Table 2. Results of the negative binomial generalized estimating equation (GEE) population-averaged models estimating the effect of net type on the number Phlebotomus argentipes captured inside the nets by CDC light traps, using cattle as bait.
| Female P. argentipes | Total P. argentipes | |||||
| Adjusted Percent Reduction (95% CI) | p-value | Expected Mean (SE) | Adjusted Percent Reduction (95% CI) | p-value | Expected Mean (SE) | |
| Trial 1 | ||||||
| Insecticide treated nets with 156 holes/inch2 (A) | Reference | 0.4 (0.20) | Reference | 1.55 (0.58) | ||
| Insecticide treated nets with 625 holes/inch2 (B) | 77% (56%–88%) | <0.0001 | 0.1 (0.00) | 78% (59%–88%) | <0.0001 | 0.34 (0.16) |
| Trial 2 | ||||||
| Untreated nets with 156 holes/inch2 (C) | Reference | 0.65 (0.23) | Reference | 1.76 (0.50) | ||
| Insecticide treated nets with 156 holes/inch2 (A) | 77% (27%–93%) | 0.0132 | 0.15 (0.07) | 61% (36%–76%) | 0.0001 | 0.69 (0.16) |
Estimates for net type used in each trial (Trial 1: A vs B and Trial 2: A vs C) were adjusted by night and sequence. The effect of nets is presented as efficacy in percent reduction and its 95% confidence interval (CI) for both female and total P. argentipes. The marginal estimates: mean count and Standard Error (SE) by net used are also presented.