4  Findings

4.1 Demographics

The Kano Zero-Dose Baseline Survey sample is representative of Nigerian children who were aged 12–23 months in early 2025. The tables in this section summarize the estimated number of children, households and proportion of households with age-eligible children (Section 4.1.1), as well as selected demographic characteristics of the children (Section 4.1.2), caregivers (Section 4.1.3), and households (Section 4.1.4) in the sample.

4.1.1 Age-Eligible Children

Table 4.1 estimates the total and average number of children per household and survey area. It shows that there are an estimated 194 thousand children aged 0-23 months in the three sentinel Local Government Areas (LGAs), with roughly 69 thousand in each of Gaya and Gabasawa and 57 thousand in Nassarawa LGA. There are large differences in the average number of children per household in the three sentinel LGAs, ranging from 0.73 in Gabasawa (a mostly rural LGA) to 0.30 in Nassarawa (a highly urban LGA).

Table 4.1: Estimated number of children (in thousands), by age cohort and LGA

	Number of children 0-23 months ('000)		Number of children 6-23 months ('000)		Number of children 12-23 months ('000)		Average children 0-23 months per HH		Sample Sizea
	Estimate	MOE	Estimate	MOE	Estimate	MOE	Estimate	MOE
Sentinel	194.9	±7.8	144.5	±6.2	94.5	±4.9	0.50	±0.02	30,142
Gaya	69.3	±3.0	52.1	±2.6	33.6	±2.1	0.65	±0.03	10,523
Gabasawa	68.6	±4.3	50.8	±3.3	32.2	±2.4	0.73	±0.04	10,688
Nassarawa	57.0	±2.4	41.6	±1.9	28.7	±1.7	0.30	±0.01	8,931
Non-sentinel	998.0	±85.1	734.8	±58.8	473.4	±32.2	0.56	±0.04	16,353
Tudun Wada	136.0	±30.1	99.5	±22.0	62.2	±12.8	0.68	±0.04	1,387
Ungogo	134.7	±26.3	98.5	±18.7	60.7	±12.2	0.45	±0.05	1,733
Gezawa	114.5	±24.8	87.0	±21.5	59.8	±15.5	0.70	±0.11	815
Sumaila	96.7	±19.6	68.4	±14.4	45.8	±9.9	0.62	±0.05	1,666
Kumbotso	82.7	±9.0	63.7	±7.0	40.6	±5.1	0.37	±0.03	2,917
Takai	82.5	±17.8	60.2	±13.0	39.6	±8.4	0.68	±0.06	1,060
Dawakin Tofa	80.9	±18.9	59.0	±13.9	36.1	±8.9	0.59	±0.05	1,072
Dawakin Kudu	75.6	±14.2	53.5	±10.7	33.4	±7.5	0.56	±0.05	1,361
Kiru	64.8	±10.6	49.7	±8.4	34.9	±6.2	0.70	±0.04	1,447
Bebeji	58.5	±12.3	44.7	±10.0	26.9	±7.0	0.66	±0.06	1,022
Dambatta	48.5	±11.0	33.4	±7.0	22.4	±5.2	0.53	±0.11	1,223
Tarauni	22.6	±9.5	17.2	±7.8	11.1	±5.2	0.29	±0.07	650
All 15 LGAs	1,192.9	±90.4	879.4	±62.7	567.9	±34.6	0.55	±0.03	46,495
Abbreviations: MOE = Margin of Error; HH = Household
aSample size of households answering the screening question upon initial contact

Across the three sentinel LGAs, approximately half (49.7%) of all households had at least one child under two years of age, with 36.9% reporting a child aged 6–23 months and 24.1% reporting a child aged 12–23 months. As shown in Table 4.2, the age-group distributions vary considerably across the LGAs: Gabasawa and Gaya show higher concentrations of young children, with 72.8% and 64.7% of households reporting at least one child under two, respectively, compared to just 29.9% in Nassarawa. These differences reflect underlying demographic patterns and fertility levels that influence household composition and program reach.

Table 4.2: Proportions of household with age-eligible children, by age cohort and survey area

	Proportion of HH with Children 0-23 months			Proportion of HH with Children 6-23 months			Proportion of HH with Children 12-23 months			Sample Sizea
	Estimate	MOE	DEFF	Estimate	MOE	DEFF	Estimate	MOE	DEFF
Sentinel	49.7%	±2.0	6.9	36.9%	±1.6	5.6	24.1%	±1.3	5.2	30,142
Gabasawa	72.7%	±4.4	9.6	53.9%	±3.4	7.2	34.2%	±2.5	5.5	10,688
Gaya	64.6%	±2.5	3.3	48.5%	±2.3	3.3	31.3%	±1.9	3.4	10,523
Nassarawa	29.9%	±1.1	1.0	21.9%	±0.9	0.9	15.0%	±0.8	1.0	8,931
Non-sentinel	56.0%	±3.6	12.5	41.2%	±2.4	6.8	26.6%	±1.1	2.1	16,353
Gezawa	70.4%	±11.2	4.4	53.5%	±10.6	4.7	36.7%	±7.9	3.5	815
Kiru	70.3%	±4.5	1.2	53.9%	±4.8	1.5	37.9%	±2.9	0.8	1,447
Tudun Wada	67.9%	±3.9	1.2	49.7%	±2.9	0.8	31.0%	±2.0	0.4	1,387
Takai	67.8%	±5.6	1.7	49.5%	±4.4	1.2	32.6%	±2.6	0.6	1,060
Bebeji	65.7%	±5.6	1.7	50.2%	±4.8	1.6	30.2%	±4.0	1.6	1,022
Sumaila	62.5%	±4.9	2.4	44.2%	±3.9	1.7	29.6%	±2.7	1.2	1,666
Dawakin Tofa	58.7%	±5.1	1.8	42.8%	±3.5	1.0	26.2%	±2.7	0.9	1,072
Dawakin Kudu	56.1%	±5.2	2.4	39.6%	±4.0	1.8	24.7%	±3.6	2.1	1,361
Dambatta	52.9%	±10.6	8.6	36.5%	±5.9	3.7	24.4%	±4.8	3.4	1,223
Ungogo	45.2%	±4.6	2.7	33.1%	±2.3	0.8	20.4%	±1.5	0.5	1,733
Kumbotso	37.5%	±2.9	2.2	28.9%	±2.3	1.7	18.4%	±1.8	1.5	2,917
Tarauni	28.6%	±7.0	3.4	21.8%	±6.3	3.6	14.1%	±4.4	2.5	650
All 15 LGAs	54.9%	±3.3	29.3	40.4%	±2.2	16.5	26.1%	±1.1	5.6	46,495
Abbreviations: MOE = Margin of Error; DEFF = Design Effect; HH = Household
aSample size of households answering the screening question upon initial contact

Table 4.3 presents household estimates across 15 target LGAs in Kano. The three sentinel LGAs collectively represent an estimated 389,200 households, accounting for only 18% of the total household population across the study area. In contrast, the 12 non-sentinel LGAs represent the remaining 1.7 million households, underscoring the concentration of population outside the sentinel sites.

The margin of error is low in the sentinel areas due both to the high sample size and the relatively direct sampling mechanism. In contrast, the non-sentinel LGAs show greater uncertainty and variability, reflected in a higher Coefficient of Variation (CV) (2.5%) and a Design Effect (DEFF) of 9.6. These elevated DEFFs in non-sentinel LGAs stem from two main factors: the use of cluster-based sampling (as opposed to direct stratified sampling of buildings in sentinel LGAs), and the presence of higher-density housing structures in some areas–particularly apartment buildings–which can increase within-cluster homogeneity.

Table 4.3: Estimated number of households, by LGA

	Number of Households ('000)				Mean HH per Building		Sample Sizea
	Estimate	MOE	CV	DEFF	Estimateb	MOE
Sentinel	389.8	±4.4	0.6%	1.0	0.93	±0.01	45,770
Nassarawa	188.2	±3.7	1.0%	1.0	0.85	±0.02	15,257
Gaya	108.0	±2.0	1.0%	0.9	1.05	±0.02	15,256
Gabasawa	93.6	±1.6	0.9%	1.0	1.01	±0.02	15,257
Non-sentinel	1,772.4	±85.5	2.5%	9.6	0.89	±0.04	27,602
Ungogo	289.3	±45.6	8.0%	22.9	0.76	±0.12	3,149
Kumbotso	223.3	±15.5	3.5%	8.5	0.68	±0.05	5,748
Tudun Wada	193.0	±37.2	9.8%	7.6	1.46	±0.28	1,911
Gezawa	170.1	±18.0	5.4%	2.9	1.05	±0.11	1,287
Sumaila	153.8	±34.7	11.5%	15.0	0.99	±0.22	2,556
Dawakin Kudu	134.2	±19.8	7.5%	7.2	0.91	±0.13	2,196
Dawakin Tofa	131.6	±24.7	9.6%	13.0	0.93	±0.17	1,951
Takai	121.4	±19.8	8.3%	6.1	1.00	±0.16	1,628
Kiru	93.8	±12.2	6.6%	5.6	0.84	±0.11	2,283
Dambatta	89.1	±14.0	8.0%	8.3	0.76	±0.12	1,976
Tarauni	88.0	±13.7	7.9%	6.8	0.91	±0.14	1,404
Bebeji	84.9	±14.6	8.8%	6.5	0.88	±0.15	1,513
All 15 LGAs	2,162.2	±85.6	2.0%	17.6	0.90	±0.04	73,372
Abbreviations: MOE = Margin of Error; CV = Coefficient of Variation; DEFF = Design Effect
aSample size of building footprints visited by the field team
bAverages include non-residential buildings with zero households

4.1.2 Infant Characteristics

This section tabulates background characteristics of infants 0-23 months in the 15 LGAs covered by the study.

Table 4.4: Distribution of the sex of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Boys	51.2%	±1.4	(49.8; 52.7)	50.0%	52.5%	4.97	659,943	±34,540	11,522
Girls	48.8%	±1.4	(47.3; 50.2)	47.5%	50.0%	4.97	628,081	±39,946	11,347
All	100.0%			100.0%	100.0%		1,288,024	±64,766	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 22,869

Table 4.5: Distribution of the birth order of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
1	59.5%	±2.4	(57.1; 61.9)	57.5%	61.5%	13.28	725,597	±48,450	12,856
2	9.6%	±0.9	(8.7; 10.6)	8.8%	10.4%	5.83	116,558	±13,107	2,083
3	6.7%	±0.8	(6.0; 7.5)	6.1%	7.4%	5.05	82,157	±9,308	1,609
4	5.9%	±0.6	(5.3; 6.6)	5.4%	6.4%	4.20	71,616	±8,672	1,339
5	5.3%	±0.6	(4.7; 6.0)	4.8%	5.9%	4.12	64,921	±7,604	1,138
6+	13.0%	±1.1	(11.9; 14.1)	12.0%	13.9%	6.42	158,004	±16,051	2,607
All	100.0%			100.0%	100.0%		1,218,853	±61,211	21,632
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 21,632

Table 4.6: Proportion of infants aged 0-23 months with disabilities

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Gabasawa	0.8%	±0.2	(0.6; 1.1)	0.6%	1.0%	1.15	603	±172	58	6,677
Nassarawa	0.7%	±0.4	(0.4; 1.3)	0.4%	1.1%	1.54	452	±259	16	2,554
Non-sentinel LGAs	0.7%	±0.3	(0.4; 1.0)	0.5%	0.9%	1.96	7,075	±2,690	49	7,520
Gaya	0.5%	±0.2	(0.3; 0.6)	0.3%	0.6%	0.89	337	±119	33	6,118
All	0.7%	±0.2	(0.5; 0.9)	0.5%	0.9%	4.23	8,466	±2,706	156	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 1.489 (p=0.172)

Table 4.7: Distribution of disability types among infants aged 0-23 months with disabilities

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Physical disabilities	60.8%	±13.1	(47.3; 72.9)	49.5%	71.1%	2.88	5,149	±1,853	91
Other	22.1%	±11.1	(13.0; 35.1)	14.2%	32.8%	2.86	1,875	±1,070	35
Visual impairment	**	**	**	**	**	2.82	1,210	±945	22
Hearing impairment	**	**	**	**	**	1.87	330	±394	7
Refuse to answer	0.2%	±0.4	(0.0; 1.5)	0.1%	1.1%	0.32	20	±36	2
Don't know	0.2%	±0.2	(0.1; 0.7)	0.1%	0.6%	0.13	16	±19	2
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 156; Potentially unreliable values have been suppressed and marked with asterisks (reason(s): The absolute CI width is ≥5% and the relative CI width is ≥130%); Note: Because answers follow a select-multiple format, proportions in this table do not necessarily sum to 100%

Table 4.8: Proportion of infants aged 0-23 months that are twins

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	5.1%	±1.2	(4.0; 6.5)	4.2%	6.3%	2.11	3,245	±813	125	2,554
Non-sentinel LGAs	4.1%	±0.8	(3.3; 5.0)	3.5%	4.8%	3.29	43,970	±8,951	308	7,520
Gabasawa	4.0%	±0.7	(3.4; 4.8)	3.4%	4.6%	2.24	2,941	±531	270	6,677
Gaya	3.7%	±0.6	(3.1; 4.4)	3.2%	4.3%	1.79	2,767	±488	210	6,118
All	4.1%	±0.7	(3.5; 4.9)	3.6%	4.7%	7.16	52,923	±9,095	913	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 1.178 (p=0.309)

Table 4.9: Proportion of infants aged 0–23 months registered at birth

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	59.2%	±2.2	(57.0; 61.4)	57.4%	61.0%	1.31	37,494	±1,996	1,519	2,554
Gaya	32.7%	±1.6	(31.2; 34.3)	31.4%	34.1%	1.80	24,246	±1,354	2,021	6,118
Non-sentinel LGAs	27.3%	±2.1	(25.2; 29.4)	25.6%	29.1%	4.31	293,954	±25,553	2,354	7,520
Gabasawa	26.8%	±1.3	(25.5; 28.2)	25.7%	28.0%	1.54	19,728	±1,060	1,949	6,677
All	29.1%	±1.8	(27.4; 31.0)	27.7%	30.7%	9.12	375,422	±25,750	7,843	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 99.622 (p<0.001)

Table 4.10: Distribution of infants aged 0-23 months by place of birth

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Home	78.5%	±1.5	(77.0; 79.9)	77.2%	79.7%	7.54	1,010,697	±58,396	17,854
Public Health facility	18.0%	±1.4	(16.6; 19.4)	16.8%	19.1%	7.52	231,424	±19,181	4,374
Private health facility	3.2%	±0.5	(2.7; 3.8)	2.8%	3.7%	5.41	41,104	±6,828	569
Other	0.4%	±0.2	(0.2; 0.6)	0.3%	0.5%	3.97	4,800	±2,034	72
All	100.0%			100.0%	100.0%		1,288,024	±64,766	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 22,869

Table 4.11: Proportion of infants aged 0-23 born in a health facility, by survey area

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	57.7%	±2.0	(55.7; 59.7)	56.0%	59.4%	1.12	36,551	±2,035	1,465	2,554
Non-sentinel LGAs	20.0%	±1.7	(18.3; 21.8)	18.6%	21.5%	3.64	215,561	±20,447	1,659	7,520
Gaya	15.5%	±1.0	(14.5; 16.6)	14.7%	16.4%	1.32	11,498	±815	965	6,118
Gabasawa	12.7%	±0.9	(11.8; 13.6)	11.9%	13.5%	1.30	9,318	±713	864	6,677
All	21.2%	±1.5	(19.8; 22.7)	20.0%	22.4%	7.59	272,928	±20,485	4,953	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 397.12 (p<0.001)

Table 4.12: Distribution of facility types where vaccinated infants aged 0-23 months received inoculations

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Public Health facility	94.5%	±1.0	(93.3; 95.4)	93.5%	95.3%	8.40	872,897	±45,789	14,889
Other	4.3%	±1.0	(3.5; 5.4)	3.6%	5.3%	9.46	40,093	±9,765	373
Private health facility	1.2%	±0.3	(0.9; 1.6)	0.9%	1.5%	3.78	10,869	±3,100	200
All	100.0%			100.0%	100.0%		923,859	±48,899	15,462
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 15,462

4.1.3 Caregiver Characteristics

This section tabulates background characteristics of the caregivers of infants 0-23 months in the 15 LGAs covered by the study.

Table 4.13: Distribution of the sex of caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Female	98.2%	±0.4	(97.7; 98.5)	97.8%	98.4%	3.87	1,084,914	±83,860	18,859
Male	1.8%	±0.4	(1.5; 2.3)	1.6%	2.2%	3.87	20,438	±4,459	512
All	100.0%			100.0%			1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.14: Distribution of the relationship of caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Mother	90.6%	±1.8	(88.6; 92.3)	89.0%	92.1%	23.45	1,165,299	±52,410	20,836
Other	7.3%	±1.9	(5.6; 9.4)	5.8%	9.0%	30.95	93,480	±27,148	1,375
Grandparent	1.2%	±0.2	(0.9; 1.4)	1.0%	1.4%	2.87	14,813	±3,008	348
Father	1.0%	±0.3	(0.7; 1.3)	0.8%	1.2%	4.36	12,284	±3,397	252
Refuse to answer							2,149	±1,011	58
All	100.0%			100.0%	100.0%		1,288,024	±64,766	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 22,869

Table 4.15: Proportion of infants aged 0-23 months whose reporting caregiver is their mother, by survey area

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	93.8%	±1.1	(92.5; 94.8)	92.7%	94.7%	1.48	59,374	±2,579	2,401	2,554
Non-sentinel LGAs	90.4%	±2.2	(87.9; 92.3)	88.4%	92.1%	10.81	973,290	±52,431	6,872	7,520
Gaya	90.3%	±1.0	(89.2; 91.2)	89.4%	91.1%	1.82	66,885	±2,033	5,535	6,118
Gabasawa	89.4%	±1.0	(88.4; 90.4)	88.6%	90.3%	1.92	65,750	±1,911	6,028	6,677
All	90.5%	±1.8	(88.5; 92.2)	88.8%	91.9%	23.25	1,165,299	±52,410	20,836	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 2.732 (p=0.085)

Table 4.16: Distribution of the marital status of caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Married	97.9%	±0.4	(97.5; 98.3)	97.6%	98.2%	3.80	1,082,559	±84,150	19,006
Divorced	1.0%	±0.3	(0.7; 1.3)	0.8%	1.2%	3.83	10,752	±3,251	176
Widowed	0.4%	±0.2	(0.3; 0.6)	0.3%	0.6%	2.74	4,820	±2,052	80
Never married	0.4%	±0.2	(0.2; 0.6)	0.3%	0.5%	3.98	4,104	±2,295	67
Separated	0.2%	±0.1	(0.1; 0.4)	0.1%	0.4%	4.01	2,469	±1,823	26
Other	0.1%	±0.0	(0.0; 0.1)	0.0%	0.1%	1.70	650	±672	16
All	100.0%			100.0%			1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.17: Distribution of the age of caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
< 18	1.0%	±0.3	(0.7; 1.2)	0.8%	1.2%	2.91	9,458	±2,477	230
18-20	18.5%	±1.3	(17.2; 19.8)	17.4%	19.6%	4.76	182,971	±20,077	3,248
21-25	28.2%	±1.5	(26.7; 29.7)	26.9%	29.5%	4.93	278,974	±26,770	4,556
26-30	26.0%	±1.2	(24.8; 27.2)	25.0%	27.0%	3.31	257,401	±22,442	4,570
31-35	13.1%	±1.0	(12.1; 14.1)	12.3%	13.9%	3.85	129,328	±14,325	2,083
36-40	8.9%	±0.8	(8.2; 9.8)	8.3%	9.6%	3.57	88,484	±10,553	1,516
above 40	4.4%	±0.7	(3.7; 5.1)	3.8%	5.0%	4.72	43,240	±7,685	694
Refusal or don't know							115,495	±18,104	2,474
All	100.0%			100.0%	100.0%		1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.18: Proportion of caregivers of infants aged 0-23 months who were born in Kano state

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Gabasawa	94.4%	±0.7	(93.7; 95.0)	93.8%	94.9%	1.19	55,403	±993	5,110	5,438
Gaya	94.4%	±0.8	(93.6; 95.1)	93.7%	95.0%	1.40	56,569	±1,002	4,723	4,993
Non-sentinel LGAs	90.6%	±1.6	(89.0; 92.1)	89.2%	91.9%	4.84	841,792	±79,481	5,938	6,597
Nassarawa	75.0%	±1.9	(73.1; 76.9)	73.4%	76.6%	1.21	43,450	±1,398	1,769	2,343
All	90.2%	±1.3	(88.8; 91.4)	89.1%	91.3%	9.75	997,214	±79,161	17,540	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 34.276 (p<0.001)

Table 4.19: Proportion of caregivers of infants aged 0-23 months who report having ever attended school

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	77.6%	±1.9	(75.7; 79.5)	76.0%	79.2%	1.29	44,951	±1,323	1,836	2,343
Non-sentinel LGAs	40.5%	±3.8	(36.7; 44.3)	37.3%	43.7%	10.28	376,010	±44,291	2,913	6,597
Gabasawa	26.0%	±1.3	(24.7; 27.3)	24.9%	27.0%	1.19	15,239	±761	1,507	5,438
Gaya	21.6%	±1.2	(20.4; 22.8)	20.6%	22.6%	1.18	12,934	±768	1,096	4,993
All	40.6%	±3.2	(37.5; 43.9)	38.0%	43.3%	21.38	449,133	±44,132	7,352	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 143.951 (p<0.001)

