. 2019 Dec 9;17:220. doi: 10.1186/s12916-019-1456-9

Table 3.

Key predictors of clinical malaria episodes due to P. falciparum and P. vivax in 2013

	P. falciparum					P. vivax
		Base model		_molFOB adjusted			Base model		_molFOB adjusted
	IR	IRR (CI₉₅)	p	IRR (CI₉₅)	p	IR	IRR (CI₉₅)	p	IRR (CI₉₅)	p
Areas of residence
Ilahita 1–4, 6, 7	0.05	Reference group				0.06	Reference group
Balanga & Balif	0.03	0.63 (0.13–3.14)	0.575	0.56 (0.11–2.81)	0.485	0.06	1.15 (0.32–4.06)	0.833	1.08 (0.30–3.91)	0.909
Kamanokor & Ilahita 5	0.25	4.30 (1.59–11.6)	0.004	3.96 (1.46–10.8)	0.007	0.52	8.01 (3.23–19.9)	< 0.001	3.86 (1.44–10.3)	0.007
Sunuhu 1&2	0.50	8.15 (3.40–19.6)	< 0.001	6.48 (2.65–15.8)	< 0.001	0.33	3.71 (1.53–8.99)	0.004	2.00 (0.77–5.17)	0.152
		p < 0.0001^a		p < 0.0001^a			p < 0.0001^a		p < 0.0234^a
Age		1.38 (1.10–1.73)	0.006	1.30 (1.03–1.64)	0.026
Haemoglobin		0.52 (0.35–0.77)	0.001	0.61 (0.40–0.92)	0.017		0.31 (0.19–0.48)	< 0.001	0.38 (0.24–0.59)	< 0.001
FOB^b				1.10 (1.02–1.18)	0.008				1.17 (1.09–1.25)	< 0.001

Multivariate negative binomial regression model-based estimates predicting risk of clinical P. falciparum and P. vivax. Backward selection approach was used to derive significant associations. Base models included all variables except _molFOB. Incidence is based on aggregated clinical data for entire 10-month study period thus precluding analysis of recent antimalarial treatment as a covariate. Bednet use was not analysed due to non-converge of data when included into models. ^aOverall significance level for the variable estimated using wald chi-square test; _molFOB: molecular force of blood-stage infections; ^b_molFOB was included as a rate; IR: Incidence rate; IRR: Incidence rate ratio. CI₉₅ 95% confidence interval; p p value