Indoor adult Mosquito collections
Overall, 15,286 mosquitoes were collected over the two years using CDC light trap in 160 houses distributed over 16 study villages (mean = 4.35, SD = 11.48). Of all mosquitoes collected, 2,872 were anophelines (2,869 An. gambiae s.l., 1 An. funestus and 2 other anopheles), and 12,414 were culicines. Of the total 12,414 culicines collected in the two years, 20.59% were collected from screened houses compared to 79.41% collected from unscreened (control) houses respectively. There was overall low number of Anopheles mosquitoes collected indoor in experimental houses compared to control houses during the sampling period despite non-significant difference in the mean number of mosquitoes [Table 1].
Table 1: Number of indoor adult mosquitoes collected in Nyabondo for two years, separated by species and the study arm
Study arm
|
An. gambiae
N (mean; SD)
|
An. funestus
N (mean; SD)
|
Other anopheles
N (mean; SD)
|
Culicines
N (mean; SD)
|
Total
N (mean; SD)
|
Control
|
2,325 (0.82; SD=3.77)
|
1 (0; SD=0.02)
|
2 (0; SD=0.03)
|
9,858 (3.47; SD=9.11)
|
12,186 (4.29; SD=10.84)
|
Experimental
|
544 (0.80; SD=3.17)
|
0 (0; SD=0)
|
0 (0; SD=0)
|
2,556 (3.77; SD=12.85)
|
3,100 (4.57; SD=13.83)
|
Diff (t-test; p-value)
|
t=-0.108, p=0.914
|
t=-0.489, p=0.625
|
t=-0.691, p=0.489
|
t=0.698, p=0.486
|
t=0.567, p=0.571
|
Overall
|
2,869 (0.82; SD=3.66)
|
1 (0; SD=0.02)
|
2 (0; SD=0.02)
|
12,414 (3.53; SD=9.94)
|
15,286 (4.35; SD=11.48)
|
Diff: P-value for the statistical difference in the mean number of mosquito collections in the control and experimental groups was calculated using student t-test at 95% confidence interval.
|
A total of 14,952 mosquitoes were unfed (2,749 An. gambiae, 0 An. funestus, and 12,203 culicines), 311 mosquitoes were fed (106 An. gambiae, 0 An. funestus, and 205 culicines), 6 were half gravid (6 An. gambiae, 0 An. funestus, and 0 culicines), and 15 were gravid (8 An. gambiae, 1 An. funestus, and 6 culicines). Overall, among An. gambiae collected, 92.6% were non-blood fed, 3.57% were blood fed and the remaining 0.47% were composed of gravid and half gravid females. However, irrespective of the species, more unfed, fed, half gravid and gravid mosquitoes were collected in the control than experimental arms of the study [Table 2].
Table 2: Number of indoor adult mosquitoes collected in the study area, separated by species and physiological state
Mosquito species
|
Unfed
N (mean; SD)
|
Fed
N (mean; SD)
|
Half gravid
N (mean; SD)
|
Gravid
N (mean; SD)
|
Control
|
|
|
|
|
An. gambiae
|
2,208 (0.78; SD=3.13)
|
103 (0.04; SD=0.90)
|
6 (0; SD=0.11)
|
8 (0; SD=0.09)
|
An. funestus
|
0
|
0
|
0
|
1 (0; SD=0.02)
|
Other anopheles (unidentified)
|
0
|
0
|
0
|
0
|
Culicines
|
9,658 (3.40; SD=8.88)
|
197 (0.07; SD=1.01)
|
0
|
3 (0; SD=0.03)
|
Experimental
|
|
|
|
|
An. gambiae
|
541 (0.80; SD=3.15)
|
3 (0; SD=0.09)
|
0
|
0
|
An. funestus
|
0
|
0
|
0
|
0
|
Other anopheles
|
0
|
0
|
0
|
0
|
Culicines
|
2,545 (3.75; SD=12.82)
|
8 (0.01; SD=0.15)
|
0
|
3 (0; SD=0.09)
|
Overall
|
|
|
|
|
An. gambiae
|
2,749 (0.78; SD=3.14)
|
106 (0.03; SD=0.81)
|
6 (0; SD=0.10)
|
8 (0; SD=0.08)
|
An. funestus
|
0
|
0
|
0
|
1 (0; SD=0.02)
|
Other anopheles
|
0
|
0
|
0
|
0
|
Culicines
|
12,203 (3.47; SD=9.76)
|
205 (0.06; SD=0.91)
|
0
|
6 (0; SD=0.05)
|
Figure 3 gives the time series trend and comparison of the monthly collections of the anopheline mosquitoes. Baseline collections were done during the first nine months (i.e. January to September of 2017), and it cumulatively showed higher number of mosquitoes especially during the months of April and May. Further, the follow-up collection periods showed a gradual increase in the number of mosquitoes which peaked around the months of July and August of the following year, before sudden decline to very low levels. Importantly, at most months, the total number of mosquitoes in the experimental group were lower than those in the control group.
Figure 3: Time series trend and comparison of the monthly collections of the anopheline mosquitoes for two years.