Table 4.20: Distribution of the highest education level attended for caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
No education	62.6%	±3.3	(59.3; 65.8)	59.8%	65.3%	23.13	692,085	±70,807	12,662
Primary	9.6%	±1.0	(8.6; 10.7)	8.7%	10.5%	6.18	105,841	±12,969	1,901
Secondary	24.0%	±2.8	(21.3; 27.0)	21.7%	26.5%	22.21	265,770	±35,363	4,106
More than secondary	3.8%	±0.6	(3.2; 4.4)	3.3%	4.3%	5.21	41,657	±6,628	702
All	100.0%			100.0%			1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.21: Proportion of caregivers of infants aged 0-23 months who report being able to read

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	76.4%	±1.9	(74.4; 78.2)	74.7%	77.9%	1.21	44,213	±1,377	1,787	2,343
Non-sentinel LGAs	36.5%	±3.7	(32.9; 40.3)	33.5%	39.7%	10.04	339,452	±41,263	2,639	6,597
Gaya	18.1%	±1.1	(17.0; 19.3)	17.2%	19.1%	1.13	10,847	±696	941	4,993
Gabasawa	18.1%	±1.1	(17.0; 19.2)	17.2%	19.1%	1.19	10,622	±667	1,033	5,438
All	36.7%	±3.1	(33.6; 39.8)	34.1%	39.3%	20.86	405,134	±41,118	6,400	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 262.216 (p<0.001)

Table 4.22: Distribution of the employment status of caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Unemployed	54.2%	±2.1	(52.1; 56.3)	52.5%	56.0%	8.88	599,338	±51,566	11,241
Self-employed	40.8%	±2.2	(38.6; 42.9)	39.0%	42.6%	9.83	450,539	±43,177	7,168
Other: ____	3.2%	±1.1	(2.3; 4.5)	2.5%	4.3%	18.45	35,791	±12,685	538
Employed full-time	0.8%	±0.2	(0.6; 1.0)	0.7%	1.0%	2.03	8,798	±2,029	210
Student	0.4%	±0.2	(0.3; 0.6)	0.3%	0.6%	3.05	4,445	±1,416	92
Employed part-time	0.4%	±0.1	(0.3; 0.5)	0.3%	0.5%	2.36	4,144	±1,675	100
Retired	0.2%	±0.1	(0.1; 0.4)	0.1%	0.4%	4.43	2,297	±1,681	22
All	100.0%			100.0%			1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.23: Distribution of the ethnic group of caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Hausa	82.7%	±2.2	(80.4; 84.8)	80.8%	84.5%	17.36	914,074	±77,959	15,288
Fulani	14.8%	±2.2	(12.8; 17.1)	13.1%	16.8%	19.21	163,701	±26,261	3,638
Other	1.3%	±0.3	(1.0; 1.7)	1.1%	1.7%	4.39	14,759	±4,465	237
Kanuri	0.9%	±0.2	(0.7; 1.2)	0.7%	1.2%	3.38	10,314	±3,507	157
Yoruba	0.1%	±0.1	(0.1; 0.2)	0.1%	0.2%	1.50	1,190	±474	25
Igbo	0.0%	±0.0	(0.0; 0.1)	0.0%	0.1%	2.51	482	±214	9
Nupe	0.0%	±0.0	(0.0; 0.1)	0.0%	0.1%	2.25	432		5
Refuse to answer							401	±94	12
All	100.0%			100.0%	100.0%		1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.24: Distribution of the religious affiliation of caregivers of infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Islam	98.9%	±0.7	(97.9; 99.4)	98.1%	99.4%	24.12	1,092,510	±85,812	19,204
Other Christian	0.6%	±0.4	(0.4; 1.1)	0.4%	1.0%	10.75	7,035	±5,041	97
Catholic	0.4%	±0.4	(0.2; 1.1)	0.2%	1.0%	19.70	4,828	±8,888	50
Other	0.0%	±0.0	(0.0; 0.1)	0.0%	0.1%	2.12	265		5
Traditionalist	0.0%	±0.0	(0.0; 0.1)	0.0%	0.0%	1.37	126	±332	4
Refuse to answer							589	±66	11
All	100.0%			100.0%	100.0%		1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.25: Percent of caregivers of infants aged 0-23 months self-identifying as Muslim, by survey area

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Gabasawa	99.9%	±0.1	(99.8; 99.9)	99.8%	99.9%	0.71	58,622	±946	5,429	5,438
Gaya	99.9%	±0.1	(99.8; 99.9)	99.8%	99.9%	0.81	59,853	±960	4,986	4,993
Non-sentinel LGAs	98.9%	±0.9	(97.6; 99.5)	97.9%	99.4%	11.31	918,321	±86,167	6,519	6,597
Nassarawa	96.2%	±0.9	(95.2; 97.1)	95.4%	96.9%	1.48	55,714	±1,145	2,270	2,343
All	98.8%	±0.7	(97.8; 99.4)	98.0%	99.3%	23.01	1,092,510	±85,812	19,204	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 13.987 (p<0.001)

Table 4.26: Proportion of caregivers of infants aged 0-23 months who report having ever used the Internet from any location on any device

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	37.3%	±2.2	(35.1; 39.5)	35.5%	39.1%	1.24	21,574	±1,349	860	2,343
Non-sentinel LGAs	12.6%	±2.1	(10.7; 14.8)	11.0%	14.5%	6.51	117,350	±20,650	934	6,597
Gaya	2.9%	±0.5	(2.4; 3.4)	2.5%	3.3%	1.06	1,716	±312	160	4,993
Gabasawa	2.3%	±0.4	(1.9; 2.8)	2.0%	2.7%	1.24	1,348	±264	131	5,438
All	12.8%	±1.7	(11.2; 14.7)	11.5%	14.4%	13.40	141,987	±21,367	2,085	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 110.673 (p<0.001)

4.1.4 Household Characteristics

This section tabulates background characteristics of households in the 15 LGAs covered by the study. Unless explicitly mentioned otherwise, statistics pertain to households with infants 0-23 months.

Table 4.27: Estimated total number of households, by LGA (includes households with no infants)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Ungogo	13.7%	±2.3	(11.6; 16.1)	11.9%	15.7%	52.83	306,936	±56,836	1,733
Kumbotso	10.1%	±0.9	(9.3; 11.0)	9.4%	10.9%	10.46	226,309	±18,827	2,917
Tudun Wada	9.1%	±1.7	(7.5; 11.0)	7.8%	10.7%	43.22	204,516	±41,032	1,387
Nassarawa	8.7%	±0.4	(8.3; 9.1)	8.4%	9.1%	2.60	194,930	±4,785	8,931
Gezawa	7.5%	±0.9	(6.6; 8.5)	6.7%	8.3%	15.58	167,480	±21,803	815
Sumaila	7.1%	±1.2	(6.0; 8.4)	6.2%	8.1%	25.47	158,725	±27,424	1,666
Dawakin Tofa	6.3%	±1.3	(5.1; 7.7)	5.3%	7.4%	32.82	140,707	±29,583	1,072
Dawakin Kudu	6.2%	±0.9	(5.4; 7.2)	5.5%	7.0%	16.80	139,250	±20,239	1,361
Takai	5.6%	±0.9	(4.7; 6.5)	4.9%	6.4%	17.98	124,511	±20,698	1,060
Gaya	5.0%	±0.2	(4.7; 5.2)	4.8%	5.2%	1.31	111,004	±2,157	10,523
Kiru	4.5%	±0.6	(3.9; 5.1)	4.0%	5.0%	11.30	99,690	±14,189	1,447
Gabasawa	4.3%	±0.2	(4.1; 4.6)	4.2%	4.5%	1.28	97,390	±1,838	10,688
Dambatta	4.2%	±0.7	(3.6; 5.0)	3.7%	4.8%	14.15	94,533	±15,589	1,223
Bebeji	4.1%	±0.8	(3.4; 4.9)	3.5%	4.8%	18.35	91,002	±17,400	1,022
Tarauni	3.7%	±0.9	(2.9; 4.6)	3.0%	4.5%	25.20	82,028	±19,750	650
All	100.0%			100.0%	100.0%		2,239,009	±95,273	46,495
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 46,495

Table 4.28: Distribution of the urban/rural category of households (includes households with no infants)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Rural	54.2%	±4.1	(50.1; 58.3)	50.7%	57.7%	82.59	1,213,934	±109,136	26,160
Urban	45.8%	±4.1	(41.7; 49.9)	42.3%	49.3%	82.59	1,025,075	±98,886	20,335
All	100.0%			100.0%	100.0%		2,239,009	±95,273	46,495
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 46,495

Details on the urban and rural classifications can be found in Appendix E.

Table 4.29: Proportion of rural households, by survey area (includes households with no infants)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Gabasawa	94.3%	±0.9	(93.3; 95.2)	93.5%	95.0%	4.47	91,843	±1,975	9,992	10,688
Gaya	69.8%	±2.1	(67.6; 71.9)	68.0%	71.6%	5.96	77,511	±2,847	7,222	10,523
Non-sentinel LGAs	56.9%	±4.9	(52.0; 61.7)	52.8%	60.9%	41.21	1,044,580	±106,898	8,946	16,353
Nassarawa	0.0%	±0.0	(0.0; 0.7)	0.0%	0.0%		0	±0	0	8,931
All	54.2%	±4.1	(50.1; 58.3)	50.7%	57.7%	82.59	1,213,934	±109,136	26,160	46,495
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 207.177 (p<0.001)

Table 4.30: Distribution of the degree of urbanization of households (detailed classification based on the GHS-SMOD data, includes households with no infants)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Mostly uninhabited area	0.6%	±0.2	(0.4; 0.9)	0.5%	0.9%	11.28	14,142	±5,487	207
Dispersed rural area	8.1%	±1.9	(6.4; 10.2)	6.6%	9.8%	58.95	180,860	±43,821	3,008
Village	4.0%	±1.5	(2.8; 5.7)	3.0%	5.4%	65.85	89,774	±33,008	1,485
Suburban or peri-urban area	28.0%	±3.8	(24.4; 31.9)	24.9%	31.2%	84.25	625,821	±89,507	13,282
Semi-dense town	5.2%	±1.5	(3.9; 6.9)	4.1%	6.6%	52.53	116,160	±33,563	2,822
Dense town	14.0%	±2.6	(11.6; 16.7)	12.0%	16.3%	65.31	313,348	±57,828	7,688
City	40.1%	±5.3	(35.0; 45.5)	35.8%	44.6%	138.57	898,905	±119,762	18,003
All	100.0%			100.0%	100.0%		2,239,009	±95,273	46,495
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 46,495

Table 4.31: Proportion of households of infants aged 0-23 months who have a member of the household with a disability

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Gabasawa	9.4%	±0.9	(8.5; 10.3)	8.7%	10.2%	1.37	5,509	±539	521	5,438
Gaya	8.7%	±0.9	(7.9; 9.7)	8.0%	9.5%	1.26	5,231	±530	452	4,993
Non-sentinel LGAs	8.1%	±1.0	(7.1; 9.2)	7.3%	9.1%	2.50	75,495	±12,719	492	6,597
Nassarawa	5.3%	±1.0	(4.3; 6.4)	4.5%	6.2%	1.31	3,058	±637	119	2,343
All	8.1%	±0.9	(7.2; 9.0)	7.4%	8.8%	5.27	89,294	±12,851	1,584	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 2.793 (p=0.012)

Table 4.32: Proportion of households of infants aged 0-23 months with electricity

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	87.2%	±1.5	(85.7; 88.6)	85.9%	88.3%	1.15	50,477	±1,304	2,037	2,343
Non-sentinel LGAs	35.9%	±5.1	(31.0; 41.2)	31.8%	40.3%	19.40	333,777	±49,914	2,689	6,597
Gaya	22.4%	±1.2	(21.2; 23.6)	21.3%	23.4%	1.16	13,405	±785	1,118	4,993
Gabasawa	4.0%	±0.6	(3.5; 4.7)	3.5%	4.5%	1.29	2,356	±350	225	5,438
All	36.2%	±4.3	(32.0; 40.6)	32.7%	39.9%	40.23	400,015	±49,722	6,069	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 345.619 (p<0.001)

Table 4.33: Proportion of households of infants aged 0-23 months with a radio

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	61.3%	±2.2	(59.1; 63.5)	59.5%	63.1%	1.22	35,508	±1,402	1,443	2,343
Non-sentinel LGAs	26.7%	±2.6	(24.3; 29.4)	24.6%	28.9%	5.72	248,398	±31,331	1,912	6,597
Gaya	17.9%	±1.2	(16.8; 19.1)	16.9%	18.9%	1.27	10,737	±740	897	4,993
Gabasawa	15.6%	±1.1	(14.5; 16.7)	14.6%	16.5%	1.41	9,132	±695	808	5,438
All	27.5%	±2.2	(25.4; 29.7)	25.7%	29.3%	11.79	303,774	±31,243	5,060	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 234.757 (p<0.001)

Table 4.34: Proportion of households of infants aged 0-23 months with a television

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	54.0%	±2.2	(51.7; 56.2)	52.1%	55.8%	1.20	31,241	±1,374	1,269	2,343
Non-sentinel LGAs	13.9%	±1.8	(12.1; 15.8)	12.4%	15.5%	4.86	128,944	±17,318	1,128	6,597
Gaya	4.4%	±0.6	(3.8; 5.1)	3.9%	5.0%	1.18	2,644	±391	230	4,993
Gabasawa	2.2%	±0.5	(1.7; 2.7)	1.8%	2.6%	1.48	1,272	±278	118	5,438
All	14.8%	±1.6	(13.3; 16.5)	13.6%	16.2%	9.92	164,101	±17,514	2,745	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 547.982 (p<0.001)

Table 4.35: Proportion of households of infants aged 0-23 months with a fixed telephone line

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	8.0%	±1.2	(6.8; 9.3)	7.0%	9.1%	1.26	4,624	±723	172	2,343
Non-sentinel LGAs	3.9%	±1.1	(3.0; 5.1)	3.1%	4.9%	4.99	36,408	±10,481	215	6,597
Gaya	3.2%	±0.6	(2.6; 3.8)	2.7%	3.7%	1.55	1,889	±408	141	4,993
Gabasawa	2.5%	±0.5	(2.1; 3.1)	2.2%	3.0%	1.25	1,495	±303	137	5,438
All	4.0%	±0.9	(3.2; 5.0)	3.3%	4.8%	10.20	44,416	±10,683	665	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 15.429 (p<0.001)

Table 4.36: Proportion of households of infants aged 0-23 months with a computer

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	9.6%	±1.3	(8.4; 11.0)	8.6%	10.7%	1.13	5,572	±736	230	2,343
Non-sentinel LGAs	2.1%	±0.5	(1.7; 2.7)	1.8%	2.6%	1.88	19,680	±4,439	183	6,597
Gabasawa	0.9%	±0.3	(0.6; 1.2)	0.7%	1.1%	1.27	509	±206	49	5,438
Gaya	0.7%	±0.3	(0.5; 1.1)	0.6%	1.0%	1.32	449	±191	37	4,993
All	2.4%	±0.4	(2.0; 2.8)	2.0%	2.7%	3.71	26,211	±4,851	499	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 128.353 (p<0.001)

Table 4.37: Proportion of households of infants aged 0-23 months with a refrigerator

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	33.6%	±2.1	(31.6; 35.7)	31.9%	35.4%	1.19	19,459	±1,260	778	2,343
Non-sentinel LGAs	7.6%	±1.2	(6.5; 8.8)	6.7%	8.6%	3.26	70,461	±10,033	652	6,597
Gaya	1.8%	±0.4	(1.5; 2.3)	1.5%	2.2%	1.02	1,101	±258	107	4,993
Gabasawa	0.7%	±0.3	(0.5; 1.1)	0.5%	1.0%	1.57	427	±196	42	5,438
All	8.3%	±1.0	(7.3; 9.3)	7.5%	9.2%	6.64	91,449	±10,749	1,579	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 417.879 (p<0.001)

Table 4.38: Proportion of households of infants aged 0-23 months with a member of the family that owns a mobile phone

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	89.9%	±1.4	(88.4; 91.2)	88.7%	91.0%	1.34	52,058	±1,218	2,118	2,343
Non-sentinel LGAs	71.2%	±2.8	(68.4; 73.9)	68.8%	73.5%	6.42	661,649	±65,603	4,814	6,597
Gaya	55.3%	±1.6	(53.8; 56.9)	54.0%	56.6%	1.28	33,163	±1,088	2,766	4,993
Gabasawa	48.7%	±1.5	(47.2; 50.3)	47.5%	50.0%	1.33	28,604	±1,025	2,639	5,438
All	70.2%	±2.3	(67.8; 72.4)	68.2%	72.1%	13.07	775,474	±65,347	12,337	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 152.351 (p<0.001)

Table 4.39: Proportion of households of infants aged 0-23 months with a member of the family that owns a car or truck

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	9.4%	±1.3	(8.2; 10.7)	8.4%	10.5%	1.15	5,433	±739	221	2,343
Gabasawa	6.7%	±0.8	(5.9; 7.6)	6.0%	7.4%	1.60	3,926	±503	343	5,438
Non-sentinel LGAs	5.0%	±1.0	(4.1; 6.1)	4.3%	5.9%	3.56	46,811	±10,266	323	6,597
Gaya	3.9%	±0.7	(3.3; 4.7)	3.4%	4.5%	1.52	2,358	±402	184	4,993
All	5.3%	±0.8	(4.5; 6.2)	4.6%	6.0%	7.13	58,527	±10,268	1,071	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 15.819 (p<0.001)

Table 4.40: Proportion of households of infants aged 0-23 months with a member of the family that owns a motorcycle or scooter

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	40.0%	±2.2	(37.9; 42.2)	38.2%	41.9%	1.22	23,183	±1,352	933	2,343
Non-sentinel LGAs	36.3%	±2.0	(34.3; 38.3)	34.6%	38.0%	2.97	336,822	±37,664	2,453	6,597
Gabasawa	28.4%	±1.4	(27.0; 29.8)	27.2%	29.6%	1.39	16,640	±884	1,547	5,438
Gaya	26.5%	±1.4	(25.1; 27.9)	25.3%	27.7%	1.36	15,859	±904	1,323	4,993
All	35.5%	±1.7	(33.8; 37.2)	34.1%	36.9%	6.33	392,503	±37,546	6,256	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 34.816 (p<0.001)

Table 4.41: Proportion of households of infants aged 0-23 months with a member of the family that owns any land that can be used for agriculture

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Gabasawa	90.2%	±0.9	(89.3; 91.1)	89.4%	90.9%	1.28	52,934	±1,018	4,870	5,438
Gaya	88.8%	±1.0	(87.7; 89.7)	87.9%	89.6%	1.31	53,195	±1,037	4,436	4,993
Non-sentinel LGAs	73.5%	±3.7	(69.6; 77.1)	70.3%	76.6%	12.28	683,048	±73,013	4,780	6,597
Nassarawa	19.9%	±1.8	(18.2; 21.7)	18.5%	21.4%	1.18	11,521	±1,032	462	2,343
All	72.4%	±3.1	(69.2; 75.5)	69.7%	75.0%	24.94	800,699	±72,718	14,548	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 220.968 (p<0.001)

Table 4.42: Proportion of households of infants aged 0-23 months that own livestock, animal herds, poultry, or other animals

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Gabasawa	84.4%	±1.1	(83.3; 85.4)	83.5%	85.2%	1.20	49,506	±1,060	4,507	5,438
Gaya	76.1%	±1.3	(74.8; 77.4)	75.0%	77.2%	1.19	45,599	±1,114	3,735	4,993
Non-sentinel LGAs	65.5%	±3.1	(62.3; 68.4)	62.9%	68.0%	7.02	607,993	±65,174	4,240	6,597
Nassarawa	35.7%	±2.2	(33.6; 37.9)	33.9%	37.6%	1.23	20,686	±1,326	820	2,343
All	65.5%	±2.6	(62.9; 68.0)	63.3%	67.6%	14.61	723,785	±64,923	13,302	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 188.392 (p<0.001)

Table 4.43: Distribution of the main material of the roof for households with infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Finished roofing (metal / tin, wood calamine / cement fibre, ceramic tile, cement or roofing shingles)	63.1%	±2.9	(60.1; 65.9)	60.6%	65.5%	18.22	695,856	±65,574	10,036
Natural roofing (thatch /palm leaf or sod)	17.5%	±2.2	(15.3; 19.8)	15.7%	19.4%	17.58	192,613	±27,705	4,572
Rudimentary roofing (rustic mat, palm / bamboo, wood planks, cardboard)	17.3%	±1.7	(15.7; 19.1)	15.9%	18.8%	10.14	191,082	±23,690	4,230
No roof	1.8%	±0.8	(1.2; 2.8)	1.3%	2.6%	16.32	20,191	±9,958	377
Other (specify)	0.3%	±0.2	(0.1; 0.7)	0.2%	0.6%	9.93	3,403	±906	95
Don't know or cannot be determined from outside							2,207	±1,284	61
All	100.0%			100.0%	100.0%		1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.44: Distribution of the main material of exterior walls for households with infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Finished walls (cement stone with lime / cement, bricks, cement blocks, covered adobe, wood planks / shingles)	46.9%	±3.5	(43.4; 50.3)	44.0%	49.8%	24.27	517,883	±54,825	7,462
Rudimentary walls (bamboo, with mud, stone with mud, uncovered adobe, plywood, cardboard, reused wood )	32.7%	±2.9	(29.9; 35.6)	30.4%	35.1%	18.69	361,380	±43,569	6,637
Natural walls (cane / palm / trunks or dirt)	19.7%	±2.4	(17.4; 22.3)	17.8%	21.9%	18.87	217,980	±31,524	4,959
No walls	0.5%	±0.1	(0.3; 0.6)	0.4%	0.6%	2.20	5,111	±1,699	223
Other (specify)	0.2%	±0.1	(0.2; 0.3)	0.2%	0.3%	1.92	2,536	±864	80
Don't know or cannot be determined from outside							462	±139	10
All	100.0%			100.0%	100.0%		1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371

Table 4.45: Wealth index quintiles for households with infants aged 0-23 months

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Poorest	18.4%	±2.4	(16.2; 21.0)	16.5%	20.5%	19.22	203,811	±31,638	5,238
Poorer	19.1%	±2.1	(17.1; 21.3)	17.4%	21.0%	14.93	211,280	±32,011	3,824
Middle	20.3%	±2.0	(18.4; 22.4)	18.7%	22.0%	11.93	224,797	±29,193	3,473
Richer	20.6%	±2.3	(18.4; 23.0)	18.7%	22.6%	16.50	227,904	±30,740	3,118
Richest	21.5%	±2.8	(18.8; 24.4)	19.2%	23.9%	23.31	237,561	±33,324	3,718
All	100.0%			100.0%			1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371
Note: Wealth quintiles are approximated using principal components analysis, which collapses information on household characteristics and amenities into a single wealth index variable, following the well-established DHS methodology. Further details on the wealth index construction are provided in the corresponding Annex.

