Table 7. Multivariable pooled regression analysis for the association between vaccination schedule and vaccine type on hepatitis B vaccination timing among children aged 12–60 months in 47 countries.

Variable	First dose			Third dose
	No. of children vaccinateda	No. of children with delays	aOR   (95% CI)	No. of children vaccinateda	No. of children with delays	aOR   (95% CI)
Vaccination schedule start week
≤ 1	14 437	4 353	Ref.	9 565	5 602	Ref.
4	3 972	1 353	0.91 (0.80 to 1.03)	3 810	2 355	1.14 (1.00 to 1.30)
6	44 647	12 525	0.81 (0.75 to 0.88)	43 932	23 336	0.97 (0.91 to 1.03)
9	29 151	4 482	0.45 (0.41 to 0.50)	33 273	10 688	0.50 (0.46 to 0.53)
13	791	338	1.11 (0.92 to 1.34)	1 016	565	1.21 (1.03 to 1.42)
Vaccine type
Monovalent	37 763	8 305	Ref.	32 297	14 007	Ref.
Combination	60 055	14 746	0.76 (0.71 to 0.81)	59 299	28 538	0.99 (0.94 to 1.05)

aOR: adjusted odds ratio; CI: confidence interval; Ref.: reference category.

^a The number of children included in the analyses were adjusted for the covariates stated below.

Notes: Data were extracted from the most recent demographic and health survey (DHS) in each country (survey year range: 2005–2014). Total number of children (weighted counts) were 97 818 (first dose) and 91 596 (third dose). Total observations were 100 167 (first dose) and 93 807 (third dose). Denominators vary across variables because of item non-response. Model was adjusted for child’s age (yearly increments), sex, residence (urban versus rural), birth order of the child (1 versus > 1), mother’s age (yearly increments), mother’s marital status (married versus unmarried), mother’s education (none, primary, secondary and higher), birth place (home versus institutional), household wealth index (5 quintiles of wealth; poorest, poor, medium, rich, richest), family size (increments per member), country income level as per the World Bank (categorized as low income, lower-middle income and upper-middle income;²² and survey year. The variance inflation factors for the multivariable models were 1.06 for first dose (delayed) and 1.09 for third dose (delayed), respectively, indicating the absence of multicollinearity among explanatory variables.