Impact of house screening on malaria prevalence
Overall, malaria prevalence was 5.6% (95%CI: 4.2-7.5) n=1,918, with baseline prevalence rate of 6.1% (95%CI: 3.9-9.4), n=481 and 3rd follow-up survey prevalence of 3.6% (95%CI: 2.0-6.8) n=494. At all the three follow-up survey points, house screening significantly reduced the malaria prevalence by 100% (p<0.001), 63.6% (p=0.026), and 100% (p<0.001) in the 1st, 2nd and 3rd follow-ups respectively. Analysis of the prevalence when baseline survey was not taken into account is also given in table 3, and overall, there was no noticeable difference in results with and without baseline survey. When baseline survey was not taken into account, greater impact of house screening in reducing malaria prevalence was observed at 80% (p = 0.005) [Table 3].
Figure 4 compares the malaria prevalence in both control and experimental groups at different survey points. Overall, house screening significantly reduced malaria prevalence by 54% (OR = 0.46, 95%CI: 0.24-0.87, p = 0.017), there were also significant impacts observed at each of the three follow-up survey points.
Figure 4: Comparison of malaria prevalence in both control and experimental groups at different survey points
Table 3: House eaves screening impact on malaria parasite prevalence rate in the study area
Characteristic
|
Baseline
survey
|
1st follow-up
|
2nd follow-up
|
3rd follow-up
|
Overall
|
Demographics
|
No. of villages (No. of houses)
|
16 (160)
|
15 (150)
|
16 (160)
|
16 (160)
|
16 (160)
|
Household membership: median (range)
|
4 (1-8)
|
4 (1-8)
|
4 (1-8)
|
4 (1-8)
|
4 (1-8)
|
Age: mean (range); N
|
20.9 (0-90); 478
|
21.0 (0-91); 481
|
21.2 (0-91); 457
|
21.3 (0-91); 493
|
21.1 (0-91); 1909
|
Gender: No. infected / No. sampled (%)
|
Male
|
9/208 (4.3%)
|
11/208 (5.3%)
|
15/190 (7.9%)
|
7/206 (3.4%)
|
42/812 (5.12%)
|
|
Female
|
20/273 (7.3%)
|
11/275 (4.0%)
|
23/272 (8.5%)
|
11/288 (3.8%)
|
65/1108 (5.87%)
|
|
Malaria parasite Prevalence, % (95%CI)*
|
|
|
|
|
|
|
With baseline
|
Without baseline
|
Control
|
5.2 (3.4-8.1)
|
6.9 (5.0-9.4)
|
10.1 (6.6-15.3)
|
4.1 (2.2-7.9)
|
6.3 (4.8-8.3)
|
6.7 (4.9-9.2)
|
Experimental
|
10 (4.5-22.2)
|
0
|
3.7 (1.3-10.7)
|
0
|
3.0 (1.5-6.0)
|
1.4 (0.4-4.5)
|
Risk reduction, % (p-value)
|
increase
|
100% (p<0.001)
|
63.6% (p=0.026)
|
100% (p<0.001)
|
52.4% (p=0.019)
|
79.1% (p=0.006)
|
Total prevalence
|
6.1 (3.9-9.4)
|
4.6 (3.5-5.9)
|
8.2 (5.1-13.2)
|
3.6 (2.0-6.8)
|
5.6 (4.2-7.5)
|
5.4 (3.9-7.5)
|
House screening impact on malaria parasite prevalence, OR (95%CI), p-value@
|
|
|
|
|
|
|
With baseline
|
Without baseline
|
Control
|
Reference
|
|
|
|
|
|
Experimental
|
2.01 (0.88-4.57), p=0.096
|
0, p<0.001
|
0.35 (0.14-0.85), p=0.021
|
0, p<0.001
|
0.46 (0.24-0.87), p=0.017
|
0.20 (0.06-0.61), p=0.005
|
*Malaria parasite prevalence was calculated and 95% confidence intervals estimated using binomial logistic regression model while accounting for house clustering
@The impact of house screening on malaria prevalence was calculated and odds ratios (ORs) estimated using multilevel mixed effects logistic regression model while accounting for house clustering
|
Vector Densities and malaria cases in relation to rainfall pattern
Female Anopheles mosquitoes collected indoors per trap/night reduced significantly from 2017 to 2018. In 2017, there was relatively high numbers of captured indoor malaria vectors that corresponded with a similar increase in recorded average rainfall especially for the months of March-June. The same year and months also recorded a similar increase in recorded malaria cases in the nearby Nyabondo Mission Hospital compared to year 2018 [Figure 5]. The high numbers of recorded malaria cases at the hospital for the months of September-November 2017 was attributed to national wide strike of nurses and clinicians working in government health facilities in the country. It was observed when the local government health facilities were not functioning, more local people sought health services from the mission hospital which also doubled as a referral facility serving people from other surrounding sub-counties. Comparatively, the year 2018 recorded reduced amount of rainfall, less numbers of malaria vectors.
Figure 5: Female Anopheles mosquito densities/night/trap in relation to rainfall and malaria cases in the study area