4.2 Routine Immunization

This chapter presents a routine immunization assessment for the 15 LGAs targeted in the Kano Zero-Dose Baseline Survey. It will examine how well the Kano health system is delivering vaccines to age-eligible children according to the national schedule, looking at coverage rates for different vaccines, zero-dose prevalence (both on a crude and valid basis), dropout rates between doses, vaccination timeliness, and missed opportunities for vaccination.

4.2.1 Overview and Definitions

Following the Vaccination Coverage Quality Indicators (VCQI) standard, the crude basis of calculation considers an infant to be vaccinated if there is either evidence of immunization in home-based records or if the child’s caregiver recalls that their child received the given vaccination from memory. The crude basis may describe specific vaccinations or the vaccination status of a child as a whole (e.g., crude-basis zero-dose, crude-basis fully-vaccinated, etc.).

Following the VCQI standard, this report defines the valid basis of calculation considers an infant to be vaccinated only if there is documented evidence of immunization in home-based records. In addition, for the valid basis, the following must be true: (a) The child had reached the minimum age of eligibility for this dose; (b) If the schedule specifies a maximum age of eligibility, then the child was within the allowable age range when they received the dose; and (c) If the dose is number 2 or 3 (or higher) in a sequence, then the minimum interval had passed since receiving the earlier dose, so the child was eligible to receive the next dose.

4.2.2 Vaccination Cards

For health planners in Nigeria, reliable vaccine card record keeping is crucial for generating accurate data on routine immunization coverage. Home-based records (HBRs), such as vaccination cards, are a central source of verification during household surveys and are essential for accurately assessing whether most children have received vaccines according to the national schedule, as recall from memory can be prone to human error. When cards are incomplete, unavailable, or contain unclear or inconsistent entries, it becomes difficult to determine true coverage levels, identify gaps, or plan targeted interventions. In a context where administrative data systems may have limitations, high-quality HBRs are a key tool for informing evidence-based decisions, allocating resources efficiently, and ultimately strengthening immunization program performance at local administrative levels. The following section reports on the observed frequency and quality of HBRs by the survey team.

4.2.2.1 Card availability

RI_QUAL_01

NoteSurvey Question 412 (select multiple)

Text: Do you have a National Child Immunisation Record immunisation records from a private health provider or any other document where {childname}’s vaccinations are written down? YES → May I see the card(s) (and/or) other document?

Scope: All caregivers of infants 0-23 (once per child)

Table 4.46: Proportion of the 12-23 population estimated to have a home-based record available for interviewers to view at the time of the survey

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
RI Card Availability (%)	55.4%	±3.1	(52.3; 58.4)	52.8%	57.9%	10.67	339,637	±25,477	6,250
RI Card with Dates or Ticks (%)	55.3%	±3.1	(52.3; 58.4)	52.8%	57.9%	10.67	339,502	±25,479	6,239
RI Card with Dates (%)	54.6%	±3.0	(51.5; 57.6)	52.0%	57.1%	10.54	334,884	±25,408	5,986
RI Card with Only Clean Dates (%)	20.7%		(19.4; 21.9)
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 10,895; Note: Because answers follow a select-multiple format, proportions in this table do not necessarily sum to 100%

An estimated 55% of children aged 12–23 months who were eligible for the survey had a HBR available at the time of interview. In the vast majority of cases, caregivers presented the national immunization card, as fewer than 1% of households with HBRs showed the field team some other document. In nearly all those cases, the record contained at least one vaccination entry—whether as a date or a tick mark. However, none of the available records were found to contain only “clean” vaccination dates, defined as dates that fall within the plausible window (after birth and before the interview) and follow the correct chronological order for multi-dose series.

Table 4.47: Proportion of infants 12-23 where the caregiver showed their vaccination card to the field team if it was reported to exist

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Yes, I can see the card(s)	55.1%	±3.0	(52.0; 58.1)	52.5%	57.6%	10.58	337,990	±25,470	6,191
No, no card and no other document seen	44.6%	±3.1	(41.6; 47.7)	42.1%	47.2%	10.67	273,953	±27,466	4,645
Yes, I can see the other document(s)	0.7%	±0.2	(0.5; 0.9)	0.5%	0.9%	1.46	4,151	±1,142	131
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 10,895; Note: Because answers follow a select-multiple format, proportions in this table do not necessarily sum to 100%

As shown in Figure 4.1, the likelihood that HBRs are available is associated with the age of the child. It is the highest in the 70%s range during the infants’ first three months of life, hovers above 60% from three to 16 months, and then drops steadily starting at 16 months to 33% likelihood at 23 months. The visualization underscores not only that recall-based assessment may be more important for vaccinations administered later in life, but also for assessing the early-infancy vaccination history of children who are nearing their second birthday.

The likelihood of seeing vaccination cards is higher for those born in the health facility (estimated at 21.2%, or 272,928 infants, as shown in Table 4.48) compared to those born elsewhere (estimated at 78.8%, or 1,015,096 infants). For both groups, this likelihood begins to drop noticeably at 16 months.

Figure 4.1: Home-based record availability over time. The dashed lines represent smoothed trend lines estimating the probability that a child’s HBR were shown to the field team. A 95% confidence band is shown around the estimate. The likelihood of seeing vaccination cards is consistently higher for children born in the health facility, and begins to drop noticeably at 16 months.

Table 4.48: Distribution of delivery location

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Elsewhere	78.8%	±1.5	(77.3; 80.2)	77.6%	80.0%	7.59	1,015,096	±58,517	17,916
Health facility	21.2%	±1.5	(19.8; 22.7)	20.0%	22.4%	7.59	272,928	±20,485	4,953
All	100.0%			100.0%	100.0%		1,288,024	±64,766	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 22,869

4.2.2.2 Ever had a card

RI_QUAL_02

NoteSurvey Question 411 (select one)

Text (Q411): Did you ever have a National Child Immunization Record or immunisation records from a private health provider for {childname}?

Scope: All caregivers of infants 0-23 (once per child)

Table 4.49: Proportion of vaccinated infants 12-23 where the caregiver claims to have a vaccination card (whether shown or not)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Yes, has only card(s)	84.9%	±2.0	(82.8; 86.7)	83.1%	86.4%	6.27	385,665	±28,507	6,935
No, has no card and no other document	13.9%	±1.9	(12.1; 16.0)	12.4%	15.6%	6.47	63,398	±9,114	945
Yes, has card(s) and other document	0.9%	±0.3	(0.6; 1.2)	0.7%	1.2%	1.99	3,979	±1,273	126
Yes, has only other document	0.3%	±0.1	(0.2; 0.5)	0.2%	0.5%	1.22	1,451	±611	75
All	100.0%			100.0%	100.0%		454,493	±29,749	8,081
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 8,081

Table 4.50: Proportion of infants 12-23 where the caregiver claims to have owned a vaccination card for the child at some point (whether shown or not), by area

	Proportiona	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	62.7%	±3.2	(59.5; 65.8)	60.0%	65.3%	1.22	17,148	±1,450	684	1,103
Gabasawa	49.5%	±2.6	(46.9; 52.1)	47.4%	51.7%	1.59	12,777	±935	1,135	2,276
Gaya	46.9%	±2.6	(44.3; 49.5)	44.7%	49.0%	1.39	11,501	±819	960	2,003
Non-sentinel LGAs	43.7%	±3.8	(39.9; 47.5)	40.5%	46.9%	4.13	164,558	±19,527	1,173	2,699
All	45.3%	±3.2	(42.2; 48.5)	42.7%	48.0%	8.53	205,984	±19,604	3,952	8,081
aProportion of infants aged 12-23 months whose caregives had home-based vaccination records at some point in time
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 24.783 (p<0.001)

Forty-five percent (45%) of the population who were eligible for the survey are estimated to have received at least one home-based record (vaccination card), even if they no longer have it. At 63%, the rate of HBR ownership is considerably higher for Nassarawa LGA compared to other survey areas \((d=18.5\), \(t_{(498)}=7.62\), \(p<.001)\) Although it is expected that an LGA with higher vaccination coverage would have a higher proportion of vaccination cards, what makes Nassarawa interesting is that it also had a higher than usual percentage of caregivers reporting vaccination via recall from memory \((d=11.5\), \(t_{(498)}=5.49\), \(p<.001)\).

4.2.3 Estimated Coverage

4.2.3.1 Vaccine-wise Coverage

We begin with a summary of coverage estimates for each of the principal vaccines included in Nigeria’s routine immunization schedule for infants aged 12-23 months. This vaccine-wise presentation serves as a foundation for understanding broader patterns of immunization. Table 4.51 displays point estimates for each antigen by geographic area, with more detailed statistics such as sampling errors presented in the annex (Appendix A). Table 4.52 explores how these estimates vary across key background characteristics. Characteristics include those reported in the 2023-24 Nigeria Demographic and Health Survey (NDHS) tabulation, including the household wealth quintile. Routine immunization coverage across Pathways vulnerability segments is displayed for completeness, but covered in greater detail in Section 4.2.3.5.

The LGA-based estimates in Table 4.51 provide a detailed snapshot of routine immunization performance across different geographic areas in Kano for key antigens. Within the sentinel areas, Nassarawa generally has higher coverage rates, a pattern that is generally expected given the fact it stands out as the only urban sentinel LGA. Within non-sentinel areas, Tarauni, Dawakin Tofa, and Kumbotso have the highest coverage rates, whereas Tudun Wada and Bebeji stand out as LGAs with particularly low coverage rates. Although some caution should be used in interpreting the LGA-level estimates in non-sentinel areas due to smaller sample sizes, we can see that there is wide variation in the coverage from one LGA to another. As a group, sentinel LGAs do have vaccination rates that tend to be slightly higher than the non-sentinel areas. Indeed, using Penta-1 as a representative case, we find that the 6.0-point difference between sentinel and non-sentinel areas is small but statistically significant in a design-based \(t\)-test \((d=-6.0\), \(t_{(498)}=-3.61\), \(p<.001)\). However, it should be noted that pairwise comparisons from specific LGAs in each grouping– sentinel and non-sentinel– do not allows follow the same patterns observed by comparing their umbrella groups.

Table 4.51: Vaccination coverage by antigen, dose, and area

	Infants 12-23 months																									15-23 months
	BCG	HepB	OPV				Penta			IPV		Pneumococcal			Rotavirus			MCV1	YF	Meni.	Fully vaccinated		Zero-dose		Sampled # of children
			0	1	2	3	1	2	3	1	2	1	2	3	1	2	3				B.A.a	N.I.S.b	Truec	Gavid		MCV2
Sentinel	70.4	57.2	67.6	68.1	65.5	61.0	67.7	66.3	64.8	65.3	42.8	67.2	65.9	63.9	66.7	65.1	61.7	63.4	32.5	31.1	61.9	21.8	28.9	32.3	7,287	58.9
Nassarawa	79.7	72.2	77.0	75.8	70.8	63.7	74.7	72.1	70.0	74.0	40.0	73.9	71.1	68.4	73.1	70.1	66.5	68.2	22.2	19.8	66.2	16.4	19.6	25.3	1,283	61.6
Gabasawa	68.7	53.3	66.2	66.9	64.8	60.7	66.8	65.9	64.6	65.0	47.5	66.5	65.7	63.9	66.4	65.4	61.9	63.1	38.1	38.0	61.3	25.8	30.5	33.2	3,092	58.2
Gaya	63.6	47.6	60.5	62.5	61.4	58.7	62.1	61.5	60.5	57.9	40.8	62.0	61.3	59.9	61.4	60.2	57.4	59.5	36.3	34.5	58.7	22.9	35.7	37.9	2,912	57.2
Non-sentinel	64.6	53.6	65.7	63.2	60.4	56.5	61.7	60.0	58.2	61.1	55.0	61.7	60.1	57.6	61.4	59.6	56.3	57.1	49.0	48.5	56.0	37.3	33.6	38.3	3,608	50.1
Tarauni	84.3	78.5	81.7	79.5	74.7	71.4	81.5	78.8	75.3	80.1	78.6	80.2	78.8	74.0	81.5	80.1	77.1	74.8	62.7	60.8	72.8	58.4	15.7	18.5	85	63.6
Kumbotso	79.8	72.4	79.6	74.9	71.5	67.0	76.3	74.0	72.0	75.5	67.5	76.1	74.2	70.0	75.6	73.0	68.0	70.2	59.6	59.0	69.0	51.8	19.4	23.7	484	66.9
Dawakin Tofa	77.5	63.0	78.8	77.6	76.7	71.8	76.3	75.7	74.1	75.7	71.5	76.0	75.4	73.8	75.1	74.5	73.1	73.5	64.7	63.6	73.5	49.0	21.2	23.7	231	65.2
Ungogo	74.0	67.5	77.2	73.5	69.6	63.4	69.5	67.4	62.3	68.4	57.4	69.7	67.6	62.5	68.8	67.2	61.4	62.6	52.8	52.1	59.4	41.8	22.8	30.5	310	55.6
Dambatta	71.1	58.4	72.0	70.7	66.2	61.8	68.7	65.7	62.5	67.2	62.0	68.9	66.5	62.3	68.3	64.7	61.0	61.3	46.6	47.0	59.1	35.1	27.1	31.3	219	54.7
Takai	63.3	47.4	67.1	63.2	61.3	53.8	61.0	60.3	58.2	60.2	51.6	61.0	60.3	57.0	59.9	59.2	55.4	57.0	50.6	50.7	56.5	33.8	32.9	39.0	280	46.4
Gezawa	61.5	46.7	61.1	59.9	58.6	55.7	59.7	58.7	58.0	59.6	52.5	59.7	58.7	57.8	59.6	58.7	55.3	55.8	48.9	*	55.8	31.1	38.5	40.3	228	46.1
Kiru	60.6	48.7	63.2	61.2	57.9	55.5	58.9	56.9	56.2	58.5	53.7	59.2	56.7	56.3	58.4	56.4	53.9	55.3	43.3	45.1	54.1	33.2	35.9	41.1	422	49.2
Sumaila	58.5	43.0	59.3	58.0	55.0	52.2	56.4	53.4	52.8	54.4	48.4	56.4	53.4	52.5	56.1	52.6	50.6	50.6	42.2	41.5	49.7	29.0	40.1	43.6	395	45.6
Dawakin Kudu	57.2	48.0	57.7	55.4	53.9	51.3	54.2	53.4	52.1	54.3	49.5	54.5	53.4	51.3	53.8	53.1	48.8	52.3	46.9	46.5	51.0	35.7	40.6	45.8	300	45.4
Bebeji	53.6	48.3	52.2	50.3	48.8	44.9	49.7	48.1	46.6	50.4	46.5	49.7	48.3	46.7	49.7	48.1	45.2	45.9	40.9	39.2	45.2	34.8	45.3	50.3	267	34.1
Tudun Wada	51.4	42.7	52.9	49.4	45.3	43.2	47.1	45.4	44.7	47.0	44.9	47.3	45.6	43.3	48.0	46.1	44.2	43.4	39.9	39.8	43.1	32.1	46.7	52.9	387	35.4
All	65.5	54.2	66.0	64.0	61.2	57.2	62.7	61.0	59.3	61.8	53.0	62.6	61.0	58.6	62.3	60.5	57.2	58.1	46.3	45.6	57.0	34.7	32.8	37.3	10,895	51.5
Note: Children are considered to have received the vaccine if it was either written on the child’s vaccination card or reported by the caregiver. For children whose vaccination information is based on the caregiver's report, date of vaccination is not collected. The proportions of vaccinations given during the first and second years of life are assumed to be the same as for children with a written record of vaccination. Unreliable proportions have been suppressed and replaced with asterisks. For more detailed statistical output, please see the appendices. Abbreviations: BCG = bacille Calmette-Guérin; DPT = diphtheria-pertussis-tetanus; Penta = Pentavalent (DPT-HepB-Hib); HepB = hepatitis B (birth dose); Meni = Meningitis; YF = Yellow Fever; MCV = Measles-containing vaccine; OPV = oral polio vaccine; IPV = inactivated polio vaccine; HBR = Home-based records (Vaccination card, booklet, or other).
aBasic antigens: BCG, three doses of pentavalent (DPT-HepB-Hib), a complete series of either IPV or OPV (excluding polio vaccine given at birth) or a combination of at least one each of IPV and OPV, and one dose of measles vaccine.
bNigeria immunization schedule: BCG, HepB (birth dose), three doses of DPT-HepB-Hib (pentavalent), four doses of OPV, two doses of IPV, three doses of pneumococcal vaccine, three doses of rotavirus, one dose of measles vaccine, one dose of yellow fever vaccine, and one dose of meningitis vaccine. The second measles dose is excluded as children become eligible to receive it at 15 months.
cTrue zero-dose children lacking evidence of any vaccine. Cases where the cargiver initially reports the child as having received vaccines (Q409) but is subsequently unable to recall which antigens were administered are recorded as zero-dose children.
dGavi operationalizes zero-dose children as those who have not received a DPT-containing vaccine.

Coverage of birth doses across the entire survey area was mixed. Bacille Calmette-Guérin (BCG) coverage was 65.6%, followed by 54.2% for the Hepatitis B birth dose, and 66.1% for Oral Polio Vaccine at Birth (OPV0). While a majority of infants received at least one of the birth vaccines, the lower coverage for Hepatitis B (HepB) suggests potential gaps in timely administration, particularly for doses requiring cold chain management and coordination at birth. The discrepancy between BCG and OPV0 versus HepB may reflect operational challenges in administering HepB within the critical 24-hour window after birth. If true, these explanations may suggest that strengthening birth services and early outreach could improve protection against early-life infections.

Coverage for vaccines administered between 6 weeks and 6 months of age, the above table shows a relatively consistent pattern across the survey area. For each of these doses, coverage levels cluster in the range of approximately 55% to 63%, with only modest attrition from the first to the third dose in multi-dose series. This suggests that once children are engaged with the immunization system in early infancy, a majority tend to receive the subsequent doses in a timely manner. The consistency across these antigens likely reflects routine contact points during early infancy and the bundling of multiple vaccines at the same visit.

In contrast, coverage begins to decline sharply for vaccines administered at or after 9 months of age. Yellow Fever (YF) coverage stands at just 46.3%, and meningitis coverage is even lower at 45.7%, representing a significant drop compared to earlier vaccines.¹ The drop-off may be due to fewer scheduled contact points beyond 6 months, weaker demand or recall systems for later doses, and a potential lower emphasis of targeted outreach for the 9-month visit. This pattern raises concerns about missed opportunities for protection against high-burden diseases and highlights the need for programmatic attention to sustaining vaccination engagement beyond early infancy.

Turning to Table 4.52, we can also observe some stark differences in coverage patterns—as well as areas of ostensible parity—by background characteristic. The sex of the child does not appear to have a significant bearing on the rate of vaccination, and a \(t\)-test fails to find a statistically significant difference between the sexes for Penta-1 prevalence \((d=-3.0\), \(t_{(498)}=-1.72\), \(p=.087)\). The birth order likewise shows trivially small and non-significant differences in vaccination patterns across the different antigens and doses in a design-based Wald test \((F_{(3, 496)}=0.3\), \(p=.791)\).

