Table 5. Predictors of confirmed reports at the city block level and predictors of observed infestation by Triatoma infestans bugs among infested households during surveillance phase of vector control campaign in urban districts of Arequipa, Peru, 2009–2012*.

Confirmed reports at city blocks level					Infestation among inspected households
Predictor	OR	2.5% CI	97.5% CI		Predictor	OR	2.5% CI	97.5% CI
Intercept	8.2 × 10–5†	4.5 × 10–6	1.5 × 10–3		Intercept	0.08†	0.05	0.12
Log(no. infested + 1)	2.08†	1.42	3.04		Infested	2.24‡	1.28	3.91
Time, y	1.69§	1.10	2.58		–	–	–	–
Log(no. untreated + 1)	4.85§	1.34	17.49		Untreated	5.88†	3.18	10.88
Time × log(no. untreated + 1)	0.82¶	0.66	1.02		–	–	–	–

*For confirmed reports at city blocks level, outcome is the existence of ≥1 confirmed report on the city block: reports leading to observing ≥1 infested household on the city block in subsequent inspections. Infested is number of households observed to be infested at least once during the treatment phase. Untreated is number of households not treated in the treatment phase in the city block. Time corresponds to number of years as of 2012 since the end of the initial treatment phase for the locality. A locality level random effect is included for the city block level analysis. OR, odds ratio. N = 3,727 city blocks (80 with confirmed reports) Nagelkerke pseudo-R² = 0.71. For infestation among inspected households, outcome is the infestation status during the first inspection during the surveillance phase. Infested is households infested in ≥1 inspection of the treatment phase. Untreated is households not treated during the treatment phase. We used a random effect on inspections batches around a reporting house. –, because all houses in a batch are inspected at the same time, time is not included as a predictor. CIs are + values and assume asymptotic normality. n = 613 inspected (102 infested). Nagelkerke pseudo-R² = 0.91.
†p<0.001. 
‡p<0.01. 
§p<0.05. 
¶p<0.1.