Table 4.
Univariable analysis | Multivariable analysisc | |||||
---|---|---|---|---|---|---|
Variable | N (n)a | Random effectsb | Crude OR (95 % CI) | P value | Adjusted OR (95 % CI) | P value |
Baseline parasitaemia (2-fold rise) | 27,496 (1,853) | 2.31 | 1.30 (1.26–1.34) | <0.001 | 1.27 (1.24–1.31) | <0.001 |
Baseline anaemia | ||||||
Non-anaemic (reference)d | 8,838 (544) | 2.14 | 1 | - | - | - |
Moderate | 9,652 (714) | 1.07 (0.94–1.22) | 0.274 | 1.07 (0.94–1.22) | 0.289 | |
Severe | 1,668 (124) | 1.24 (0.99–1.55) | 0.056 | 1.33 (1.06–1.67) | 0.014 | |
Unknown | 7,338 (471) | - | - | - | - | |
Gametocytes presence | ||||||
No (reference) | 18,672 (1,358) | 2.08 | 1 | - | - | - |
Yes | 1,979 (102) | 0.95 (0.74–1.2) | 0.650 | - | - | |
Febrile on presentation (temperature >37.5 °C) | ||||||
No (reference) | 9,355 (433) | 2.06 | 1 | - | - | - |
Yes | 17,217 (1,412) | 1.72 (1.52–1.95) | <0.001 | 1.46 (1.28–1.66) | <0.001 | |
Gendere | ||||||
Female (reference) | 12,873 (835) | 2.22 | 1 | - | - | - |
Male | 13,995 (982) | 1.11 (1.00–1.23) | 0.052 | - | - | |
Age category | ||||||
≥12 years (reference) | 4,245 (202) | 2.22 | 1 | - | - | - |
<1 year | 2,014 (139) | 1.89 (1.40–2.57) | <0.001 | 1.49 (1.09–2.05) | 0.013 | |
1 to <5 years | 15,677 (1,176) | 1.94 (1.52–2.46) | <0.001 | 1.54 (1.21–1.97) | 0.001 | |
5 to <12 years | 5,528 (334) | 1.49 (1.20–1.85) | <0.001 | 1.25 (1.00–1.56) | 0.048 | |
Transmission settings | ||||||
High (reference) | 10,368 (455) | 2.12 | 1 | - | - | - |
Low/moderate | 17,128 (1,398) | 1.50 (0.88–2.55) | 0.135 | 1.88 (1.09–3.24) | 0.024 | |
Treatmentf | ||||||
DP (reference) | 4,420 (340) | 2.12 | 1 | - | - | - |
AL | 12,255 (729) | 1.19 (1.00–1.42) | 0.050 | 1.21 (1.01–1.44) | 0.040 | |
ASAQ-FDC | 4,997 (246) | 0.94 (0.75–1.19) | 0.619 | 0.90 (0.71–1.14) | 0.388 | |
ASAQ-coblistered NFDC | 1,574 (167) | 1.80 (0.84–3.85) | 0.130 | 1.87 (0.86–4.04) | 0.113 | |
ASAQ-loose NFDC | 4,250 (371) | 1.62 (1.18–2.22) | 0.003 | 1.46 (1.05–2.01) | 0.022 |
aN, number of patients with non-missing data; n, number of patients with positive blood smear on day 2; bvariance of the random effects for the univariable analyses; cN = 26,544 for the final multivariable model with 1,843 cases of positive parasitaemia. Likelihood ratio test for random effect (P <0.001). Variance of random effect = 2.05. Proportion of total variance contributed by the site-level variance component (ρ) = 0.38. Coefficient (standard error) of intercept = −7.95 (0.3539). The coefficient of variation in parameter estimates was calculated by excluding one study site at a time and expressed as relative standard deviation (RSD). Distributions of the adjusted odds ratio (AOR) were generated from 250 bootstrap samples. The RSD and bootstrap distribution are shown in Additional file 4: Table S8 and Figure S3); dmultiple imputation was performed on missing anaemia status using ordinal logistic regression with age, gender and parasitaemia as covariates. The estimates derived using 100 imputations for moderate and severe anaemia are: AOR = 1.05 (95 % CI: 0.93–1.19), P = 0.446; and AOR = 1.24 (95 % CI: 0.99–1.55), P = 0.056, respectively; egender (AOR = 1.10 (95 % CI: 0.99–1.22), P = 0.079 using likelihood ratio test) was no longer significant in the presence of the other variables shown in the multivariable model and hence dropped; ffor AL compared to ASAQ-FDC (AOR = 1.33 (95 % CI: 1.08–1.63, P = 0.005). For ASAQ-loose NFDC compared to ASAQ-FDC (AOR = 1.61 (95 % CI: 1.14–2.29), P = 0.007). AL, artemether-lumefantrine; AS-AQ, artesunate-amodiaquine; ASAQ-coblistered NFDC, non-fixed dose combination in a co-blister formulation; ASAQ-FDC, fixed dose combination; ASAQ-loose NFDC, non-fixed dose combination in a loose formulation; DP, dihydroartemisinin-piperaquine