Table 4.52: Vaccination coverage by antigen, dose, and background characteristic

	Infants 12-23 months																									15-23 months
	BCG	HepB	OPV				Penta			IPV		Pneumococcal			Rotavirus			MCV1	YF	Meni.	Fully vaccinated		Zero-dose		Sampled # of children
			0	1	2	3	1	2	3	1	2	1	2	3	1	2	3				B.A.a	N.I.S.b	Truec	Gavid		MCV2
Sex of child
Girls	67.5	55.6	67.9	65.5	62.6	58.6	64.3	63.0	61.0	63.5	54.6	64.2	63.0	60.3	63.9	62.5	59.1	60.0	48.1	47.6	58.9	36.1	30.9	35.7	5,366	53.8
Boys	63.7	52.9	64.3	62.6	59.9	56.0	61.3	59.2	57.7	60.2	51.5	61.2	59.2	57.1	60.7	58.7	55.4	56.4	44.6	43.8	55.3	33.4	34.5	38.7	5,529	49.4
Birth order
1	65.9	55.8	67.0	64.7	62.4	57.4	63.1	61.5	59.5	61.9	52.2	63.0	61.6	59.1	62.7	61.1	57.5	57.9	45.3	44.5	56.7	34.7	32.0	36.9	5,956	50.9
2-3	65.6	54.5	65.2	64.0	60.3	57.4	63.5	60.6	59.0	62.7	55.2	63.4	60.5	58.1	62.8	60.1	57.9	59.0	47.5	47.4	57.6	37.3	33.7	36.5	1,862	55.1
4-5	64.3	51.4	64.7	61.4	59.3	56.3	60.8	60.3	59.9	60.3	53.3	61.0	60.2	59.0	60.5	59.9	56.6	58.6	48.9	48.3	58.4	34.3	34.3	39.2	1,221	51.5
6+	64.8	51.2	64.5	63.1	60.4	57.6	62.1	60.4	59.4	61.2	53.5	61.9	60.2	58.0	62.0	60.0	55.7	58.4	47.3	46.0	57.2	33.2	33.5	37.9	1,264	50.3
Degree of urbanization
City	76.0	70.3	75.7	72.4	68.8	63.3	71.9	69.6	66.8	70.6	57.6	71.7	69.4	65.8	71.2	69.0	64.8	65.5	46.7	45.9	63.4	40.3	22.8	28.1	2,926	59.2
Town	68.8	54.1	70.3	68.2	65.4	62.8	65.8	64.8	63.6	65.9	58.2	65.8	64.8	62.8	65.1	64.0	61.2	63.4	51.9	52.1	62.2	38.1	28.3	34.2	2,802	56.9
Suburban	62.2	48.8	62.7	61.2	59.2	56.2	59.7	58.2	56.8	58.5	50.9	59.7	58.3	56.7	59.3	57.8	54.8	55.5	47.1	46.2	54.8	32.7	36.5	40.3	3,775	47.2
Village	50.3	38.0	50.9	50.0	47.2	41.7	49.1	47.4	45.9	48.1	42.7	49.1	47.4	45.2	49.5	47.3	44.0	44.3	36.8	35.7	44.2	24.9	48.2	50.9	1,392	39.1
Card status
HBR seen	99.9	81.6	99.1	97.9	96.5	93.9	97.8	96.2	94.3	96.8	85.7	97.7	96.2	94.2	97.5	95.8	91.1	92.0	82.5	82.0	91.4	62.6	0.0	2.2	6,178	84.6
HBR not seen	80.1	72.8	77.3	68.8	55.7	37.9	68.3	63.4	58.1	63.4	45.3	67.4	62.8	50.6	64.8	59.7	55.0	56.6	6.8	4.4	53.0	2.8	18.6	31.7	669	44.1
Never had HBR	14.4	12.2	17.2	15.0	12.1	8.2	11.9	10.6	9.8	11.6	7.9	12.0	10.7	9.5	11.8	10.6	9.4	10.3	*	0.7	8.9	0.4	81.8	88.1	4,048	9.9
Highest education level completed (caregiver)
No education	58.0	45.2	58.8	57.3	55.0	51.3	55.5	54.3	53.1	54.9	48.1	55.5	54.3	52.6	55.3	54.0	51.3	52.0	42.5	41.6	51.4	30.1	40.0	44.5	7,294	45.8
Primary	71.8	56.8	71.9	69.2	65.2	61.3	69.2	65.7	62.9	67.2	53.5	68.8	65.6	61.9	67.6	64.5	60.2	61.9	47.3	47.8	59.8	32.1	26.7	30.8	1,131	52.9
Secondary	81.0	73.9	80.9	77.5	74.2	69.1	76.8	74.3	71.4	75.5	63.6	76.6	74.3	71.0	76.1	73.6	69.0	70.3	55.0	54.7	68.3	46.7	18.1	23.2	2,119	62.6
More than secondary	86.6	84.6	85.9	84.6	81.1	76.3	85.2	84.8	83.4	84.2	72.1	85.1	84.8	80.3	84.8	84.3	80.8	81.1	56.3	55.2	79.9	49.5	13.3	14.8	351	73.5
Place of delivery
Health facility	81.0	71.7	80.4	77.3	74.9	69.2	76.9	74.8	72.0	75.3	63.7	76.5	74.7	71.0	76.4	74.5	70.3	70.4	53.0	52.2	69.2	43.1	18.1	23.1	2,299	61.5
Elsewhere	61.5	49.7	62.3	60.5	57.7	54.1	59.0	57.5	56.0	58.2	50.2	59.0	57.5	55.5	58.6	56.9	53.8	55.0	44.5	43.9	53.9	32.6	36.6	41.0	8,596	49.0
Pathways type
U1-NN	91.8	87.7	90.6	88.5	84.2	75.9	88.8	85.6	79.8	85.9	67.6	88.0	85.4	79.1	87.9	84.9	77.4	81.3	57.7	56.7	77.8	48.8	8.2	11.2	360	72.9
U3-NN	79.3	74.1	78.4	76.0	72.2	67.2	75.7	74.0	71.5	74.2	62.8	75.1	73.9	70.3	75.1	73.6	70.1	70.3	51.9	50.9	68.3	46.4	20.4	24.3	1,328	64.5
U2-NN	79.4	74.4	82.9	79.7	77.2	74.3	77.2	73.7	72.3	75.4	60.3	76.7	73.2	70.5	75.6	72.0	69.8	66.9	52.5	50.3	65.6	43.2	15.8	22.8	354	55.7
R2-NN	69.5	58.8	70.1	66.7	64.0	59.9	65.3	63.3	61.5	65.2	57.0	65.6	63.3	61.2	64.9	62.5	59.2	60.9	47.6	49.1	59.5	37.4	29.3	34.7	1,214	53.1
R3.1-NN	65.7	51.0	65.3	64.9	62.3	58.8	63.6	62.6	61.1	62.2	54.2	63.8	62.8	60.5	63.0	61.8	58.2	59.9	47.1	45.8	58.6	33.3	33.1	36.4	1,367	50.3
R3.2-NN	60.5	46.0	60.6	59.4	57.0	54.0	58.4	57.1	56.2	57.8	51.4	58.4	57.1	55.6	58.1	56.8	53.9	54.8	46.4	45.3	54.4	31.8	38.2	41.6	4,106	47.5
U4-NN	62.3	51.7	64.5	60.7	58.1	52.9	57.8	55.9	53.5	56.5	45.3	57.9	56.0	53.0	57.5	55.7	52.0	53.1	41.1	40.8	51.6	30.7	34.3	42.2	1,430	48.6
R4-NN	43.5	32.2	46.1	44.2	42.0	38.2	42.5	40.9	39.6	42.7	37.3	42.5	40.9	39.5	42.3	40.7	38.8	37.8	31.7	32.1	37.7	21.9	53.5	57.5	598	33.4
Ethnic group
Hausa	67.6	55.9	67.9	65.9	62.9	59.4	64.7	62.9	61.3	63.9	55.2	64.6	62.9	60.7	64.3	62.5	59.3	60.0	48.3	47.9	58.9	36.4	31.0	35.3	8,552	52.7
Other	71.7	66.5	71.6	64.5	63.5	57.3	64.9	64.4	63.8	64.1	51.9	64.3	63.6	60.8	64.4	63.6	60.8	62.5	42.7	42.1	61.8	35.0	28.3	35.1	225	58.7
Fulani	54.8	44.6	56.3	54.7	53.0	46.7	52.7	51.6	49.1	51.1	42.6	52.7	51.6	48.5	52.2	50.9	46.3	48.5	37.0	34.8	47.4	26.3	42.4	47.3	2,118	45.0
Religious affiliation
Muslim	65.3	54.0	65.8	63.8	61.0	57.0	62.5	60.8	59.1	61.6	52.9	62.4	60.8	58.4	62.1	60.4	57.0	58.1	46.3	45.7	56.9	34.8	33.0	37.5	10,823	51.4
Other	88.0	*	89.3	87.5	87.1	84.5	86.0	85.7	85.7	81.6	*	85.8	85.4	82.6	83.8	82.5	81.2	*	49.4	41.3	*	28.9	*	14.0	67	62.0
Wealth quintile
Poorest	55.6	42.6	56.0	56.3	53.8	50.7	54.7	53.1	51.7	53.7	45.7	54.8	53.2	51.3	54.3	52.4	49.0	50.1	41.5	39.3	49.6	28.5	42.7	45.3	2,981	43.0
Poorer	57.2	45.1	58.3	55.8	53.1	49.8	53.6	52.6	51.5	53.3	46.2	53.7	52.7	51.3	53.4	51.9	50.1	50.0	40.3	39.8	49.2	28.5	40.1	46.4	2,115	43.3
Middle	63.3	48.1	64.4	62.9	60.9	57.2	61.1	59.6	58.5	59.9	53.6	60.8	59.4	57.8	61.0	59.6	56.4	58.0	48.1	47.8	57.6	33.3	34.5	38.9	2,114	50.9
Richer	67.6	57.3	68.4	65.8	62.9	58.7	64.3	62.3	60.8	63.9	55.3	64.6	62.5	60.0	63.8	62.0	59.4	59.2	48.5	48.8	57.6	38.3	30.9	35.7	1,742	51.2
Richest	82.8	77.0	81.8	78.0	74.3	68.5	78.6	76.4	72.9	77.0	63.0	78.2	76.2	71.8	77.7	75.5	70.0	72.1	51.9	51.1	70.0	44.2	16.9	21.4	1,943	67.1
All	65.5	54.2	66.0	64.0	61.2	57.2	62.7	61.0	59.3	61.8	53.0	62.6	61.0	58.6	62.3	60.5	57.2	58.1	46.3	45.6	57.0	34.7	32.8	37.3	10,895	51.5
Note: Children are considered to have received the vaccine if it was either written on the child’s vaccination card or reported by the caregiver. For children whose vaccination information is based on the caregiver's report, date of vaccination is not collected. The proportions of vaccinations given during the first and second years of life are assumed to be the same as for children with a written record of vaccination. Unreliable proportions have been suppressed and replaced with asterisks. For more detailed statistical output, please see the appendices. Abbreviations: BCG = bacille Calmette-Guérin; DPT = diphtheria-pertussis-tetanus; Penta = Pentavalent (DPT-HepB-Hib); HepB = hepatitis B (birth dose); Meni = Meningitis; YF = Yellow Fever; MCV = Measles-containing vaccine; OPV = oral polio vaccine; IPV = inactivated polio vaccine; HBR = Home-based records (Vaccination card, booklet, or other).
aBasic antigens: BCG, three doses of pentavalent (DPT-HepB-Hib), a complete series of either IPV or OPV (excluding polio vaccine given at birth) or a combination of at least one each of IPV and OPV, and one dose of measles vaccine.
bNigeria immunization schedule: BCG, HepB (birth dose), three doses of DPT-HepB-Hib (pentavalent), four doses of OPV, two doses of IPV, three doses of pneumococcal vaccine, three doses of rotavirus, one dose of measles vaccine, one dose of yellow fever vaccine, and one dose of meningitis vaccine. The second measles dose is excluded as children become eligible to receive it at 15 months.
cTrue zero-dose children lacking evidence of any vaccine. Cases where the cargiver initially reports the child as having received vaccines (Q409) but is subsequently unable to recall which antigens were administered are recorded as zero-dose children.
dGavi operationalizes zero-dose children as those who have not received a DPT-containing vaccine.

Other background characteristics in Table 4.52 show much stronger differences.

Not surprisingly, children in urban areas tend to have a consistently higher rate of vaccination than their peers in rural areas \((d=-10.9\), \(t_{(498)}=-4.12\), \(p<.001)\). This is further confirmed by examining how the observed vaccination prevalence generally goes up as the degree of urbanization² increases. This pattern may be related to observed difference in vaccination based on the delivery location of the child, with infants born in non-clinical settings being showing a considerably lower rate of vaccination compared to those delivered in a health facility \((d=-17.9\), \(t_{(498)}=-7.86\), \(p<.001)\).

Other characteristics also exhibit strong association with vaccination trends. The education level of the caregiver—typically the mother—is strongly predictive of the vaccination status of a given child. For example, the survey finds that infants in households where the caregiver has no education have a 55% chance of being Penta-1 vaccinated, compared to 85.3% for households where the mother attended post-secondary education \((d=-21.3\), \(t_{(498)}=-9.93\), \(p<.001)\). Association with ethnic groups also shows some notable differences, with ethnic Fulanis exhibiting a 10-point lower vaccination rate compared to Hausa households \((d=-11.9\), \(t_{(498)}=-4.61\), \(p<.001)\). Infants in non-Muslim households were found to have a Penta-1 coverage rate that was 24 percentage points higher compared infants in Muslim households \((d=23.5\), \(t_{(498)}=3.93\), \(p<.001)\). Many of these dimensions are captured by the Pathways type, discussed more in depth in Section 4.2.3.5.

Vaccination coverage also associated with the socio-economic status of the household \((F_{(4, 495)}=21.9\), \(p<.001)\). Using wealth quintile estimates created via Principal Components Analysis (PCA), the survey found that the Penta-1 vaccination coverage increases with a household’s wealth: Whereas the wealthiest 20% of households have an average Penta-1 vaccination rate of 78%, the poorest 20% have a considerably lower rate at 56%.

Figure 4.2 presents the correlations between individual vaccine doses, along with three indicator variables: anyvax, fv_basicand fv_full. These respectively denote children who have received at least one vaccine (i.e., the complement of truly zero-dose children), children who have received all basic antigens defined in Table 4.51, and fully-vaccinated children according to the Nigeria immunization schedule, excluding Vitamin A or the second dose of Measles-Containing Vaccine (MCV)³. All estimated correlations are strongly positive with design-based 95% confidence interval bands that exclude zero. The consistently high correlations suggest a clear pattern: children tend to be either largely up to date with their vaccinations or missing several doses. Two birth doses—OPV0 and BCG—stand out as the strongest predictors of a child’s zero-dose status, with correlations of 0.97 and 0.96, respectively. Penta-1 (Diphtheria-Pertussis-Tetanus (DPT)-1), a commonly used proxy indicator, also shows a strong association (0.91), though slightly weaker than the birth doses. By contrast, the HepB birth dose is less strongly correlated with other vaccines, a pattern also observed for YF and meningitis vaccines, both administered after nine months of age. Interestingly, the first measles dose, also given at nine months, demonstrates stronger correlations with other vaccinations than these two later vaccines. Early and mid-schedule vaccines such as DPT-1, Pneumococcal Conjugate Vaccine (PCV)-1, Oral Polio Vaccine (OPV)-1, Rotavirus (RV)-1, DPT-2, and PCV-2 are exceptionally tightly linked, with correlations ranging from 0.95 to 0.99. A part of the explanation is that this reflects opportunities for simultaneous administration, whereby children receiving one of these vaccines often receive several others during the same visit.

Figure 4.2: Design-based Pearson correlation coefficients between vaccinations (card or recall)

4.2.3.2 Crude vs. Valid

The comparison between crude and valid coverage highlights important discrepancies in the reliability and timeliness of reported vaccination uptake. Crude coverage figures– which include both documented and caregiver-reported doses regardless of timing– can overstate immunization levels for several vaccines if caregiver reports are flawed. Table 4.53 compares crude and valid coverage estimates for vaccines in the Nigeria immunization schedule, highlighting the extent to which apparent coverage drops when accounting for documentation and timeliness criteria. Across all antigens, valid coverage is consistently lower than crude coverage, with the largest discrepancies observed for birth and early infancy doses such as OPV0 (66% crude vs. 26.7% valid, a 39.3 percentage point gap) and HepB (54.2% vs. 26%).

This pattern suggests that a substantial portion of reported or recorded vaccinations either occurred outside of the recommended window or lacked adequate documentation. As children progress through the schedule, the gap between crude and valid coverage narrows. By the time of DPT-1, valid coverage reaches 51.9% compared to 62.7% crude, and by DPT-3 the gap reduces further. This trend may reflect improvements in both timeliness and record-keeping as children engage more consistently with routine services. Still, the persistent gaps—even for later doses like MCV-1 and YF—underscore the importance of strengthening both timely vaccine delivery and documentation practices to more accurately reflect true immunization system performance.

Table 4.53: Comparison of valid vs. crude coverage by antigen and dose (95% CIs shown), sorted by magnitude of difference in methods.

	Crude coverage: card or recall	Valid coverage: card only	Difference
OPV0	66.0% (63.7; 68.3)	26.7% (24.4; 29.3)	-39.3
HEPB	54.2% (51.6; 56.7)	26.0% (23.7; 28.5)	-28.2
RV3	57.2% (54.6; 59.8)	30.7% (28.3; 33.2)	-26.5
DPT3	59.3% (56.6; 61.9)	35.1% (32.7; 37.6)	-24.2
OPV3	57.2% (54.5; 59.8)	33.1% (30.9; 35.3)	-24.1
PCV3	58.6% (56.0; 61.2)	34.9% (32.5; 37.3)	-23.8
IPV2	53.0% (50.3; 55.7)	30.2% (28.0; 32.6)	-22.8
MCV2	51.5% (48.3; 54.7)	34.5% (31.3; 37.8)	-17.0
IPV1	61.8% (59.2; 64.3)	47.1% (44.0; 50.2)	-14.7
RV2	60.5% (57.9; 63.1)	46.6% (43.6; 49.6)	-13.9
MENI	45.6% (42.7; 48.6)	32.5% (30.2; 34.8)	-13.2
MCV1	58.1% (55.5; 60.8)	45.2% (42.4; 48.1)	-12.9
OPV2	61.2% (58.7; 63.7)	48.5% (45.6; 51.4)	-12.7
BCG	65.5% (63.1; 67.9)	52.9% (49.9; 55.9)	-12.6
PCV2	61.0% (58.4; 63.6)	48.6% (45.6; 51.6)	-12.5
DPT2	61.0% (58.4; 63.6)	48.8% (45.8; 51.8)	-12.3
OPV1	64.0% (61.5; 66.4)	52.0% (48.9; 55.0)	-12.0
PCV1	62.6% (60.1; 65.2)	51.8% (48.8; 54.8)	-10.8
DPT1	62.7% (60.1; 65.2)	51.9% (48.9; 54.9)	-10.8
RV1	62.3% (59.6; 64.8)	51.6% (48.6; 54.5)	-10.7
YF	46.3% (43.5; 49.1)	36.9% (34.3; 39.5)	-9.4

4.2.3.3 Zero-Dose Children

Figure 4.3, Figure 4.5, and Figure 4.4 present the estimated prevalence and number of zero-dose children—those who have not received the first dose of the pentavalent vaccine (Penta-1)—across various domains in the surveyed areas in Kano State. The vertical dashed lines mark key reference points: the yellow line at 42.2% represents the 2023–24 Demographic and Health Survey (DHS) estimate for Kano, while the green line at 20% corresponds to an aspirational target for 80% Penta-1 coverage (that is, 20% zero-dose or lower). Observed zero-dose rates vary widely across domains, with substantial disparities evident by geography, education, and health access indicators. Non-sentinel areas and certain LGAs like Gaya and Gabasawa hover near or below the DHS benchmark, with Nassarawa notably at 25%.

Figure 4.3: Zero-dose prevalence by survey area

Urban–rural disparities persist, but they are less pronounced than other factors, with rural areas showing a 41% zero-dose rate versus 30% in urban areas. Based on the study, infants found in rural areas were 1.4 times as likely to be found to be zero-dose compared those in urban areas \((F_{(1, 499)}=15.7\), \(p<.001)\). Differences become more pronounced when disaggregating by the full degree of urbanization, with children in villages and sparsely inhabited areas being 1.8 times as likely as children in cities to be zero-dose \((F_{(1, 499)}=42.1\), \(p<.001)\).

By comparison, domains of interest such as sex of the child and birth order do not associate strongly with the zero-dose status. The design-based chi-square tests on these associations did not reveal statistically significant differences (\(F_{(1, 499)}=2.9\), \(p=.088\) for the sex of the child; \(F_{(2.5, 1254.3)}=0.3\), \(p=.825\) for the birth order).

Maternal education also shows a strong and statistically significant association: infants whose mothers had received no schooling whatsover were 1.8 times more likely to be zero-dose than children whose mother had a minimum of primary school \((F_{(1, 499)}=88.7\), \(p<.001)\). The delivery location and ethnic minority status also associated with a higher zero-dose prevalence above the statistically significant threshold. Children born outside of a health facility were 1.8 times as likely to be found to be zero-dose \((F_{(1, 499)}=52.2\), \(p<.001)\), and infants in Fulani households were 1.3 more likely than infants in Hausa households \((F_{(1, 499)}=20.8\), \(p<.001)\) be Penta-1 unvaccinated. These findings highlight where immunization programs might focus to close gaps—especially among vulnerable pathway segments, households without vaccination records, and mothers with limited education.

Figure 4.4: Estimated number of zero-dose children across the 15 LGAs included in the study.

Households’ access to and presentation of HBRs is strongly associated with zero-dose prevalence. Among children in households where an HBR was never available—relying entirely on caregiver recall—the estimated Penta-1 zero-dose prevalence is 64%, indicating a substantial risk of missing routine immunization services. At the other end of the spectrum, when HBRs were seen by field teams, the Penta-1-based zero-dose prevalence drops sharply to just 2%. While not strictly zero, this low figure suggests that children in these households likely received multiple vaccinations, even if Penta-1 was not recorded. Between these two extremes is a group of households where HBRs were reported to exist but were not seen by enumerators during the interview. This group exhibits an intermediate zero-dose prevalence of 32%, reflecting a mixed profile that may include both children with undocumented vaccinations and those truly unvaccinated. Details on the availability of HBRs is further discussed in Section 4.2.2.

Figure 4.5: Zero-dose prevalence by domain of interest

Figure 4.6 compares different approaches to estimating zero-dose prevalence—specifically, children who have received no vaccines by a given age—using data from 15 LGAs in Kano. The yellow line shows the true zero-dose curve based on whether a child had received any vaccine (reported via card or recall), while the green line captures zero-dose status using only Penta-1 as a proxy (Gavi definition).⁴ These are overlaid with a black line representing the ratio between the two approaches. After an initial period of divergence, the ratio stabilizes: by six months, the Penta-1 proxy consistently captures about 88% of true zero-dose children. This means that while Penta-1 is not a perfect proxy, it provides a reliable approximation for identifying zero-dose status beyond early infancy.

