Household and study participants
The total number of households which were selected to participate in the baseline survey was 2880. Of these 2270 (78.8%) consented to participate, 276 (9.6%) failed to meet the inclusion criteria (no children <15 years), 13 (0.5%) refused and 11.1% were excluded for other reasons (Fig. 1). Of the 6,918 children aged between 0.5 and 14 years, 4388 (63.4%) were selected for inclusion and of these 3871 participated in the clinical survey, 10 children had missing RDT results and 517 (11.8%) children did not turn up for testing. Therefore, of the total children selected 3,861 (88%) were included in the analysis (Fig. 1).
Almost all nets which were used were LLINs (96.8%). The proportion of individuals who reported using a net the night before was 27.1%. The proportion of nets declared to be less than 2 years old was 37.1%, those 2-3 years old was 31.3% and those over 3 years old were 31.4%. Many of the nets had been received via clinic-based distribution during the period since the last UCC. Only 26% of individuals slept under nets of good condition, 35% slept under nets of moderate condition and 38% under nets which were badly torn. Net access was very low; only 16.9% of individuals were residing in households with enough nets to cover all sleeping places. About 25.9% of the households reported to be sprayed with insecticide in the year before (Table 2).
Table 2: Factors associated with malaria infection in the univariate and multivariate analysis
|
|
Number of Children
|
% with malaria infection
|
Univariate
|
|
Multivariate
|
|
|
|
OR
|
96%CI
|
p-value
|
|
OR
|
96%CI
|
p-value
|
|
Altitude (metres)
|
|
|
|
|
|
|
|
|
|
|
|
Low (1128 – 1264)
|
1,300
|
74.7%
|
1
|
|
|
|
|
|
|
|
|
Medium (1265 -1352)
|
1,291
|
68.1%
|
0.38
|
0.28–0.50
|
0.001
|
|
0.78
|
0.42–1.43
|
0.416
|
|
|
High (1353 - 1654)
|
1,266
|
51.0%
|
0.14
|
0.10–0.20
|
0.001
|
|
0.27
|
0.14–0.55
|
<0.001
|
|
Age
|
|
|
|
|
|
|
|
|
|
|
|
< 1 year
|
384
|
50.5%
|
1
|
|
|
|
|
|
|
|
|
2 to 4 years
|
1,062
|
63.0%
|
1.86
|
1.44–2.41
|
<0.001
|
2.59
|
1.31–5.11
|
0.006
|
|
|
|
5 to 14 years
|
2,414
|
67.7%
|
2.52
|
1.99–3.20
|
<0.001
|
3.82
|
1.89–7.72
|
|
<0.001
|
|
Improved Housing Structure
|
|
|
|
|
|
|
|
|
|
|
|
No
Yes
|
589
3,272
|
76.9%
62.5%
|
1
0.32
|
0.22–0.48
|
<0.001
|
|
1
0.27
|
0.13–0.54
|
<0.001
|
|
Socio economic status
|
|
|
|
|
|
|
|
|
|
|
|
|
Poorest
|
1,261
|
72.6%
|
1
|
|
|
|
1
|
|
|
|
|
Middle
|
1,264
|
64.3%
|
0.72
|
0.59–0.86
|
0.001
|
|
0.98
|
0.51– 1.89
