Table 4. Determinants of population density of Triatoma infestans in rooms of human dwellings in a periurban community of Arequipa, Peru*.

Risk factor		Mean no. triatomines captured (range)	Univariate zero-truncated negative binomial regression		Multivariate zero-truncated negative binomial regression
			Ratio (95% CI)	p value	Adjusted ratio (95% CI)	p value
Host
	Each person sleeping in room†		1.58 (1.29–1.93)	<0.001	1.42 (1.16–1.72)	0.001
	Animal sleeping in room	22.12 (1–124)	7.34 (2.20–24.54)	0.001	5.23 (1.56–17.47)	0.007
Material
	Fully stuccoed walls and ceiling	3.03 (1–20)	0.065 (0.03–0.16)	<0.001	0.11 (0.04–0.31)	<0.001
	Adobe	1.73 (1–4)	0.055 (0.013–0.24)	<0.001	0.15 (0.032–0.73)	0.019
	Mortared brick, sillar, or basalt	12.62 (1–124)	2.25 (1.08–4.71)	0.03	0.80 (0.35–1.84)	0.599‡
	Unmortared brick, sillar, or basalt§	11.85 (1–103)	1.58 (0.65–3.87)	0.313
Insecticide use§		9.74 (1–124)	1.20 (0.32–2.44)	0.808

*Triatomine density was estimated by timed-capture of insects by trained collectors. CI, confidence interval.
†No mean is given for continuous variables.
‡The variable describing mortared brick, sillar, and basalt was left in the final model because it represents a possible intervention and was therefore a principal variable of interest. Removing the variable does not greatly affect the estimates for the other parameters or the significance level of those estimates.
§Dropped from the multivariate model.