Figure 4.6: Probability of finding a zero-dose child, by categorization approach

The shape of the curves in Figure 4.6 also reveals a key dynamic in vaccination uptake. Most of the decline in the probability of being zero-dose occurs by 14 weeks—the window for timely receipt of Penta-1 and other early childhood vaccines. After this point, the curves flatten, and by six months, the proportion of children remaining zero-dose becomes relatively stable. This suggests that most zero-dose children are effectively determined in the early postnatal period, with few corrections (catch-up vaccinations) occurring later. Consequently, expanding surveillance or estimation beyond the commonly used 12–23 month age group to include children aged 6–11 months offers meaningful statistical advantages without compromising much in terms of representativeness.

From a practical standpoint, expanding the analytic window to include 6–23-month-olds yields a roughly 50% increase in sample size, which directly improves statistical power and precision. While 12–23 months remains a standard reporting group, the minimal change in zero-dose prevalence after six months—as shown in the chart—means the 6–23-month population can serve as a robust alternative or supplement. Especially in contexts where screening for zero-dose status is logistically or financially intensive, this broader window can enhance efficiency, allowing health planners to make more confident estimates and detect more nuanced patterns across population subgroups.

4.2.3.3.1 Zero-Dose Hotspots

This section aims to identify geographic hotspots of zero-dose children. For this hotspot analysis, we define zero-dose as the absence of the first dose of pentavalent vaccine (Penta-1). To increase the precision of our spatial estimates, we expand the domain of analysis from children 12-23 months to include children aged 6–23 months. This larger sample improves precision and statistical power, particularly when identifying local hotspots within highly disaggregated geographic areas. As the zero-dose rates have been shown not to change materially after the first six months of life Figure 4.6, the hotspots identified on a 6-23 cohort are likely to be nearly identical to hotspots for the 12-23 cohort.

There are multiple ways to conceptualize and detect zero-dose hotspots, each capturing different facets of the problem. One approach focuses on prevalence, using indicator smoothing techniques⁵ to estimate the proportion of zero-dose children across geographic space. This method highlights areas with a high local burden, yet often identifies areas that are typically sparsely populated and may have far fewer zero-dose children compared to urban areas with modest zero-dose prevalence. An alternative approach examines the absolute number of zero-dose children by generating a smoothed density surface that estimates the number of zero-dose children per hectare. This more commonly draws attention to densely populated settings, such as urban centers, where even moderate prevalence rates can result in a large number of unvaccinated children.

As shown in Figure 4.7, the prevalence and count of zero dose children tends to be positively correlated at the ward level (\(r = 0.55\)): When zero-dose prevalence is high, the total number of children also tends to be high. While this pattern holds for ward-level aggregates, the reverse can sometimes be found when plotting highly granular maps, as sparsely populated areas typically have few zero-dose children but a high prevalence of zero dose children. The difference is explained by the fact that the ward-level estimates use a weighted approach where dense urban centers count for a lot more in the aggregate, whereas mapping approaches showcase many areas with fewer inhabitants.

This implies that these two perspectives–prevalence and absolute burden– can move together but do not necessarily coincide, and neither should be viewed as universally superior. Rather, they offer complementary insights for program planning and resource targeting: one highlights where the proportion of missed children is greatest, while the other shows where the total number of missed children is highest.

Figure 4.7: Ward-level relationship between zero-dose prevalence and absolute burden for children aged 6-23 months.

4.2.3.3.1.1 Gaya

A logical entry point for identifying zero-dose hotspots is a tabulation of the zero-dose prevalence at the administrative ward level. Focusing on the expanded age group of children aged 6–23 months yields estimates with moderate, yet acceptable, precision for comparing areas below the LGA level. Table 4.54 displays ward-level estimates for Gaya LGA, ranked by zero-dose prevalence from highest to lowest. The “Estimated Total” column shows the total number of children aged 6–23 months residing in each ward. Balan Ward stands out with the highest zero-dose prevalence at 55%. Shagogo Ward, though not as extreme, also exceeds the LGA average of 36%, with 44% of children unvaccinated with Penta-1. Conversely, wards such as Gaya South and Wudilawa demonstrate relatively high coverage, with substantially fewer zero-dose children.

Table 4.54: Prevalence of Penta-1 zero-dose children aged 6-23 months in Gaya LGA, by administrative ward

	Proportiona	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Balan	54.7%	±5.9	(48.7; 60.5)	49.7%	59.6%	1.47	2,500	±420	216	403
Shagogo	44.4%	±4.6	(39.9; 49.1)	40.6%	48.3%	1.30	3,135	±463	244	594
Kazurawa	38.7%	±5.7	(33.2; 44.6)	34.0%	43.6%	1.37	1,765	±322	163	385
Maimakawa	37.4%	±4.5	(33.0; 42.0)	33.7%	41.3%	1.27	2,770	±425	211	562
Gaya North	36.5%	±4.4	(32.2; 41.0)	32.9%	40.3%	1.70	3,503	±585	311	793
Gamarya	33.7%	±7.1	(27.0; 41.2)	28.0%	39.9%	1.31	1,027	±238	89	225
Gamoji	31.5%	±7.6	(24.5; 39.6)	25.5%	38.2%	1.63	1,051	±322	68	238
Kademi	29.6%	±4.8	(25.1; 34.6)	25.8%	33.8%	1.78	2,229	±468	182	631
Wudilawa	27.4%	±5.7	(22.1; 33.5)	22.9%	32.4%	1.25	999	±251	75	296
Gaya South	24.8%	±4.4	(20.7; 29.5)	21.4%	28.7%	1.17	1,214	±235	116	444
All	36.3%	±1.7	(34.6; 38.1)	34.9%	37.8%	1.53	20,195	±1,275	1,675	4,571
aProportion of zero-dose children out of all infants 6-23 months
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 9.335 (p<0.001)

The coverage map in Figure 4.8 presents a smoothed surface highlighting areas with high zero-dose prevalence. Dark purple regions indicate rates above 70%, often in sparsely populated areas. This map helps to identify local hotspots that may span across ward boundaries or exist as smaller pockets within them. In contrast, areas shown in lime green and yellow represent zones with high vaccination coverage—also exceeding 70%.

Notable high-coverage areas include the northern border of Gaya (south of the Kafi settlement), Niima Yamma, the eastern edge of Balan, and the southern part of Shagogo ward near the Alkalawa settlement. In the southeast of the LGA, coverage hotspots appear in Kazurawa ward (outside Gidan Geri) and in Maimakawa ward, particularly in the settlements of Hausawa Hadi, Hausawa Magaji Gari Mai Unguwa, and Kurta.

While these coverage hotspots are important, the zero-dose density map in Figure 4.9 provides a complementary view by identifying areas where large numbers of zero-dose children are concentrated. It highlights the neighboring settlements of Zango Layin Gidan Mai Unguwar and Unguwar Mahaukaci as significant clusters. Despite relatively high coverage rates, the high population density in these areas means that even a small proportion of zero-dose children translates into a substantial absolute number of unvaccinated children. Looking at the Lower Confidence Bound (LCB), none of the wards in Gaya LGA can confidently be said to have a zero-dose rate above 50% (a commonly-chosen threshold in Lot Quality Assurance Sampling (LQAS) studies).

Figure 4.10 combines the two perspectives of prevalence and absolute burden. Areas in green identify places where the Penta-1 prevalence is below 40%. Areas in yellow identify places with at least five zero-dose children per hectare.⁶

Coverage

Figure 4.8: Penta-1 vaccination coverage in Gaya LGA among children 6-23 months. The visualization uses an Indicator Kriging (IK) method to interpolate and smooth coverage estimates over the entire area of the LGA. Health facilities offering immunization shown in red.

Zero-Dose Child (ZDC) Density

Figure 4.9: Density of Penta-1 unvaccinated infants aged 6-23 months in Gaya LGA. Darker areas represent zones with a high concentration of zero-dose children, although these are often in areas that are already densely populated.

Hotspots

Figure 4.10: Zero-dose hotspots in Gaya LGA. Areas where the local Penta-1 vaccination rate is less than 40% are shown in green, whereas areas where there are 5+ Penta-1 zero-dose children per hectare are shown in yellow

4.2.3.3.1.2 Nassarawa

As a distinctly urban LGA, Nassarawa stands apart from the other two sentinel LGAs in many regards. As discussed in earlier sections, the area generally has a lower zero-dose rate. As shown in Table 4.55, all its administrative wards have a Penta-1 zero dose rate of 32% or lower, with some wards like Gama even reaching a particularly low prevalence of 11%. Hotoro North stands out as a ward that has a zero-dose prevalence that is on the higher end for the LGA, as well as being a ward with a markedly higher absolute count of zero-dose children aged 6-23 months (~3000). It should be noted that uncertainty around these ward-level estimates is materially higher because Nassarawa yielded a smaller sample compared to the other sentinel LGAs due to its lower birth rate and lower survey response rate. The Krigging techniques used in estimating the geographic maps also reveal that the spatial correlation in Nassarawa is virtually nil: past 10 meters, distance to neighbouring areas does not correlate any more or less to the zero-dose status of the child. Combined, this means that the spatial data is not as informative as other sentinel areas at estimating where zero-dose rates are particularly high or low in the LGA.

Table 4.55: Prevalence of Penta-1 zero-dose children aged 6-23 months in Nassarawa LGA, by administrative ward

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Dakata	31.1%	±11.1	(21.2; 43.1)	22.7%	41.1%	1.09	608	±252	23	77
Hotoro South	28.5%	±9.8	(19.8; 39.1)	21.0%	37.3%	1.16	640	±251	30	99
Tudun Murtala	27.2%	±6.6	(21.1; 34.4)	22.1%	33.1%	1.14	1,228	±360	58	200
Hotoro North	26.9%	±4.3	(22.8; 31.5)	23.5%	30.7%	1.11	3,063	±594	120	449
Kawaji	26.7%	±5.7	(21.4; 32.8)	22.2%	31.8%	1.06	1,681	±407	69	250
Gwagwarwa & Kaura Gojea	22.9%	±5.2	(18.2; 28.5)	18.9%	27.6%	1.31	1,694	±445	76	337
Tudun Wada & Giginyua	22.3%	±5.4	(17.4; 28.2)	18.1%	27.2%	1.30	1,933	±517	66	301
Gama & Gawunaa	10.7%	±5.4	(6.3; 17.4)	6.9%	16.1%	1.23	417	±225	18	158
All	24.3%	±2.1	(22.3; 26.5)	22.6%	26.1%	1.17	11,265	±1,113	460	1,871
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 2.85 (p=0.006)
aIn this table, administrative wards with small sample sizes have been collapsed with neighbouring wards.

Nevertheless, some spatial insights can be gleaned from a mapping exercise. Figure 4.11 suggest that on a relative basis, the prevalence of zero dose children is slightly lower in the eastern part of the LGA. The density map Figure 4.12 helps highlight neighborhoodss with large numbers of zero-dose children. We observe a strong concentration of zero-dose children per hectare in the heart of Kaura Goje and Gama settlements in the northern part of the LGA. Because the prevalence map is relatively uniform, this concentration in zero-dose child density is likely largely a function of the population density of these areas. The south-eastern part of Nassarawa also has a relatively higher-than-average spatial concentration of zero-dose children, particularly in Hotoro North and the easternmost part of Giginyu ward. Figure 4.13 combines these two perspectives into a single plot based on the arbitrary cutoffs of areas with less than 75% coverage and areas with a zero-dose density per hectare greater than 10 children.

Coverage

Figure 4.11: Penta-1 vaccination coverage in Nassarawa LGA among children 6-23 months. The visualization uses an Indicator Kriging (IK) method to interpolate and smooth coverage estimates over the entire area of the LGA. Health facilities offering immunization shown in red.

ZDC Density

Figure 4.12: Density of Penta-1 unvaccinated infants aged 6-23 months in Nassarawa LGA. Darker areas represent zones with a high concentration of zero-dose children, although these are often in areas that are already densely populated.

Hotspots

Figure 4.13: Zero-dose hotspots in Nassarawa LGA. Areas where the local Penta-1 vaccination rate is less than 75% are shown in green, whereas areas where there are 10+ Penta-1 zero-dose children per hectare are shown in yellow

4.2.3.3.1.3 Gabasawa

With an average zero-dose prevalence of 30.7%, Gabasawa LGA has wards that range within roughly 10 percentage points of this average. The wards of Tarauni, Yautar Arewa, Joda, and Yautar Kudu stand well above this LGA average, as shown in Table 4.56. Karmami ward and Zugachi ward have the lowest zero-dose prevalence, at 20.5% and 22.3% respectively. In terms of estimated absolute numbers, Yautar Kudu and Mekiya are likely to capture the most zero-dose children of all the ward in Gabasawa.

Table 4.56: Prevalence of Penta-1 zero-dose children aged 6-23 months in Gabasawa LGA, by administrative ward

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Tarauni	41.4%	±7.1	(34.5; 48.6)	35.6%	47.4%	1.37	948	±192	109	256
Yautar Arewa	40.1%	±5.8	(34.5; 45.9)	35.3%	45.0%	1.55	1,869	±405	186	436
Joda	38.4%	±5.3	(33.3; 43.8)	34.1%	43.0%	1.42	2,139	±380	182	465
Yautar Kudu	38.3%	±4.0	(34.3; 42.4)	34.9%	41.7%	1.24	2,647	±369	266	691
Mekiya	31.3%	±4.6	(26.9; 36.0)	27.6%	35.3%	1.67	2,548	±466	212	656
Garun Danga	27.3%	±4.7	(22.9; 32.2)	23.6%	31.4%	1.30	1,531	±303	130	460
Gabasawa	26.7%	±3.8	(23.1; 30.7)	23.6%	30.1%	1.09	1,431	±241	159	557
Zakirai	25.7%	±6.0	(20.1; 32.1)	21.0%	31.0%	0.93	475	±123	53	192
Yumbu	25.0%	±7.9	(18.0; 33.8)	19.0%	32.2%	2.39	880	±343	72	277
Zugachi	22.3%	±4.7	(18.0; 27.3)	18.6%	26.5%	1.36	1,121	±253	104	418
Karmami	20.5%	±4.0	(16.7; 24.8)	17.3%	24.1%	1.33	1,133	±240	101	514
All	30.7%	±1.6	(29.1; 32.3)	29.4%	32.1%	1.55	16,723	±1,100	1,574	4,922
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 7.748 (p<0.001)

Figure 4.14 shows that the north-western part of the LGA tends to have the lowest vaccination coverage. Some settlements with higher-than-average zero-dose prevalences include Bariya, Nassarawar Yarzabaina, Nassarawar Yauta, and Yautar Arewa. In other parts of Gabassawa, Badawa Danduwa settlement appears to have lower vaccination coverage, as well as Binoni in the south. However, many of the low-coverage areas appear to be in the rural areas surrounding villages, as shown by Figure 4.16. On the basis of geographic density, Karami and Kumbotso settlments are likely to have concentrations of unvaccinated children, as well as the stretch from Zugachi to Gunduwa. Gumawa in the center of the LGA also appears to have higher concentrations, as well as Kumbo, Mekiya, Ballagaza, and Nasarawa settlement (to the south of Tahade). Many of these areas correlate strongly with areas that have the highest population density in general.

Coverage

Figure 4.14: Penta-1 vaccination coverage in Gabasawa LGA among children 6-23 months. The visualization uses an Indicator Kriging (IK) method to interpolate and smooth coverage estimates over the entire area of the LGA. Health facilities offering immunization shown in red.

ZDC Density

Figure 4.15: Density of Penta-1 unvaccinated infants aged 6-23 months in Gabasawa LGA Darker areas represent zones with a high concentration of zero-dose children, although these are often in areas that are already densely populated.

Hotspots

Figure 4.16: Zero-dose hotspots in Gabasawa LGA. Areas where the local Penta-1 vaccination rate is less than 45% are shown in green, whereas areas where there are 2+ Penta-1 zero-dose children per hectare are shown in yellow

In Gabasawa LGA, children’s Penta-1 vaccination status shows signs of spatial clustering– children who live near each other are more likely to have similar vaccination outcomes. This spatial correlation extends up to about 1.6 kilometers, meaning that within this range, households tend to show more similar vaccination patterns the closer they are to each other. Beyond that distance, the relationship weakens, and vaccination status of other households has no greater bearing on the likelihood that a given household is zero-dose. At the same time, there’s still considerable variability even between neighbouring households (a high so-called nugget effect). This suggests that while location plays a role, other factors– such as household characteristics or community-specific dynamics– also influence whether a child is vaccinated.

4.2.3.3.1.4 Non-Sentinel Areas

The identification of non-sentinel hotspots proceeds in a different manner. In contrast to the three sentinel LGAs where buildings were sampled directly in the first stage, non-sentinel LGAs were preceded by an additional level of cluster sampling: gridded Enumeration Areas (EAs) were selected first, followed by sub-sampling of buildings. Although non-sentinel estimates at the aggregate level are approximately unbiased, fine-grained analysis is not appropriate for identifying geographic hotspots as a result of this different sampling approach. Indeed, sampled data for non-sentinel zones are highly clustered and some wards would have zero data if plotted on a geographic map. Consequently, the analysis of zero-dose hotspots in non-sentinel areas is limited to identifying sampled EAs with surprisingly high zero-dose prevalence.

To increase statistical power for detecting unusually high rates, Table 4.57 below expands the domain of analysis to all children aged 6–23 months. This boosts the sample size by roughly 50% in each EA. An unweighted one-sided binomial test is used to assess how likely it would be to find the recorded zero-dose prevalence at least as high as the one observed, under the null hypothesis that the EA has the same Penta-1 zero-dose rate as the weighted LGA average. This makes a simplifying assumptions that the data within the EA behave as though they were drawn as a simple random sample, and that the propensity for a child to be vaccinated is independent of all other children in the EA. An area is flagged if such the observed vaccination prevalence (or a lower prevalence) would occur by chance alone in fewer than 1 out of 10 repeated samples (i.e., a \(p\)-value below 0.1 in Table 4.57). Geographic coordinates are provided which identify the geometric median of all children assessed in the EA.

For example, looking at the first few rows of the table, we can see that EA #8894 is flagged within Gediya Ward in Sumaila LGA. Its weighted estimated Penta-1-based zero-dose prevalence is 96.8%, based on 21 children in the cluster out of 22 aged 6-23 months who showed no evidence of receiving a Penta-1 vaccine (card or recall). This rate is considered very high, given that the LGA-wide zero-dose rate is 39.6%. The adjusted p-values tells us we would expect such a result very infrequently—less than 0.1 percent of the time—if this cluster had a zero-dose rate equal to 39.6%. Because this is so far from the LGA average, this EA is flagged as a zone where there is unusually low coverage, and the coordinates (11.2703, 8.9393) tell us the latitude and longitude of this zone if we wanted to send teams to this area to target households with catch-up interventions.

Table 4.57: Non-sentinel enumeration areas with suprisingly high zero-dose rates (children aged 6-23)

Area	EA	Coordinates (Lat, Lon)	Weighted	Unweighted			p-value
			Prop.a	Prop.	ZD Casesd	ne	Unadjustedb	Adjustedc
Sumaila LGA			39.5%
Gediya	#8894	11.2703, 8.9392	96.8%	95.2%	20	21	<0.001	<0.001
Magami	#9020	11.203, 8.8566	91.1%	88.9%	16	18	<0.001	<0.001
Masu	#9069	11.2277, 8.8784	86.3%	91.7%	11	12	<0.001	0.001
Masu	#9101	11.145, 8.8327	73.9%	76.5%	13	17	0.002	0.006
Magami	#9008	11.2504, 9.0033	73.1%	76.2%	16	21	<0.001	0.003
Sitti	#9337	11.305, 8.7992	65.0%	63.0%	17	27	0.012	0.024
Garfa	#8739	11.5364, 8.9981	61.2%	56.0%	14	25	0.071	0.079
Kiru LGA			39.1%
Yalwa	#5382	11.4369, 8.2005	100.0%	100.0%	14	14	<0.001	<0.001
Yalwa	#5385	11.4567, 8.169	82.8%	94.1%	16	17	<0.001	<0.001
Yalwa	#5438	11.4586, 8.1982	61.7%	62.2%	23	37	0.004	0.009
Kumbotso LGA			22.8%
Unguwar Rimi	#7310	11.9096, 8.5687	62.7%	50.0%	5	10	0.055	0.065
Kureken Sani	#6717	11.8839, 8.5745	59.0%	58.1%	18	31	<0.001	<0.001
Naibawa	#6976	11.9063, 8.5399	52.3%	50.0%	7	14	0.024	0.038
Chiranchi	#5838	11.9296, 8.4746	51.7%	50.0%	6	12	0.036	0.049
Garun Gawa	#6606	11.9238, 8.4809	41.5%	42.9%	6	14	0.077	0.084
Takai LGA			40.8%
Kachako	#9709	11.5684, 9.3431	84.4%	75.8%	25	33	<0.001	<0.001
Takai	#9851	11.5451, 9.0805	78.2%	70.6%	24	34	<0.001	0.002
Gezawa LGA			40.3%
Gawo	#3783	12.0051, 8.7724	68.5%	68.6%	24	35	<0.001	0.003
Mesar Tudu	#3918	12.1325, 8.7427	67.7%	62.0%	31	50	0.002	0.005
Dawakin Kudu LGA			40.2%
Yan Barau	#1872	11.8603, 8.6294	88.4%	70.0%	7	10	0.056	0.065
Dosan	#1541	11.7908, 8.6908	79.6%	80.0%	12	15	0.002	0.006
Tsakuwa	#1749	11.7569, 8.6409	76.8%	71.4%	15	21	0.004	0.009
Dawaki	#1342	11.8616, 8.6054	61.9%	62.2%	23	37	0.006	0.012
Tsakuwa	#1750	11.7901, 8.6261	58.7%	58.8%	10	17	0.095	0.097
Bebeji LGA			48.4%
Gwarmai	#258	11.4882, 8.3487	64.9%	69.2%	18	26	0.026	0.040
Gwarmai	#230	11.4721, 8.2518	62.5%	62.5%	30	48	0.035	0.049
Kuki	#349	11.3777, 8.3508	59.5%	73.0%	27	37	0.002	0.006
Dambatta LGA			30.1%
Kore	#1087	12.36, 8.746	61.1%	52.9%	9	17	0.041	0.054
Saidawa	#1121	12.3885, 8.6312	48.2%	55.0%	11	20	0.017	0.032
Kore	#1053	12.4189, 8.7243	43.1%	42.4%	14	33	0.090	0.095
Dawakin Tofa LGA			26.3%
Dawanau	#2141	12.0854, 8.4335	49.2%	50.0%	9	18	0.027	0.040
Ganduje	#2228	12.2161, 8.4392	45.1%	52.9%	9	17	0.018	0.032
Tudun Wada LGA			50.6%
Sabon Gari	#10870	11.2548, 8.3641	69.1%	67.9%	19	28	0.049	0.063
Baburi	#10456	11.3481, 8.7209	68.4%	69.7%	23	33	0.020	0.036
Tarauni LGA			18.3%
Hotoro	#10192	11.9625, 8.5742	36.0%	35.7%	10	28	0.022	0.038
Ungogo LGA			27.6%
Rijiyar Zaki	#11680	11.9846, 8.4438	66.6%	54.5%	6	11	0.054	0.065
Gayawa	#11246	12.0508, 8.5665	55.2%	57.1%	4	7	0.097	0.097
aEA zero-dose prevalence: proportion of children aged 6-23 months who have not received a Penta-1 vaccine (card or recall).
bThe unadjusted p-value is the probability that at least x zero-dose children would be observed out of n cases in the EA, under the assumption that the cluster has a true zero-dose rate equal to the LGA-average, r. Mathematically, this is the binomial probability P(X ≥ x | n, r).
cThe adjusted p-value inflates the raw p-value to account for repeated testing, using the Benjami-Hochberg procedure that controls for the False Discovery Rate. Adjustment is applied after sub-setting cases to those where the unadjusted p-value is 0.1 or lower.
dNumber of Penta-1 zero-dose children found in the EA
eNumber of children aged 6-23 assessed in the EA

4.2.3.4 Fully-Immunized Children

The concept of a fully immunized child provides a standardized way to assess completion of essential childhood vaccinations for children 12-23 months old. It captures whether children have received the full sequence of key antigens required for protection against major vaccine-preventable diseases, and serves as a central indicator for monitoring routine immunization program performance.

