Assessing the effect of indoor residual spraying (IRS) on malaria morbidity in Northern Uganda: a before and after study

Table 3 Regression results showing percentage point change in total SPR adjusted for variations at district level

Timing in relation to IRS	Percentage point change	p value	95% confidence interval
Timing in relation to IRS	Percentage point change	p value	Lower boundary	Upper boundary
1 month after IRS	−4.18	0.157	−9.98	1.63
2 month after IRS	−6.48	0.037	−12.55	−0.41
3 month after IRS	−7.11	0.038	−13.80	−0.41
4 month after IRS	−0.39	0.912	−7.27	6.50
5 month after IRS	−2.97	0.459	−10.87	4.93
6 month after IRS	8.37	0.510	−16.66	33.39

Two linear fixed effects regression models were regressed on the SPR as the outcome variable both at district and hospital level. The models are shown below
Hospital model SPR = β₀ + β₁ (time) + β₂ (months past after spraying) + β₃ (Hospital)
District model SPR = β₀ + β₁ (time) + β₂ (months past after spraying) + β₃ (District)
The regression results adjusted for variations at district level and the adjusted confidence intervals of the percentage changes in the SPR 6 months after IRS reveal the same results with a decrease in the SPR 1–3 months after IRS which wanes out in the fourth month following IRS. The same results are obtained when the SPR is analysed by age category as shown in Table 6. The SPR increases by the sixth month when compared to the spray month, the reference month = zero

ISSN: 1475-2875