|
0.961
|
|
|
Least poor
|
1,258
|
57.6%
|
0.50
|
0.42–0.60
|
<0.001
|
|
1.07
|
0.52– 2.19
|
0.862
|
|
Adequate access to LLINs
|
|
|
|
|
|
|
|
|
|
|
No
|
3,167
|
66.0%
|
1
|
|
|
|
1
|
|
|
|
Yes
|
643
|
58.8%
|
0.68
|
0.57–0.83
|
<0.001
|
|
0.57
|
0.34–0.98
|
0.042
|
|
Individual net age
|
|
|
|
|
|
|
|
|
|
|
< 2years
|
350
|
52.9%
|
1
|
|
|
|
1
|
|
|
|
2 -3 years
|
295
|
64.1%
|
2.32
|
1.26-4.28
|
0.007
|
|
1.89
|
1.01–3.53
|
0.044
|
|
4 years and more
|
299
|
65.9%
|
2.37
|
1.28-4.37
|
0.006
|
|
1.90
|
1.02–3.58
|
0.004
|
|
Sex
|
|
|
|
|
|
|
|
|
|
|
Male
|
1,869
|
66.0%
|
1
|
|
|
|
|
|
|
|
Female
|
1,870
|
63.4%
|
0.87
|
0.75–1.01
|
0.061
|
|
|
|
|
|
Eaves of the house
|
|
|
|
|
|
|
|
|
|
|
Open
Closed
|
2,334
1,478
|
69.2%
57.9%
|
1
0.59
|
0.51–0.69
|
<0.001
|
|
|
|
|
|
Livestock present
|
|
|
|
|
|
|
|
|
|
|
No
|
1,026
|
67.6%
|
1
|
|
|
|
|
|
|
|
Yes
|
2,786
|
63.7%
|
0.77
|
0.65–0.91
|
0.002
|
|
|
|
|
|
Animals kept inside the house
|
|
|
|
|
|
|
|
|
|
|
No
|
1,857
|
63.5%
|
1
|
|
|
|
|
|
|
|
Yes
|
935
|
63.6%
|
0.97
|
0.81–1.17
|
0.757
|
|
|
|
|
|
Head of household attended school
|
|
|
|
|
|
|
|
|
|
|
No
|
1,007
|
71.4%
|
1
|
|
|
|
|
|
|
|
Yes (any schooling)
|
2,791
|
62.4%
|
0.74
|
0.62–0 .88
|
0.001
|
|
|
|
|
|
IRS 18 months before
|
|
|
|
|
|
|
|
No
|
2,798
|
67.0%
|
1
|
|
|
|
|
Yes
|
979
|
59.1%
|
0.84
|
0.70– 1.01
|
0.06
|
|
Household owns at least 1 net
|
|
|
|
|
|
|
|
|
|
|
No
|
1,165
|
69.1%
|
1
|
|
|
|
|
|
|
|
Yes
|
2,647
|
62.6%
|
0.71
|
0.60–0.83
|
<0.001
|
|
|
|
|
|
Individual usage of LLIN
|
|
|
|
|
|
|
|
|
|
|
Don’t use net
Use Net
|
2,276
1,045
|
66.4%
61.2%
|
1
0.69
|
0.50–0.94
|
0.018
|
|
|
|
|
|
Individual LLIN Condition
|
|
|
|
|
|
|
|
|
|
|
|
Use good condition net
(PHI > 64)
Use moderate condition net (PHI 65 – 642)
Use torn net (PHI >643)
|
154
206
221
|
68.2%
61.7%
63.8%
|
1
0.65
0.77
|
0.28–1.49
0.34–1.77
|
0.307
0.544
|
|
|
|
|
|
Of the 1008 households selected for indoor entomological collection 770 participated, and of 144 selected for outdoor collection 120 participated (Fig. 1). Altitude of study households ranged from 1128 to 1654 metres above sea level. Around 76.8% of households were constructed with mud walls and floors, and 61.2% of houses had open eaves. Most households (94.3%) relied on farming or fishing and selling cash crops for income. Most individuals came from houses with livestock (73.1%) and 33.5% of these kept animals inside the household. 73.3% of heads of households had attended primary school (Table 3).