The specific definition of fully immunized can vary depending on the set of antigens included and national program timelines. In this chapter, we begin by presenting the basic antigen approach, in which a child must receive at least:

• One dose of BCG vaccine;
• Three doses of polio vaccine given as OPV, Inactivated Polio Vaccine (IPV), or a combination of OPV and IPV⁷;
• Three doses of DPT-containing vaccine (pentavalent); and
• One dose of MCV

This contrasts with the stricter national schedule approach, which follows the Nigeria immunization schedule for children aged 12–23 months. To be considered fully-immunized under this standard, a child must have received:

• One dose of BCG
• One (birth) dose of HepB
• Three doses of DPT (pentavalent)
• Four doses of OPV
• Two doses of IPV
• Three doses of PCV vaccine
• Three doses of RV vaccine
• One dose of MCV vaccine
• One dose of YF vaccine
• One dose of meningitis vaccine

The 2023-24 NDHS–often used as a benchmark–did not include rotavirus immunizations in prevalences of calculations. This is because its data collection overlapped with the tail end of the RV roll-out in phases in August and October 2022. Since not all children age 12–23 months were eligible for RV at the time of the NDHS survey, their estimates excluded them. The current zero-dose study took place later and did not have this problem. For comparability with the NDHS, variants with and without RV are presented in Table 4.59 and Table 4.60. As the present only collected data on infants up to 23 months, our report does not present the NDHS fully vaccinated cohort for 24–35 months, which includes the second measles dose. Including rotavirus, the proportion of children fully vaccinated according to the national schedule is 34.8%.

Nassarawa LGA, in spite of having lowest proportion of zero-dose children out of the three sentinel LGAs, nevertheless has a far lower proportion of fully-immunized children (national schedule) than Gaya or Gabassawa. This is mainly explained by the fact that children in this state have high coverage for basic antigens, but much lower coverage for YF and meningitis.

Basic Antigens

Table 4.58: Proportion of children aged 12-23 months who have received all basic antigens (card or recall)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Nassarawa	66.2%	±2.7	(63.4; 68.9)	63.9%	68.5%	1.11	21,126	±1,644	834	1,283
Gabasawa	61.3%	±2.1	(59.3; 63.4)	59.6%	63.0%	1.43	21,165	±1,132	1,856	3,092
Gaya	58.7%	±2.2	(56.6; 60.9)	56.9%	60.5%	1.47	21,067	±1,121	1,712	2,912
Non-sentinel LGAs	56.0%	±3.2	(52.8; 59.2)	53.4%	58.7%	3.81	286,510	±27,614	2,071	3,608
All	57.0%	±2.7	(54.3; 59.6)	54.8%	59.2%	8.10	349,868	±27,771	6,473	10,895
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 11.2 (p<0.001)

Nigeria Immunization Schedule

With RV

Table 4.59: Proportion of children aged 12-23 months who have all vaccinations and doses described in the Nigeria Immunization Schedule (card or recall), excluding MCV2

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Non-sentinel LGAs	37.3%	±3.1	(34.3; 40.4)	34.8%	39.9%	3.75	190,755	±20,198	1,416	3,608
Gabasawa	25.8%	±1.9	(23.9; 27.7)	24.2%	27.4%	1.51	8,897	±757	758	3,092
Gaya	22.9%	±1.7	(21.2; 24.7)	21.5%	24.4%	1.29	8,216	±673	681	2,912
Nassarawa	16.4%	±2.3	(14.2; 18.8)	14.5%	18.4%	1.30	5,224	±802	206	1,283
All	34.7%	±2.6	(32.2; 37.3)	32.6%	36.9%	8.15	213,092	±20,206	3,061	10,895
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 74.198 (p<0.001)

Without RV

Table 4.60: Proportion of children aged 12-23 months who have all vaccinations and doses described in the Nigeria Immunization Schedule (card or recall), excluding MCV2 and Rotavirus (for comparability with NDHS 2023-24)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases	Domain Sample Size
Non-sentinel LGAs	38.0%	±3.1	(35.0; 41.2)	35.4%	40.7%	3.88	194,480	±20,465	1,444	3,608
Gabasawa	26.3%	±1.9	(24.4; 28.2)	24.7%	27.9%	1.50	9,066	±762	773	3,092
Gaya	23.2%	±1.7	(21.5; 25.0)	21.7%	24.7%	1.29	8,312	±677	688	2,912
Nassarawa	16.6%	±2.3	(14.4; 19.0)	14.7%	18.6%	1.29	5,292	±801	209	1,283
All	35.4%	±2.6	(32.8; 38.0)	33.2%	37.6%	8.40	217,150	±20,468	3,114	10,895
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Pearson's Chi-square test with a Rao & Scott adjustment: F = 76.33 (p<0.001)

The survey finds that 57% of children aged 12–23 months are fully vaccinated for basic antigens. This rate is considerably higher than the 35.8% for Kano state estimate from the 2023-24 NDHS. The explanations may be considered. First, the inferential population of this study only captures 15 LGAs out of 40 in Kano state. Second, the NDHS estimate may have more sampling error (only 271 children assessed). Similarly, the estimate of fully immunized children (without rotavirus) in this survey is 35%–considerably higher than the NDHS estimate of 13.6% for Kano. Finally, fully immunized child rates tend to fluctuate more than individual antigen coverage or zero-dose measures because they depend on a child receiving all required vaccines in the schedule. Even small shifts in coverage for a single antigen—such as a stockout, delayed introduction, or change in eligibility—can cause large swings in FIC estimates across time a space.

Figure 4.17: Proportion of fully-immunized children aged 12-23 months (card or recall)

Detailed disaggregations by different LGA are presented in Section A.1.2. Related insights on vaccination drop-out are also discussed in Section 4.2.4.

4.2.3.4.1 Predictors of Vaccination Status (Ordinal Ladder)

A binary fully-immunized indicator treats two very different children the same: the one who has missed every antigen and the one who has received every basic antigen but not the full national schedule. To capture the gradient between them, we built a four-level ordinal vaccination status, ordered from least to most immunized:

1. Zero-dose — no documented antigens
2. Partially immunized — at least one antigen, but not all basic antigens
3. Basic antigens — BCG, three DPT, full polio sequence, and MCV-1, but not the full national schedule
4. National schedule — the full Nigerian schedule (BCG, HepB birth dose, three DPT, full polio sequence, IPV, PCV, RV, MCV-1, YF, and meningitis)

We do not split the top level by MCV-2 status: MCV-2 is scheduled at 15 months, and many children in the 12–23 month cohort are not yet age-eligible. The analysis covers the full 12–23 month cohort.

The candidate pool has eight background characteristics: sex of child, birth order, caregiver education, household wealth quintile, settlement type, place of delivery, ethnic group, and religion. The complete-case sample is 10,300 children.

We use forward-then-backward stepwise selection on a design-weighted proportional-odds regression to pick the model with lowest Akaike Information Criterion (AIC).⁸

Table 4.61: Forward+backward stepwise selection path on the design-weighted proportional-odds regression. ΔAIC is the change in AIC from the action (negative values indicate the action improves AIC); the AIC-equivalent inclusion criterion is Wald χ² > 2·df.

Step	Action	Variable	Wald χ²	df	ΔAIC
1	Add	Wealth quintile	79.82	4	-71.82
2	Add	Caregiver education	30.61	3	-24.61
3	Add	Settlement type	16.12	3	-10.12
4	Add	Place of delivery	13.27	1	-11.27
5	Add	Ethnic group	9.15	2	-5.15
6	Add	Sex of child	5.39	1	-3.39
7	Add	Religion	2.04	1	-0.04
8	Drop	Wealth quintile	2.86	4	-5.14

Selection retained 6 of the 8 candidate predictors. Wealth quintile and birth order were dropped because they did not improve fit enough to be kept: wealth quintile is plausibly captured by caregiver education and settlement type — both retained — and birth order adds little once the other predictors are in.

Table 4.62: Design-based proportional-odds regression for the 4-level vaccination status ladder. Estimates are on the cumulative log-odds scale; positive values raise the odds of falling at or above the listed threshold (i.e., being more immunized).

Term	Estimate	exp(Estimate)	SE(Estimate)	z-statistic	p-value	FDR p-value
Threshold cutpoints (intercepts)
Zero-dose | Partially immunized	0.118	1.13	0.156	0.761	0.446
Partially immunized | Basic antigens	-0.339	0.71	0.157	-2.162	0.031
Basic antigens | National schedule	-1.289	0.28	0.161	-8.007	<0.001
Caregiver education
Primary	0.216	1.24	0.115	1.874	0.061	0.084
Secondary	0.676	1.97	0.114	5.922	<0.001	<0.001
More than secondary	0.877	2.40	0.202	4.339	<0.001	<0.001
Settlement type
Suburban	0.313	1.37	0.184	1.702	0.089	0.109
Town	0.603	1.83	0.181	3.329	<0.001	0.002
City	0.290	1.34	0.193	1.507	0.132	0.145
Place of delivery
Health facility	0.351	1.42	0.094	3.727	<0.001	<0.001
Ethnic group
Fulani	-0.273	0.76	0.104	-2.619	0.009	0.019
Other	-0.389	0.68	0.194	-2.009	0.045	0.070
Sex of child
Girls	0.165	1.18	0.070	2.358	0.018	0.034
Religion
Non-Muslim	0.281	1.32	0.198	1.420	0.156	0.156
Reference levels: Caregiver education = No education; Settlement type = Village; Place of delivery = Elsewhere; Ethnic group = Hausa; Sex of child = Boys; Religion = Muslim.
The reference categories describe a common low-coverage child: no caregiver education, born outside a health facility, Hausa Muslim, living in a village. For this child the model predicts a 47% chance of being zero-dose, 11% partially immunized, 20% at basic antigens, and 22% on the full national schedule.

Caregiver education and settlement type are the largest movers. Children whose caregivers completed at least secondary school sit higher on the ladder than those whose caregivers had no formal education; children in towns or cities sit higher than those in villages. Place of delivery moves children in the same direction — a facility birth is associated with a more complete schedule than birth elsewhere — and ethnic group adds further structure: Fulani and other-ethnic children sit lower on the ladder than Hausa children. Girls sit higher on the ladder than boys — a gap that is small in the raw marginals (boys 32% vs girls 31% zero-dose) but widens once survey weights are applied (35% vs 31%), and which the design-weighted selection step picks up. Religion is retained on AIC grounds, but its average effect is not statistically significant: the non-Muslim group is small (~1% of the sample), and the proportional-odds average hides the strongest threshold-by-threshold heterogeneity of any predictor (see Table 4.63). The directions match the bivariate patterns in Section A.1.2.

The estimates above should be read as averages across thresholds: the proportional-odds model forces each predictor’s effect to be the same at every rung, an assumption that does not hold strictly here.⁹ Table 4.63 shows what this averaging masks.

The three intercepts also reveal something about the ladder itself. The boundary between partially immunized and basic antigens sits very close to the boundary between zero-dose and partially immunized — less than half the distance from basic antigens to the full national schedule. The partially-immunized rung is therefore a narrow band: small shifts in a child’s circumstances can move them across it in either direction, while completing the final step from basic antigens to the national schedule is a substantially larger jump.

Several patterns emerge from the threshold-specific coefficients. The place-of-delivery effect is roughly three times larger at the zero-dose / partially-immunized cut than at the basic / national cut. A few coefficients (primary education, city settlements, religion) flip sign across the three thresholds, while sex of child and Fulani ethnicity are essentially constant — for those two, the proportional-odds approximation is closer to right.

Religion is the clearest case of threshold heterogeneity: being non-Muslim is strongly associated with climbing out of zero-dose (β ≈ +1.17 at that cut), but the advantage shrinks at higher rungs and reverses slightly at the final step (β ≈ −0.30). The proportional-odds model averages these into a small, near-null coefficient — exactly the pattern the assumption is built to hide.

Table 4.63: Threshold-specific coefficients from an exploratory unweighted partial-proportional-odds fit on the same six selected predictors. Each row reports the predictor’s effect at one of three cumulative thresholds; under the proportional-odds assumption all three numbers in a row would be identical. The fit is unweighted because the design-weighted variant does not converge on this sample.

Term	Zero-dose | Partially	Partially | Basic	Basic | National
Caregiver education
Primary	0.358	0.181	-0.114
Secondary	0.692	0.438	0.417
More than secondary	1.002	0.835	0.486
Settlement type
Suburban	0.133	0.182	0.122
Town	0.474	0.534	0.329
City	0.277	0.188	-0.112
Place of delivery
Health facility	0.515	0.336	0.169
Ethnic group
Fulani	-0.348	-0.366	-0.338
Other	-0.012	-0.007	-0.096
Sex of child
Girls	0.032	0.033	0.043
Religion
Non-Muslim	1.168	0.520	-0.303
Estimates are unweighted partial-proportional-odds coefficients on the cumulative log-odds scale; positive values raise the odds of falling at or above the listed threshold (i.e., being more immunized). Reference levels and predictor groupings match the headline table above.

4.2.3.5 Pathways Typing Tool

The Pathways Segmentation Tool for Northern Nigeria is a population grouping analysis framework that classifies households in eight types based on shared characteristics and associates them with varying degrees of risk of poor health outcomes (vulnerability) (Pathways n.d.).

The segments are classified using a decision tree that Pathways fit on a set of variables ranging from elements such as types of cooking fuel used, urban/rural designations,¹⁰ frequency of TV use, and many other factors.¹¹

Each type is associated with a code, as shown in the first column of Table 4.64. The “U” and “R” letters designate urban and rural households respectively, and the digit portion increases on a range from least vulnerable (1) to most vulnerable (4).

As shown by Table 4.64, 72.9% of the respondents in the survey areas belong to either the “R3.2” or “R2” types. Table 4.65, Table 4.66 and Table 4.67 also show that there are large differences in the frequency and distribution of types from one LGA to another.

All

Table 4.64: Distribution of Pathways types among households with children 0-23 months in all 15 LGAs

	Cumulative Proportion	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Medium-high vulnerability (R3.2)	37.5%	37.5%	±3.7	(33.9; 41.3)	34.5%	40.6%	28.79	414,862	±57,941	7,136
Medium-high vulnerability (U3)	51.1%	13.6%	±2.0	(11.7; 15.7)	12.0%	15.3%	16.49	150,036	±24,661	2,424
Higher vulnerability (U4)	64.6%	13.5%	±2.8	(11.0; 16.6)	11.4%	16.0%	33.27	149,496	±31,940	2,417
Lower vulnerability (R2)	77.1%	12.5%	±1.6	(10.9; 14.2)	11.2%	13.9%	12.37	138,106	±18,559	2,285
Medium-low vulnerability (R3.1)	85.6%	8.5%	±1.3	(7.2; 9.9)	7.4%	9.7%	11.51	93,574	±16,511	2,387
Higher vulnerability (R4)	90.8%	5.2%	±1.0	(4.3; 6.3)	4.4%	6.1%	10.05	57,664	±12,314	1,029
Lower vulnerability (U1)	95.4%	4.6%	±0.9	(3.8; 5.6)	3.9%	5.4%	9.59	50,772	±11,119	805
Medium-low vulnerability (U2)	99.3%	3.9%	±0.7	(3.2; 4.7)	3.3%	4.5%	7.20	42,701	±9,381	682
UNCLASSIFIED	100.0%	0.7%	±0.3	(0.5; 1.1)	0.5%	1.0%	5.13	8,140	±2,763	206
All		100.0%			100.0%			1,105,353	±85,311	19,371
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 19,371
Note: A small proportion of cases are labelled as "unclassified." These represent cases where respondents did not answer key questions necessary for classification into Pathways types. These represent 0.5% of all cases.

Gaya

Table 4.65: Distribution of Pathways types among households with children 0-23 months in Gaya

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Medium-high vulnerability (R3.2)	41.2%	±1.4	(39.8; 42.7)	40.1%	42.4%	1.09	24,713	±940	2,042
Higher vulnerability (U4)	17.2%	±0.9	(16.3; 18.1)	16.4%	17.9%	0.72	10,288	±559	891
Medium-low vulnerability (R3.1)	16.5%	±1.2	(15.3; 17.7)	15.5%	17.5%	1.31	9,867	±729	797
Lower vulnerability (R2)	9.1%	±0.9	(8.2; 10.1)	8.4%	9.9%	1.36	5,472	±569	427
Higher vulnerability (R4)	6.6%	±0.8	(5.8; 7.4)	5.9%	7.3%	1.32	3,941	±478	315
Medium-high vulnerability (U3)	5.2%	±0.6	(4.7; 5.9)	4.8%	5.8%	0.95	3,140	±360	286
Medium-low vulnerability (U2)	2.5%	±0.4	(2.1; 2.9)	2.1%	2.9%	1.04	1,475	±262	134
UNCLASSIFIED	1.0%	±0.3	(0.8; 1.4)	0.8%	1.3%	1.00	618	±179	63
Lower vulnerability (U1)	0.7%	±0.2	(0.5; 0.9)	0.5%	0.9%	1.01	406	±137	38
All	100.0%			99.9%	100.0%		59,921	±958	4,993
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 4,993
Note: A small proportion of cases are labelled as "unclassified." These represent cases where respondents did not answer key questions necessary for classification into Pathways types. These represent 0.5% of all cases.

Gabasawa

Table 4.66: Distribution of Pathways types among households with children 0-23 months in Gabasawa

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Medium-high vulnerability (R3.2)	51.2%	±1.5	(49.7; 52.8)	50.0%	52.5%	1.32	30,066	±1,019	2,766
Medium-low vulnerability (R3.1)	20.2%	±1.2	(18.9; 21.4)	19.1%	21.2%	1.36	11,829	±760	1,073
Lower vulnerability (R2)	16.1%	±1.1	(15.0; 17.2)	15.2%	17.1%	1.29	9,452	±669	875
Higher vulnerability (R4)	7.5%	±0.8	(6.8; 8.4)	6.9%	8.2%	1.30	4,420	±473	419
Higher vulnerability (U4)	2.6%	±0.3	(2.2; 2.9)	2.3%	2.8%	0.61	1,499	±193	149
UNCLASSIFIED	1.2%	±0.3	(0.9; 1.5)	0.9%	1.5%	1.22	695	±196	74
Medium-high vulnerability (U3)	0.6%	±0.2	(0.4; 0.8)	0.5%	0.8%	0.77	346	±105	39
Medium-low vulnerability (U2)	0.5%	±0.2	(0.3; 0.7)	0.4%	0.6%	0.80	283	±97	33
Lower vulnerability (U1)	0.2%	±0.1	(0.1; 0.3)	0.1%	0.3%	0.85	90	±56	10
All	100.0%			99.9%	100.0%		58,682	±945	5,438
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 5,438
Note: A small proportion of cases are labelled as "unclassified." These represent cases where respondents did not answer key questions necessary for classification into Pathways types. These represent 0.5% of all cases.

