Table 4. Multivariable logistic regression modelling of the association between background factors and risk of false negative reporting (i.e. no report through the population-based, event-driven surveillance system when the reference method – one-week recall questionnaires – signals onset of disease) in 2008, n = 396.

		OR	95% CI
Validation week	1	1.00
	2	0.59	(0.25–1.43)
	3	0.70	(0.28–1.75)
	4	1.13	(0.45–2.82)
	5	1.01	(0.40–2.57)
	6	1.32	(0.56–3.10)
	7	1.21	(0.46–3.18)
Gender, and Time from invitation to registration	Women ≤2 weeks	0.95	(0.45–2.01)
	Women >2 weeks	1.00
	Men ≤2 weeks	1.53	(0.68–3.42)
	Men >2 weeks	6.37	(2.05–19.82)
Age group, years, and Mode of registration	0–14 Telephone	1.07	(0.36–3.19)
	15–39 Telephone	1.44	(0.46–4.52)
	40–64 Telephone	1.00
	≥65 Telephone	1.24	(0.36–4.29)
	0–14 Internet	0.38	(0.16–0.94)
	15–39 Internet	0.68	(0.27–1.67)
	40–64 Internet	1.33	(0.55–3.21)
	≥65 Internet	11.58	(1.17–114.24)
Educational level, yearsa	≤9	1.00
	10–14	0.24	(0.08–0.69)
	≥15	0.25	(0.08–0.74)
	Missing	0.24	(0.04–1.58)
Household size	1	1.00
	2	0.59	(0.25–1.41)
	3	1.76	(0.69–4.54)
	4	1.26	(0.47–3.36)
	≥5	1.48	(0.50–4.39)
	Missing	0.69	(0.03–15.71)
Household income groupb	Low & Middle/low	1.00
	Middle	1.81	(0.69–4.73)
	Middle/high	1.34	(0.53–3.39)
	High	0.80	(0.35–1.80)
	Missing	4.72	(0.45–49.27)
Event-driven disease report within 24 hours of registration	No	1.00
	Yes	1.03	(0.60–1.78)

^aEducation is the guardians’ highest education if child.

^bLow & Middle/low≤226,810; Middle = 226,811–340,466; Middle/high = 340,467–473,903; High≥473,904 in SEK in 2006.