25	26	172	15.12	5	60	8.33	3	13	23.08
Mean number of buckets of night soil used on winter crops per household
<10	1	26	3.85	24	121	19.83	14	71	100.0
10–30	19	218	8.72	7	41	17.07	0	0	--
≥30	43	229	18.78	3	35	8.57	0	0	--
Number of buckets of night soil used on winter crops by household
0	27	197	13.71	31	137	22.63	9	62	14.52
1–26	5	79	6.33	0	16	0.00	1	3	33.33
>26	31	197	15.74	3	44	6.82	4	6	66.67

A table of all candidate predictors is provided in Supplementary Table S2.

^aNo data (ND). Data was not collected on bovine sex, age or type in 2016, or on pig ownership in 2007.

^bAlthough predictors are broken into categories to facilitate a comparison of the distributions across the study period, most were assessed as continuous variable in our Random Forests models. See Table 1 for details on variable formulation for the RF models.

^cVariables with >90% of observations taking on a single value were excluded from the RF assessment. This exclusion criteria applied twice in 2016: >90% of the included households reported owning at least one dog, and >90% of households owned zero pigs.