Nassarawa

Table 4.67: Distribution of Pathways types among households with children 0-23 months in Nassarawa

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Medium-high vulnerability (U3)	48.4%	±2.2	(46.2; 50.6)	46.5%	50.3%	1.21	28,023	±1,360	1,159
Higher vulnerability (U4)	20.7%	±1.8	(19.0; 22.6)	19.2%	22.3%	1.24	11,998	±1,096	475
Lower vulnerability (U1)	16.6%	±1.6	(15.1; 18.2)	15.3%	18.0%	1.09	9,617	±902	397
Medium-low vulnerability (U2)	13.1%	±1.6	(11.6; 14.8)	11.8%	14.5%	1.39	7,583	±1,069	286
UNCLASSIFIED	1.2%	±0.5	(0.8; 1.8)	0.8%	1.7%	1.30	679	±319	26
All	100.0%			99.8%	100.0%		57,901	±1,097	2,343
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 2,343
Note: A small proportion of cases are labelled as "unclassified." These represent cases where respondents did not answer key questions necessary for classification into Pathways types. These represent 0.5% of all cases.

Non-Sentinel

Table 4.68: Distribution of Pathways types among households with children 0-23 months in non-sentinel LGAs

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Medium-high vulnerability (R3.2)	38.8%	±4.3	(34.5; 43.2)	35.2%	42.5%	13.52	360,082	±58,177	2,328
Higher vulnerability (U4)	13.5%	±3.3	(10.5; 17.2)	11.0%	16.6%	16.01	125,711	±33,989	902
Lower vulnerability (R2)	13.3%	±2.0	(11.4; 15.4)	11.7%	15.0%	5.77	123,182	±18,619	983
Medium-high vulnerability (U3)	12.8%	±2.3	(10.6; 15.3)	10.9%	14.8%	8.31	118,527	±26,218	940
Medium-low vulnerability (R3.1)	7.7%	±1.6	(6.3; 9.5)	6.5%	9.2%	6.01	71,879	±16,549	517
Higher vulnerability (R4)	5.3%	±1.2	(4.2; 6.6)	4.4%	6.4%	4.74	49,303	±12,349	295
Lower vulnerability (U1)	4.4%	±1.1	(3.4; 5.6)	3.6%	5.4%	4.79	40,658	±12,238	360
Medium-low vulnerability (U2)	3.6%	±0.9	(2.8; 4.6)	2.9%	4.4%	3.66	33,360	±9,481	229
UNCLASSIFIED	0.7%	±0.3	(0.4; 1.1)	0.4%	1.0%	2.71	6,148	±3,055	43
All	100.0%			99.9%			928,850	±85,665	6,597
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 6,597

The analysis of vaccination coverage and zero-dose prevalence across Pathway types, shown in Table 4.69 and Figure 4.18, confirms that the rate of vaccination generally rises as the measure of vulnerability increases. Types labeled R4 and R3.2—designating higher-vulnerability rural types—report the highest prevalence (58% and 42%, respectively), while less-vulnerable urban segments like Under the Age of 1 (U1) and U2 report just 11% and 23%. Separate regression-based Wald tests on the urban and rural domains confirm that these ordinal associations are statistically significant (\(F_{(3, 496)}=20.8\), \(p<.001\) for infants in urban areas, \(F_{(3, 496)}=6.0\), \(p<.001\) for rural areas).

Table 4.69: Vaccination coverage by antigen, dose, and Pathways vulnerability type

	Infants 12-23 months																									15-23 months
	BCG	HepB	OPV				Penta			IPV		Pneumococcal			Rotavirus			MCV1	YF	Meni.	Fully vaccinated		Zero-dose		Sampled # of children
			0	1	2	3	1	2	3	1	2	1	2	3	1	2	3				B.A.a	N.I.S.b	Truec	Gavid		MCV2
Urban
Lower vulnerability (U1)	91.8	87.7	90.6	88.5	84.2	75.9	88.8	85.6	79.8	85.9	67.6	88.0	85.4	79.1	87.9	84.9	77.4	81.3	57.7	56.7	77.8	48.8	8.2	11.2	360	72.9
Medium-low vulnerability (U2)	79.4	74.4	82.9	79.7	77.2	74.3	77.2	73.7	72.3	75.4	60.3	76.7	73.2	70.5	75.6	72.0	69.8	66.9	52.5	50.3	65.6	43.2	15.8	22.8	354	55.7
Medium-high vulnerability (U3)	79.3	74.1	78.4	76.0	72.2	67.2	75.7	74.0	71.5	74.2	62.8	75.1	73.9	70.3	75.1	73.6	70.1	70.3	51.9	50.9	68.3	46.4	20.4	24.3	1,328	64.5
Higher vulnerability (U4)	62.3	51.7	64.5	60.7	58.1	52.9	57.8	55.9	53.5	56.5	45.3	57.9	56.0	53.0	57.5	55.7	52.0	53.1	41.1	40.8	51.6	30.7	34.3	42.2	1,430	48.6
Rural
Lower vulnerability (R2)	69.5	58.8	70.1	66.7	64.0	59.9	65.3	63.3	61.5	65.2	57.0	65.6	63.3	61.2	64.9	62.5	59.2	60.9	47.6	49.1	59.5	37.4	29.3	34.7	1,214	53.1
Medium-low vulnerability (R3.1)	65.7	51.0	65.3	64.9	62.3	58.8	63.6	62.6	61.1	62.2	54.2	63.8	62.8	60.5	63.0	61.8	58.2	59.9	47.1	45.8	58.6	33.3	33.1	36.4	1,367	50.3
Medium-high vulnerability (R3.2)	60.5	46.0	60.6	59.4	57.0	54.0	58.4	57.1	56.2	57.8	51.4	58.4	57.1	55.6	58.1	56.8	53.9	54.8	46.4	45.3	54.4	31.8	38.2	41.6	4,106	47.5
Higher vulnerability (R4)	43.5	32.2	46.1	44.2	42.0	38.2	42.5	40.9	39.6	42.7	37.3	42.5	40.9	39.5	42.3	40.7	38.8	37.8	31.7	32.1	37.7	21.9	53.5	57.5	598	33.4
All	65.5	54.2	66.0	64.0	61.2	57.2	62.7	61.0	59.3	61.8	53.0	62.6	61.0	58.6	62.3	60.5	57.2	58.1	46.3	45.6	57.0	34.7	32.8	37.3	10,895	51.5
Note: Children are considered to have received the vaccine if it was either written on the child’s vaccination card or reported by the caregiver. For children whose vaccination information is based on the caregiver's report, date of vaccination is not collected. The proportions of vaccinations given during the first and second years of life are assumed to be the same as for children with a written record of vaccination. Unreliable proportions have been suppressed and replaced with asterisks. For more detailed statistical output, please see the appendices. Abbreviations: BCG = bacille Calmette-Guérin; DPT = diphtheria-pertussis-tetanus; Penta = Pentavalent (DPT-HepB-Hib); HepB = hepatitis B (birth dose); Meni = Meningitis; YF = Yellow Fever; MCV = Measles-containing vaccine; OPV = oral polio vaccine; IPV = inactivated polio vaccine; HBR = Home-based records (Vaccination card, booklet, or other).
aBasic antigens: BCG, three doses of pentavalent (DPT-HepB-Hib), a complete series of either IPV or OPV (excluding polio vaccine given at birth) or a combination of at least one each of IPV and OPV, and one dose of measles vaccine.
bNigeria immunization schedule: BCG, HepB (birth dose), three doses of DPT-HepB-Hib (pentavalent), four doses of OPV, two doses of IPV, three doses of pneumococcal vaccine, three doses of rotavirus, one dose of measles vaccine, one dose of yellow fever vaccine, and one dose of meningitis vaccine. The second measles dose is excluded as children become eligible to receive it at 15 months.
cTrue zero-dose children lacking evidence of any vaccine. Cases where the cargiver initially reports the child as having received vaccines (Q409) but is subsequently unable to recall which antigens were administered are recorded as zero-dose children.
dGavi operationalizes zero-dose children as those who have not received a DPT-containing vaccine.

Figure 4.18: Zero-dose prevalence by Pathways type

4.2.4 Dropouts

RI_CONT_01

Infants who only received the first dose in a series and were age-eligible to receive the subsequent dose before the survey can be considered to have “dropped out” of the immunization program.

As shown in the vaccination coverage and timeliness charts in Section 4.2.5, dropout between two crude doses is fairly common among children aged 12–23 months in Kano State. This is especially evident in Figure 4.22, where coverage bars for later doses in a series (e.g., DPT-3) are consistently shorter than those for earlier doses (e.g., DPT-1).

Figure 4.19, Figure 4.20 and Figure 4.21 illustrate dropout patterns among children aged 12–23 months for vaccines typically administered during the first year of life. Notably, Nassarawa LGA consistently shows higher dropout rates across all VCQI continuity indicators.

Dropout from DPT-1 to DPT-3 is generally low across the board, with an overall rate of 5.4% when combining sentinel and non-sentinel LGAs (Figure 4.19). This suggests that an estimated 94.6% of children who start the DPT series eventually complete it. Gabasawa and Gaya LGAs stand out with slightly lower dropout rates of 3.4% and 2.6%, respectively, compared to Nassarawa and the non-sentinel LGAs combined.

The transition from DPT-3 to MCV-1 follows a similar pattern, where the overall dropout rate is 3.4% (Figure 4.20). Again, Nassarawa exceeds this with a dropout rate of 4.7%.

The highest recorded dropout rates appear in the transition between OPV-1 to OPV-3 series (Figure 4.21). The combined dropout rate for this series is 10.6%, with Gaya standing out as a lower outlier (6.1%), and Nassarawa recording the highest proportion at 15.9%.

DPT-1 to DPT-3

Figure 4.19: Dropout from DPT-1 to DPT-3

DPT-3 to MCV-1

Figure 4.20: Weighted dropout from DPT-3 to MCV-1

OPV-1 to OPV-3

Figure 4.21: Weighted dropout from OPV-1 to OPV-3

4.2.5 Vaccination Timeliness and Simultaneity

4.2.5.1 Vaccination Coverage and Timeliness

Figure 4.22 presents the estimated proportion of children with evidence of receiving each dose in the national immunization schedule, by timeliness and across all survey areas (see Appendix C for the analysis of non-sentinel LGAs and sentinel LGAs individually). The proportions combine evidence from HBR and caregiver recall.

Just over half of the vaccination evidence in this survey comes from HBRs, which provide dates needed to assess timeliness. Overall, cards were seen for only 54.7% of children aged 12–23 months. Children for which timelines cannot be determined fall into the “Timing Unknown” group (grey bars), either because the evidence comes from caregiver recall or, in some cases, an illegible or implausible date on the card.

The color-saturated part of each bar, to the left of the vertical HBR line, shows what we know about timeliness based on the child’s date of birth and recorded vaccination dates.

Coverage of birth doses, doses scheduled up to 14 weeks, and MCV-1 is fairly consistent, at around 60–65%. HepB (birth dose) and IPV-2 are exceptions, with lower coverage OPV0 and BCG exceed 65% coverage, with BCG showing the highest rate of timely vaccination.

Early vaccination (dark purple bars) is rare across all dose types and occurs at similar levels for non-birth doses.

Dropout is visible in every dose series, as the bars for later doses tend to be shorter than those for earlier doses. For OPV, for instance, 64.0% of children received the first dose, but 57.2% received OPV-3. In other words, about 11% of age-eligible children who started the series had not completed it (on a weighted basis). For DPT, weighted dropout rates hover around 5%.¹²

Timeliness also declines across doses. For each dose in a series, the green bars—doses given within recommended 28-day window—are longest for the first dose and shorter for later doses. Timely vaccination is highest for early doses such as BCG, OPV0, and HepB, but drops for later doses (with the exception of OPV-1) and for vaccines given later in infancy.

Many children received later doses late—under two months (light pink bars) or more than two months overdue (dark pink bars)—especially for DPT-3, PCV-3, and RV-3. The dark pink bars, representing doses given 2+ months late, are smallest for first doses and grow noticeably for later doses.¹³

Figure 4.22: Vaccination coverage and timeliness, sentinel and non-sentinel LGAs¹⁴

Table 4.70: Cumulative percentage of bar segment lengths for Figure 4.22, sentinel and non-sentinel LGAs

Dose	Too early	Early or timely (<28 days)	No more than two months late	Total coverage (known timing only)	Total coverage (with timing unknown)
BCG	0.0%	36.5%	43.5%	52.9%	65.5%
HEPB	0.0%	30.2%	32.6%	34.3%	54.2%
OPV0	0.0%	34.1%	38.4%	42.3%	66.0%
OPV1	7.5%	37.6%	42.6%	49.5%	64.0%
DPT1	7.1%	38.2%	43.5%	50.6%	62.7%
PCV1	7.1%	38.3%	43.4%	50.6%	62.6%
RV1	7.2%	38.1%	43.0%	49.5%	62.3%
IPV1	6.0%	32.3%	36.8%	45.7%	61.8%
OPV2	5.2%	33.0%	40.1%	49.7%	61.2%
DPT2	5.1%	32.7%	40.2%	49.9%	61.0%
PCV2	5.1%	32.9%	40.0%	49.8%	61.0%
RV2	4.9%	32.4%	39.6%	48.3%	60.5%
OPV3	3.7%	27.1%	35.9%	48.5%	57.2%
DPT3	3.6%	27.4%	36.7%	49.1%	59.3%
PCV3	3.7%	27.0%	36.4%	49.0%	58.6%
RV3	3.2%	25.6%	34.1%	44.2%	57.2%
IPV2	2.7%	20.8%	27.9%	38.1%	53.0%
MCV1	5.1%	35.0%	41.6%	47.6%	58.1%
MENI	3.7%	25.8%	31.3%	36.2%	45.6%
YF	4.3%	30.3%	36.0%	41.2%	46.3%
MCV2	5.8%	32.5%	38.1%	39.5%	51.5%

4.2.5.2 Cumulative Coverage Curves

RI_CCC_02

The cumulative coverage curves show the survey-weighted proportion of children whose vaccination card records DPT-1, DPT-2, or DPT-3 at each day of life.

Coverage increases almost vertically on the scheduled ages of 42, 70, and 98 days, when the child becomes eligible for each particular dose. After each of these inflection points, the curves continue to rise at a lower gradient; theses segments represents doses given after the minimum age. Using a 28-day window to define timeliness, a DPT dose is classified as late if it has not been received by the start of the next scheduled dose.

Timeliness is low. For instance, at 70 days—the last available day for a timely DPT-1 vaccination and minimum age for receiving DPT-2—less than 40% of children have a documented first dose. Similarly, ~35% of children have DPT-3 recorded by 98 days. Overall, as doses increase—from DPT-1 to DPT-2 and from DPT-2 to DPT-3—the proportion of children that are late also increases.

The curves eventually reach a plateau, which indicates the final documented coverage achieved in the sample. Even after allowing for late doses, these plateaus remain substantially below 100%, confirming incomplete overall coverage.

There is minimal evidence of early DPT vaccination, as the curves remain close to 0% before the scheduled ages.

Trends are similar across survey areas.

Sentinel & Non-Sentinel LGAs

Figure 4.23: Cumulative curves of age at vaccination (days), sentinel and non-sentinel LGAs

Gabasawa

Figure 4.24: Cumulative curves of age at vaccination (days), Gabasawa

Gaya

Figure 4.25: Cumulative curves of age at vaccination (days), Gaya

Nassarawa

Figure 4.26: Cumulative curves of age at vaccination (days), Nassarawa

Non-Sentinel LGAs

Figure 4.27: Cumulative curves of age at vaccination (days), non-sentinel LGAs

4.2.5.3 Cumulative Interval Curves

RI_CIC_02

Similar insights can be derived from cumulative interval curves for DPT-1 and DPT-2. These curves show the survey-weighted proportion of all age-eligible children whose vaccination cards show an intra-dose interval no longer than 28 days.

Nearly the entire curve lies flat until the scheduled 28-day interval, confirming that almost no children receive the second dose prematurely.

At the 28-day mark, the line jumps abruptly to about one quarter of the cohort, reflecting on-schedule administration.

Beyond this point, the curve rises only gradually, adding roughly eighteen percentage points over the next month and then reaching a plateau near 42%.

Sentinel & Non-Sentinel LGAs

Figure 4.28: Cumulative interval curve between DPT-1 and DPT-2, sentinel and non-sentinel LGAs

Gabasawa

Figure 4.29: Cumulative interval curve between DPT-1 and DPT-2, Gabasawa

Gaya

Figure 4.30: Cumulative interval curve between DPT-1 and DPT-2, Gaya

Nassarawa

Figure 4.31: Cumulative interval curve between DPT-1 and DPT-2, Nassarawa

Non-Sentinel LGAs

Figure 4.32: Cumulative weighted proportion of vaccinated children, non-sentinel LGAs

4.2.6 Missed Opportunities for Simultaneous Vaccination

RI_QUAL_07 RI_QUAL_08 RI_QUAL_09

A Missed Opportunity for Vaccination (MOV) happens when a child visits a health facility but does not receive all the vaccine doses for which they are eligible at the time of the visit. A Missed Opportunity for Simultaneous Vaccination (MOSV) is a specific type of MOV, which occurs when a child receives at least one vaccine during a visit but misses one or more other eligible doses. Health facility visit dates recorded in a HBR can be used to identify these MOSVs and track how often they occur.

Figure 4.33 and Figure 4.37 show clear differences in MOSVs for children aged 12-23 months between sentinel and non-sentinel LGAs, both in how frequently they occur and how frequently they are later corrected.

MOSVs for any vaccine dose are much more common in sentinel LGAs (Figure 4.33). Nassarawa displays the highest rate at 79.8%; Gabasawa and Gaya follow closely with 74.1% and 73.6% respectively, while non-sentinel LGAs show a lower rate at 60.7%.

Figure 4.33: Proportion of children aged 12-23 months with MOSVs for any dose, by survey area

Figure 4.34: Proportion of children aged 12-23 months with MOSVs for DPT-1, by survey area

When a child first visits a health facility after becoming eligible for a vaccine, there are two possible outcomes: either the child receives the dose at the first eligible opportunity, or they experience a MOSV (Kazi et al. 2024). If the dose is later given at another visit, the MOSV is considered corrected; if the dose still has not been administered by the time of the survey, the MOSV is considered uncorrected (Kazi et al. 2024).

When focusing on the proportion of children aged 12-23 months with MOSVs who eventually had all their missed doses corrected, the pattern observed above reverses (Figure 4.37). The percentage of respondents with MOSVs that had all MOSVs later corrected is much higher in non-sentinel (72.6%) than sentinel LGAs (Gabasawa at 38.3%, Gaya at 36.3%, and Nassarawa at just 21.5%).

Figure 4.35: Proportion of visits with MOSVs for any dose, by survey area

Figure 4.36: Proportion of visits with MOSVs for DPT-1, by survey area

Overall, caregivers in sentinel LGAs (particularly Nassarawa) are not only more likely to miss opportunities for simultaneous vaccination, but also less likely to correct them compared to caregivers in non-sentinel LGAs. One possible reason is Nassarawa’s higher early vaccination coverage, which creates more opportunities for children to miss doses as they move through the schedule.

Figure 4.34 shows MOSVs DPT-1. Reassuringly, the proportion of DPT-1 MOSVs that are later corrected is very high across all LGAs, with Gaya leading at 99.6% of missed DPT-1 doses eventually corrected (Figure 4.38).

Figure 4.37: Percent of respondents with MOSVs that had all MOSVs later corrected, by survey area

Figure 4.38: Proportion of MOSVs for DPT-1 that were later corrected, by survey area

Figure 4.39 shows the estimated valid DPT-1 coverage if there had been no MOSVs and no early doses. The blue bars represent this hypothetical valid coverage, while the gray bars show the actual valid coverage.

Across sentinel and non-sentinel LGAs, the differences between actual and hypothetical coverage are small. Nassarawa shows the highest potential gain, with coverage rising from 53.8% to 56.5% (a gain of 2.7 percentage points). Gabasawa and Gaya show similar small gaps, with potential increases of around 1.9 to 2.1 percentage points. The overall combined estimate for sentinel and non-sentinel LGAs moves only slightly, from 51.4% to 53.2% (a gain of 1.8 percentage points).

These minor differences suggest that MOSVs and early doses have had little impact on DPT-1 coverage in Kano State. In other words, even if the routine immunization system had captured every opportunity to vaccinate at the right time, the increase in valid DPT-1 coverage would have been minimal. Other factors are likely behind the low DPT-1 vaccination rates.

Figure 4.39: Proportion of children who would have valid DPT-1 if no MOSVs and no early doses

Table 4.71: Missed opportunities for simultaneous vaccination, by survey area (Summary)

Survey area	No MOSVs	All MOSVs corrected	Mixture of corrected and uncorrected MOSVs	All MOSVs uncorrected
Gabasawa	25.9%	28.4%	25.1%	20.6%
Gaya	26.4%	26.7%	26.0%	20.8%
Nassarawa	20.2%	17.2%	33.4%	29.2%
Non-sentinel LGAs	39.3%	44.1%	11.0%	5.6%
Note: Calculations use crude measures of missed opportunities for simultaneous vaccination (MOSVs). Rows show percentages, calculated using the following doses: BCG, DPT1, DPT2, DPT3, HEPB, IPV1, IPV2, MCV1, MCV2, MENI, OPV0, OPV1, OPV2, OPV3, PCV1, PCV2, PCV3, RV1, RV2, RV3, YF. The table was generated using VCQI's MOSV application, available at https://biostat-global-consulting.shinyapps.io/MOV_Tool_Public/, using routine immunization dataset output by the {vcqiR} library.