Table 3: Risk factors associated with vector density in the univariate and multivariate models
|
|
Number of HH
|
Indoor vector density
|
Univariate
|
|
Multivariate
|
|
|
IRR
|
96%CI
|
p-value
|
|
IRR
|
96%CI
|
p-value
|
Altitude (metres)
|
|
|
|
|
|
|
|
|
|
|
1133-1242
|
192
|
63.2
|
1
|
|
|
|
1
|
|
|
|
1243-1302
|
191
|
22.6
|
0.63
|
0.49-0.80
|
<0.001
|
|
0.58
|
0.46-0.74
|
<0.001
|
|
1303-1373
|
192
|
6.8
|
0.51
|
0.40-0.66
|
<0.001
|
|
0.44
|
0.33-0.58
|
<0.001
|
|
1374-1651
|
191
|
5.7
|
0.39
|
0.29-0.51
|
<0.001
|
|
0.35
|
0.27-0.46
|
<0.001
|
Wind during collection
|
|
|
|
|
|
|
|
|
|
|
None/light
|
433
|
28.2
|
1
|
|
|
|
1
|
|
|
|
Moderate/Strong
|
335
|
19.9
|
0.77
|
0.64-0.93
|
0.007
|
|
0.69
|
0.57-0.83
|
<0.001
|
Livestock
|
|
|
|
|
|
|
|
|
|
|
No
|
267
|
17.5
|
1
|
|
|
|
1
|
|
|
|
Yes
|
502
|
28.3
|
1.13
|
0.95-1.35
|
0.159
|
|
1.35
|
1.12-1.62
|
0.001
|
Improved housing structure
|
|
|
|
|
|
|
|
|
|
|
No
|
579
|
29.3
|
1
|
|
|
|
1
|
|
|
|
Yes
|
183
|
10.1
|
0.78
|
0.63-0.95
|
0.016
|
|
0.83
|
0.67-1.02
|
0.072
|
Household using at least 1 LLIN
|
|
|
|
|
|
|
|
|
|
|
No
|
264
|
22.9
|
1
|
|
|
|
1
|
|
|
|
Yes
|
505
|
25.4
|
0.79
|
0.66-0.93 0.006
|
|
0.78
|
0.66-0.92
|
0.004
|
Number of rooms
|
770
|
24.9
|
0.89
|
0.84-095
|
<0.001
|
|
0.90
|
0.84-0.96
|
0.001
|
|
|
|
|
|
|
|
|
|
|
|
Household cooks inside the house
|
|
|
|
|
|
|
|
|
|
No
|
620
|
23.3
|
1
|
|
|
|
|
|
|
|
Yes
|
149
|
29.6
|
1.31
|
1.07-1.59
|
0.010
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Total time house has been sprayed
|
756
|
24.9
|
0.94
|
0.90-0.99
|
0.011
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Animal inside the house
|
|
|
|
|
|
|
|
|
|
|
No
|
308
|
29.2
|
1
|
|
|
|
|
|
|
|
Yes
|
196
|
26.5
|
1.01
|
0.82-1.25 0.933
|
|
|
|
|
Rain during collection
|
|
|
|
|
|
|
|
|
|
|
None
|
435
|
23.7
|
1
|
|
|
|
|
|
|
|
Light
|
164
|
37.1
|
0.98
|
0.78-1.24
|
0.894
|
|
|
|
|
|
Heavy
|
169
|
14.5
|
1.15
|
0.92-1.44
|
0.234
|
|
|
|
|
Eaves of the house
|
|
|
|
|
|
|
|
|
|
|
Closed
|
324
|
22.9
|
1
|
|
|
|
|
|
|
|
Open
|
445
|
25.7
|
1.17
|
0.99-1.38
|
0.071
|
|
|
|
|
Plasmodium infection prevalence and risk factors for malaria infection
Overall malaria infection prevalence (all Plasmodium species) was 64.8 % (95% CI 61.8 – 67.8) in children aged 6 months to 14 years old. The prevalence of clinical malaria, defined as fever (ear temperature, ≥ 37.5°C) and any Plasmodium infection by RDT was 6.8% (95% CI 5.5 - 8.1). Only 2.9% (114) of tested children had fever and among those who had fever 86% had a positive test result. The prevalence of moderate or severe anaemia (haemoglobin <8 g/dL – see Table 1 for variables’ definitions) in children under five was 19.5% (95%CI 17.9 – 21.4). Among anaemic children 88.9% (95%CI 85.5 – 92.2) had malaria infection. The odds of moderate-severe anaemia was six times higher if the child had malaria infection (OR 6.2, 95% CI 5.0 - 7.6).
In the univariate analysis the odds of malaria infection was lower with living at higher altitudes, in children in younger age groups, with better education, improved house structure, closed eaves, greater wealth, presence of livestock, net ownership, adequate access to LLINs, use of LLINs and with newer LLINs (Table 2).
In the multivariate model, malaria infection was independently associated with altitude, age, quality of housing structure, and adequate net access per sleeping place.
The odds of malaria infection decreased with increasing altitude. Individuals living at high altitude had much lower odds of malaria infection compared to those who were living at lower altitude (OR 0.18 95%CI 0.13 – 0.26).
