Proximity to vector breeding site and risk of Plasmodium vivax infection: a prospective cohort study in rural Ethiopia

Table 3 Fully adjusted Cox regression models with follow-up period split at net distribution campaign

	Period 1—week 1 to 48 (n = 7016)		Period 2—week 49 to 101 (n = 6695)
	RR (95% CI)	P value	RR (95% CI)	P value
Increasing household proximity to breeding site^a	1.60 (1.29–1.98)	< 0.001	1.46 (1.17–1.81)	0.001
Age at study entry
≥ 25	Ref		Ref
15–24	2.73 (0.88–8.47)	0.08	1.23 (0.50–3.00)	0.65
5–14	9.87 (3.79–25.70)	< 0.001	2.43 (1.15–5.13)	0.02
0–4	10.79 (4.03–28.92)	< 0.001	4.57 (2.13–9.79)	< 0.001
Gender
Males	Ref		Ref
Females	0.79 (0.49–1.27)	0.33	0.96 (0.60–1.53)	0.85
Persons living in household
≤ 7	Ref		Ref
≥ 8	0.83 (0.52–1.35)	0.46	1.81 (1.11–2.97)	0.02
ITN use fraction (%)^b
> 50	Ref		Ref
≤ 50	0.67 (0.37–1.20)	0.18	1.68 (1.01–2.79)	0.04

^aDistance to breeding site categorized into: > 2700 m (ref); 2400–2700; 2100–2400; and < 2100 m, and modelled linearly
^bITN use was recorded weekly (starting from week 5) during the regular follow-up visits to households. Participants were asked if they slept under a ITN the night before the visit. Use fraction was calculated as the number of nights reportedly sleeping under a ITN divided by the total number of nights asked. During week 48, there was a net distribution campaign in the Kebele thus the follow-up period was split at this point. For those with P. vivax infection, ITN use fraction was calculated only for weeks prior to infection