The odds of malaria infection increased with age, the oldest group being the most at risk (OR 3.82, 95%CI 1.89–7.72). Children living in improved housing structure had lower odds of malaria infection compared to those who were living in unimproved housing (OR 0.27, 95%CI 0.13 – 0.54). Individuals living in houses with an open eaves showed a strong association with malaria infection compared to those who were living in houses with closed eave gaps (OR 0.59; 95%CI 0.51 – 0.69).
Individuals who were living in households with adequate net access per sleeping place were better protected from malaria infection (59% vs 66%, OR 0.57; 95% CI 0.34 – 0.98, P=0.042). Age of LLINs was also significant, with only LLINs less than two years of age showing evidence of protection (53% vs 65%, P=0.004). LLINs’ usage was not associated with malaria prevalence in the multivariate analysis. There was no association between malaria protection and the hole index of LLINs over the WHO LLIN categories of good condition (32% infection prevalence, 49/154) to ‘unserviceable’ condition (36%, 80/221).
None of the other factors (gender, wealth, livestock, household head education, IRS spraying 18 months earlier, household owning at least one LLIN, and usage of LLINs of different physical conditions) were associated with malaria infection once adjusted for the other factors in the model.
Mosquito fauna and risk factors for vector density
A total of 29,401 and 2,668 mosquitoes were collected in 770 indoor light trap collections and in 120 outdoor tent trap collections respectively. Anophelines comprised 64% of the indoor collections and 57% of the outdoor collections. Mean vector density per collection night was 24.5 indoors and 12.7 outdoors. Of the Anophelines collected 4.6% were identified as An. funestus s.l. and the remaining were An. gambiae s.l.. The 969 An. gambiae s.l. were identified to species by PCR, 93.5% were An. gambiae s.s., 4.5% were An. arabiensis, and 2.2% did not amplify. There was no evidence of difference in ratio of An. gambiae s.s. to An. arabiensis between indoor and outdoor collections (X2 3.5, p=0.18). Of the 289 An. funestus identified by PCR 81.7% were An. funestus s.s., 15.6% Anopheles leesoni, 1.0% Anopheles parensis and 1.7% did not amplify. A subsample of 4,311 specimens was tested for CSP. The sporozoite rate was 4.7%. Both An. funestus and An. gambiae s.s. were found positive for CSP. The entomological inoculation rate was 0.34 (95%CI: 0.20-0.49) infectious bites per person-night during the period of collection, and this rate varied from 0 to 2.7 between village clusters.
Altitude of households selected during the entomology survey ranged from 1,133 to 1,651 metres, with a mean of 1,303 metres. It rained during 43.4% of collections and the wind was moderate or strong during 43.6% of collections. 34.7% of households did not have animals. For those that owned animals, 39% kept them indoor. 50% of families had poorly constructed houses and 24% had improved houses. 57.9% of houses had open eaves. The median number of rooms was 5 per house and median household size was 5 persons. Only 19.4% of families cooked inside the house. Houses had been sprayed 3 times since the first IRS campaign in 2007. An average of 65.7% of households were using at least one LLINs.
The univariate analyses showed evidence for negative associations between vector density and improved housing, number of rooms, households using at least one LLINs and number of times the house had been sprayed. Cooking inside the house was associated with higher vector densities. Altitude above 1242 metres was associated with a decreased risk of indoor vectors, whereas light wind or no wind during collection was associated with increased risk.
House structure had an IRR of 0.75 (95% CI 0.60-0.90), indicating a 25% decreased risk of vector exposure in improved housing. Using at least one LLIN in the household and several times sprayed was associated with reduced vector density. There was no evidence of association with rain during collection night, open eaves or animals present inside the house.
The multivariate model retained five risk factors as significant: altitude, wind during collection, presence of livestock, house construction and LLINs’ usage. Number of rooms, showed co-linearity with house construction and, was therefore removed from the model. There was strong association with vector density, altitude and wind. Vector density was negatively associated with increasing altitude, being 65% less at the higher altitudes (above 1353 metres) and with nights with moderate or strong winds. Improved housing structure was associated with reduced vector density whereas presence of livestock was associated with higher vector density. Net usage was associated with density after adjusting for other factors. The variables that showed no association with vector density in the multivariate model were: history of IRS, cooking indoors and presence of open eaves.