Table 4.72: Missed opportunities for simultaneous vaccination, by antigen, dose and correction (Detail)

dose	Gabasawa			Gaya			Nassarawa			Non-sentinel LGAs
	Vaccinated at first eligible opportunity	Later corrected MOSV	Uncorrected MOSV	Vaccinated at first eligible opportunity	Later corrected MOSV	Uncorrected MOSV	Vaccinated at first eligible opportunity	Later corrected MOSV	Uncorrected MOSV	Vaccinated at first eligible opportunity	Later corrected MOSV	Uncorrected MOSV
BCG	88.8%	10.9%	0.2%	89.9%	10.1%	0.0%	88.9%	10.9%	0.1%	90.1%	9.8%	0.1%
DPT1	82.2%	17.4%	0.5%	78.9%	21.0%	0.1%	84.9%	14.2%	0.9%	83.2%	16.1%	0.7%
DPT2	93.1%	6.7%	0.2%	94.1%	5.9%	0.0%	91.9%	7.8%	0.4%	94.2%	5.5%	0.3%
DPT3	92.2%	7.7%	0.1%	93.4%	6.5%	0.1%	92.6%	7.4%	0.0%	93.7%	6.1%	0.2%
HEPB	81.1%	5.8%	13.1%	82.7%	5.1%	12.2%	88.9%	5.2%	5.9%	89.5%	3.1%	7.3%
IPV1	63.5%	32.4%	4.1%	59.5%	30.9%	9.7%	68.5%	29.9%	1.5%	76.8%	21.9%	1.3%
IPV2	37.2%	23.1%	39.7%	38.2%	21.7%	40.1%	23.2%	10.2%	66.6%	58.9%	35.9%	5.2%
MCV1	91.7%	6.8%	1.5%	91.0%	7.7%	1.3%	91.0%	7.1%	2.0%	89.0%	9.2%	1.8%
MCV2	97.5%	2.5%	0.0%	98.3%	1.4%	0.3%	98.6%	1.0%	0.3%	97.4%	2.1%	0.5%
MENI	50.8%	6.9%	42.4%	50.4%	6.1%	43.5%	23.3%	3.7%	73.0%	83.6%	10.1%	6.3%
OPV0	86.4%	5.4%	8.3%	88.9%	4.7%	6.5%	90.3%	4.4%	5.3%	92.2%	3.6%	4.2%
OPV1	82.4%	17.1%	0.4%	79.3%	20.6%	0.1%	87.5%	12.1%	0.4%	83.4%	16.0%	0.6%
OPV2	93.5%	6.1%	0.4%	94.8%	5.2%	0.0%	92.3%	7.5%	0.2%	94.0%	5.8%	0.2%
OPV3	90.8%	7.8%	1.3%	91.4%	8.4%	0.2%	90.2%	9.4%	0.4%	91.6%	7.4%	1.0%
PCV1	82.2%	17.3%	0.5%	79.0%	20.9%	0.2%	85.1%	13.3%	1.6%	82.9%	16.5%	0.7%
PCV2	93.1%	6.9%	0.0%	94.5%	5.5%	0.0%	92.1%	7.5%	0.4%	93.7%	6.0%	0.3%
PCV3	93.3%	6.7%	0.0%	93.3%	6.6%	0.1%	93.3%	6.7%	0.0%	93.5%	6.3%	0.2%
RV1	82.4%	16.7%	1.0%	80.0%	18.9%	1.1%	85.4%	12.5%	2.1%	84.2%	15.1%	0.8%
RV2	92.8%	7.0%	0.3%	92.8%	6.1%	1.1%	93.5%	6.1%	0.4%	93.0%	6.3%	0.6%
RV3	88.7%	6.5%	4.8%	89.1%	6.7%	4.2%	90.1%	6.8%	3.1%	90.2%	6.0%	3.8%
YF	56.8%	5.5%	37.7%	58.2%	4.8%	36.9%	28.0%	3.7%	68.3%	85.6%	9.3%	5.1%
Note: Calculations use crude measures of missed opportunities for simultaneous vaccination (MOSVs). Dose rows show percentages. The table was generated using VCQI's MOSV application, available at https://biostat-global-consulting.shinyapps.io/MOV_Tool_Public/, using routine immunization dataset output by the {vcqiR} library.

4.3 Behavioral and Social Drivers of Vaccination

4.3.1 Factors Associated with No/Incomplete Vaccination

4.3.1.1 Regression Analysis of Risk Factors

Mindset estimated a design-based logistic regression for zero-dose status (Penta-1 based) among children aged 6–23 months in the three sentinel LGAs (Gaya, Gabasawa, and Nassarawa). The analysis included 11364 children, after excluding variables with substantial missing data or insufficient variation across responses.

The age range was extended to 6–23 months (rather than the conventional 12–23) to maximise available training data; zero-dose classification is stable across this range and the underlying drivers of non-vaccination are not expected to differ meaningfully between sub-groups.¹⁵ The analysis was restricted to sentinel LGAs because survey weights differ substantially between sentinel and non-sentinel areas, which would give non-sentinel observations disproportionate leverage over the model fit.¹⁶

The analysis used a forward stepwise variable selection approach based on the design-based AIC metric, applied to survey-weighted logistic regression, with candidate predictors drawn from the full household and caregiver questionnaire.¹⁷¹⁸

The forward-selection routine retained 7 predictors from the candidate set. In human-interpretable terms, these correspond to:

• vax_importance: The caregiver’s perceived importance of vaccines for the child’s health
• birth_reg: Whether the child’s birth was officially registered
• vax_ease_to_pay: Whether the caregiver finds it easy to afford the costs associated with getting their child vaccinated
• vax_effective: Whether the caregiver believes vaccines are effective
• nn1_zd: Zero-dose status of the nearest neighbouring child in the dataset (excluding siblings from the same household)
• trust_vax_info: Caregiver’s trust in available vaccine information
• community_importance: Perceived importance of vaccines in the caregiver’s community

Two referencing conventions were used for categorical predictors. For scales with an explicit neutral midpoint (e.g., trust in vaccine information), the neutral category was set as the reference so that coefficients capture departures in both positive and negative directions. For monotonic scales without a true neutral point (e.g., perceived importance of vaccination, ease of affording vaccination), the lowest category was set as the reference so that coefficients capture the progressive protective effect of more favourable responses. Binary and demographic predictors use the most common or substantively interpretable category as baseline.

Table 4.73 summarises how each selected predictor was operationalised in the survey data.

Table 4.73: Operationalisation of selected predictors in the modeling sample (sentinel LGAs, children aged 6–23 months)

#	Name	Variable	Level	n	Mean	SD
1	vax_importance	The caregiver's perceived importance of vaccines for the child's health	Not at all important (ref.)	598	5.3%	0.223
			A little important	780	6.9%	0.253
			Moderately important	1,748	15.4%	0.361
			Very important	7,266	63.9%	0.480
			Don't know / Refuse	972	8.6%	0.280
2	birth_reg	Whether the child's birth was officially registered	Registered (ref.)	4,300	37.8%	0.485
			Not Registered	7,064	62.2%	0.485
3	vax_ease_to_pay	Whether the caregiver finds it easy to afford the costs associated with getting their child vaccinated	Not at all easy (ref.)	688	6.1%	0.238
			A little easy	1,244	10.9%	0.312
			Moderately easy	1,570	13.8%	0.345
			Very easy	6,446	56.7%	0.495
			Don't know / Refuse	1,416	12.5%	0.330
4	vax_effective	Whether the caregiver believes vaccines are effective	Yes (ref.)	9,698	85.3%	0.354
			No	519	4.6%	0.209
			Don't know / Unsure	1,147	10.1%	0.301
5	nn1_zd	Zero-dose status of the nearest neighbouring child (excluding siblings from the same household)		8,271	72.8%	0.445
			Neighbour's household not zero-dose	3,093	27.2%	0.445
6	trust_vax_info	Caregiver's trust in available vaccine information	No trust at all	311	2.7%	0.163
			Little trust	292	2.6%	0.158
			Neutral (ref.)	866	7.6%	0.265
			Moderate trust	1,838	16.2%	0.368
			Complete trust	7,553	66.5%	0.472
			Don't know / Refuse	504	4.4%	0.206
7	community_importance	Perceived importance of vaccines in the caregiver's community	Yes (ref.)	9,020	79.4%	0.405
			No	1,473	13.0%	0.336
			Don't know / Refuse	871	7.7%	0.266
Note: "(ref.)" denotes the reference category for each predictor.

The tables below show the model selection path and cross-validation performance.

Model Selection Path

Table 4.74: Forward stepwise variable selection path using svyglm AIC

Step	Variable added/removed	AIC	Pseudo-R-squared
0		14,035.3	0.000
1	+ vax_importance	11,780.0	0.255
2	+ birth_reg	10,868.7	0.344
3	+ vax_ease_to_pay	10,364.4	0.392
4	+ vax_effective	10,118.9	0.414
5	+ nn1_zd	9,977.1	0.427
6	+ trust_vax_info	9,867.9	0.438
7	+ community_importance	9,789.1	0.445

Cross-Validation Performance

Table 4.75: Five-fold cross-validation performance of the final logistic regression model

Holdout Fold	AUCa	Brierb	Accuracyc	Kappad	Squared correlatione	CCCf
Fold 1	0.84	0.138	0.81	0.505	0.368	0.538
Fold 2	0.86	0.135	0.81	0.561	0.406	0.585
Fold 3	0.81	0.136	0.82	0.483	0.323	0.493
Fold 4	0.82	0.149	0.80	0.499	0.340	0.511
Fold 5	0.86	0.132	0.82	0.569	0.415	0.581
All areas	0.84	0.138	0.81	0.525	0.372	0.544
aArea Under the ROC Curve: summarises the model's ability to rank zero-dose children above vaccinated children across all probability thresholds. Values above 0.8 are considered good discriminative performance.
bBrier Score: mean squared error between the model's predicted probability and the actual zero-dose indicator (0 = perfect; 0.25 = uninformative baseline for a balanced binary outcome). Lower is better.
cAccuracy: proportion of children correctly classified as zero-dose or vaccinated at a 0.5 probability threshold.
dCohen's Kappa: classification accuracy adjusted for chance agreement (0 = no better than chance; 1 = perfect agreement).
eSquared correlation: squared Pearson correlation between the model's predicted probabilities and the observed zero-dose indicators (0 to 1; higher means predictions track observations more tightly).
fConcordance Correlation Coefficient: measures agreement between predicted probabilities and actual zero-dose status, combining both precision and accuracy of predictions.

Model fit was assessed through five-fold cross-validation.¹⁹ The model achieved an overall Area Under the Receiver Operating Characteristic Curve (AUC-ROC) of 0.84, which is considered reasonably good discriminative performance, and a cross-validated squared correlation of 0.372 between predicted probabilities and observed zero-dose indicators.²⁰

Figure 4.40: Odds ratios for risk factors identified in the design-based logistic regression for zero-dose status

Figure 4.40 plots the odds ratios for the final selected risk factors. Children whose birth was not registered have substantially elevated odds of being zero-dose — corresponding to an average predicted probability of 38% compared with 20% for children with registered births — as do children of caregivers who do not consider vaccines important. Other predictors — including place of delivery, household size, internet use, and nearest-neighbor zero-dose status (the binary vaccination status of the geographically closest child in the dataset, excluding any siblings from the same household; its selection is unsurprising and confirms that zero-dose status clusters geographically) — show more modest but directionally consistent associations. Variables whose confidence intervals cross one should be interpreted cautiously, as their estimated effects are statistically consistent with no association.

Two risk factors stand out as by far the dominant predictors: birth registration and caregiver belief in vaccine importance. Their prominence is substantively interesting, and each deserves closer examination.

Birth registration is, on its face, an administrative act — yet it emerges as one of the strongest predictors of zero-dose status in this analysis. Part of the association may reflect a convenience pathway: families who deliver in a health facility are in a setting where birth registration and vaccination can often happen together — or at least where both services are physically accessible at the same time and place. The co-location of services reduces the logistical barriers to both, creating a logical dependence between access to registration and access to early vaccination doses. Beyond this, the association likely also reflects a deeper latent construct: families who navigate civil registration processes tend to be the same families who engage more broadly with government services and the formal healthcare system. Registering a birth requires interacting with government offices, implies a degree of trust in — or at least willingness to engage with — state institutions, and reflects a pattern of health-seeking and administrative participation that predisposes households to seek routine immunization. In this sense, birth registration acts as a proxy for social and institutional integration, a characteristic that shapes access to and uptake of a wide range of public services. It would be interesting for future research to examine whether programmes that integrate birth registration and vaccination services — or that use registration records to identify and follow up unvaccinated children — could yield meaningful gains in coverage.

Caregiver belief in vaccine importance similarly points in a demand-side direction. Critically, earlier findings in this report show that many zero-dose children do not have caregivers who actively oppose vaccination; rather, caregivers frequently report uncertain or ambivalent beliefs about vaccine safety and benefits. In this context, the strong association with perceived importance suggests that the barrier for many families is not ideological rejection but motivational indifference — vaccines are simply not seen as sufficiently important or urgent to act on. The gradient in predicted risk is substantial: caregivers who consider vaccines not at all important have an estimated 45% probability of having a zero-dose child, compared with 28% among those who consider vaccines very important. This distinction matters for programme design. Interventions aimed primarily at countering anti-vaccine sentiment may be less effective in this context than approaches that raise the salience and perceived necessity of vaccination — for example, community-level social norms messaging, household visits by trusted community health workers, or the integration of vaccination counselling into birth registration touchpoints where caregiver attention is already high.

4.3.2 Perceptions of Vaccination

4.3.2.1 Knowledge, Awareness, and Perceptions

Despite high rates of zero-dose prevalence, most children have caregivers who believe vaccines are important for their child’s health, regardless of location. As shown in Figure 4.42, a large majority (ranging from an estimated 65% to 72% across different survey areas) of children has caregivers who rated vaccines as “very important.”

This view is consistent across each sentinel and non-sentinel LGA. Nassarawa records the highest positive rates, with an estimated 72% of children having caregivers who rated vaccines as “very important,” and more than one in ten rating them as at least “a little important”.

Zero-Dose Status

Figure 4.41: Perceived importance of vaccines for child health, by zero-dose status

Survey Area

Figure 4.42: Perceived importance of vaccines for child health, by survey area

A child’s vaccination status is strongly associated to caregivers’ views on vaccine importance \((F_{(3, 496)}=105.8\), \(p<.001)\), safety \((F_{(2, 497)}=216.1\), \(p<.001)\) and benefits \((F_{(1, 498)}=147.1\), \(p<.001)\), but show no significant variation across LGAs.

Notably, an estimated 96% of children vaccinated with Penta-1 has caregivers who see vaccines as “moderately” or “very important,” compared to 57% of zero-dose children (Figure 4.41 shows).

Interestingly, roughly one-third of zero-dose children have caregivers who, rather than being convinced that vaccines are unsafe (Figure 4.43) or have no benefits (Figure 4.45), report uncertain opinions regarding vaccine safety and benefits.

Zero-Dose Status

Figure 4.43: Caregivers’ belief in vaccine safety, by zero-dose status

Survey Area

Figure 4.44: Caregivers’ belief in vaccine safety, by survey area

Zero-Dose Status

Figure 4.45: Caregivers’ belief in vaccination benefits, by zero-dose status

Survey Area

Figure 4.46: Caregivers’ belief in vaccination benefits, by survey area

Caregivers’ beliefs and decisions about vaccines are shaped not only by their own knowledge and experiences, but also by their social environment. As shown in Figure 4.47, most children who received Penta-1 (92%) have caregivers who feel that people important to them approve of vaccination, compared to only 56% of zero-dose children \((F_{(1, 498)}=147.1\), \(p<.001)\).

Figure 4.47: Perceived importance of vaccination among close contacts, by zero-dose status

Across LGAs, about a third of children have caregivers who say someone in their family, community, or social network encouraged them to vaccinate their child, with the highest proportions recorded in Nassarawa and Gaya (35%) (Figure 4.48). Very few report they have been advised not to vaccinate their child. However, most children—around 60%—have caregivers who do not perceive any influence from others on their decision.

Figure 4.48: Perceived influence of others on vaccination decision, by survey area

More than 60% of children have caregivers who report no external source of influence on their vaccination decision. However, when caregivers do report such influence, it most often comes from friends and family (Figure 4.49). Health workers and religious or community leaders also appear to play a role, but to a lesser extent.

Figure 4.49: Sources of influence on vaccination decision, by zero-dose status

Figure 4.50: Presence of cultural and religious beliefs around vaccination, by survey area

As shown in Figure 4.51, about one in five caregivers report that cultural and religious beliefs in their community influence views on vaccination. This pattern is consistent across LGAs, with Gaya and Gabasawa displaying a slightly higher share of positive answers (22%).

Figure 4.51: Nature of cultural and religious beliefs around vaccination, by survey area

4.3.2.2 Trust in Vaccination

While trust in the healthcare system is generally high, Figure 4.52 reveals important gaps between caregivers of zero-dose vis à vis immunized children. In fact, more than one out of ten penta-unvaccinated children has parents who report little or no trust in the system, compared to less than 1% of vaccinated children.

Similar patterns emerge in relation to trust in healthcare workers and vaccine information. Across LGAs, 14% of zero-dose children have caregivers with little or no trust in the information they receive about vaccines, compared to about 4% of immunized children (Figure 4.55). Similarly, 12% of zero-dose children have caregivers with little or no trust in healthcare workers’ ability to provide accurate and reliable care, compared to less than 1% of vaccinated infants (Figure 4.57).

Zero-Dose Status

Figure 4.52: Caregivers’ degree of trust in the health system, by zero-dose status

Survey Area

Figure 4.53: Caregivers’ degree of trust in the health system, by survey area

Zero-Dose Status

Figure 4.54: Caregivers’ degree of trust in healthcare workers, by zero-dose status

Survey Area

Figure 4.55: Caregivers’ degree of trust in healthcare workers, by survey area

Education Level

Figure 4.56: Caregivers’ trust in information about vaccines, by education level

Zero-Dose Status

Figure 4.57: Caregivers’ trust in information about vaccines, by zero-dose status

4.3.2.3 Behavioral and Social Drivers

The survey asked caregivers why their child had not received certain vaccines, or why the child had not received any vaccines at all. Respondents could select multiple reasons based on categories defined in the WHO’s Behavioral and Social Drivers (BeSD) framework (WHO and UNICEF 2022), which groups drivers under the four domains of access, health facility, beliefs, and social processes.

Figure 4.58 shows the responses given by caregivers of truly zero-dose children, and Table 4.78 presents the same information along with more detailed statistics.

Table 4.76: Percentage of truly-zero-dose children whose caregivers report at least one barrier to vaccination, by BeSD domain

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Social processes	41.3%	±3.5	(37.8; 44.8)	38.3%	44.3%	8.06	140,551	±18,948	2,337
Beliefs	23.1%	±3.5	(19.8; 26.8)	20.3%	26.2%	10.75	78,721	±12,812	1,497
Access	14.9%	±1.9	(13.1; 17.0)	13.4%	16.6%	4.67	50,787	±7,041	1,007
Health center	4.6%	±1.1	(3.7; 5.8)	3.8%	5.6%	4.08	15,779	±3,483	304
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 6,041; Note: Because answers follow a select-multiple format, proportions in this table do not necessarily sum to 100%
Note: Categories are based on existing BeSD of vaccination tools and guidance (see WHO, 2022).

Table 4.77: Percentage of truly-zero-dose children whose caregivers report at least one barrier to vaccination, by EGH domain

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Intent	59.0%	±3.7	(55.3; 62.6)	55.9%	62.0%	8.76	200,860	±19,827	3,417
Access	10.1%	±1.5	(8.7; 11.7)	8.9%	11.4%	3.86	34,381	±5,404	678
Service delivery	4.5%	±1.0	(3.6; 5.6)	3.8%	5.4%	3.54	15,394	±3,172	357
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 6,041; Note: Because answers follow a select-multiple format, proportions in this table do not necessarily sum to 100%
Note: Categories are based on the mapping of BeSD of vaccination (WHO, 2022), used in the original survey instrument, to the EGH Vaccine Delivery Framework domains (see Carter et al., 2020).

Figure 4.58: Reasons for not getting any vaccines, by BeSD domain

Figure 4.59: Reasons for not getting any vaccines, by cr{EGH} domain and subdomain

The most cited reasons relates to social norms: about 41% of children had caregivers who cited at least one social norm (Table 4.76). More specifically, 16.9% of truly-zero-dose children had caregivers who cited descriptive peer norms (i.e., family or friends do not get vaccinated—themselves or their children), and 15.6% had caregivers who pointed to injunctive peer norms (i.e., family of friends do not support vaccination), while community and religious leader norms were cited less often (2.4% and less than 1%, respectively).

Given the evidence reviewed in Section 4.3.2.1, it remains unclear why many caregivers believe family and friends disapprove of vaccination. In fact, the vast majority of children had caregivers who reported cultural and religious beliefs as supportive of vaccination (Figure 4.51), that most felt people important to them think they should vaccinate their child (Figure 4.47), and that the influence they received from family, community, or social networks encourage vaccination (Figure 4.49).

Reasons related to beliefs (23.1%), access (14.9%), or the health facility (4.6%) were mentioned less often (Table 4.76). Regarding access barriers, 6.4% of children had caregivers who said vaccines were not available (e.g., because they were busy, away from home, or the child was sick), and 4.1% had caregivers who mentioned forgetting or not knowing the national immunization schedule. Belief-related reasons—such as fear of pain (4.4%), perceived risk of disease (3.2%), or lack of confidence in vaccine benefits (2.7%)—were also cited by a minority. Finally, health facility issues such as long waiting times, inconvenient opening hours, or being turned away were rarely mentioned.

It is important to note that these figures appear small because of the multiple-choice structure of the underlying survey question. Nevertheless, they suggest that social processes—particularly what is seen as normal or acceptable in a caregiver’s network—are closely associated with child’s zero-dose status, more so than logistical barriers or individual beliefs.

Table 4.78: q428: Reasons for not getting any vaccine (Caregivers of truly zero-dose children)

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Family or friends don’t get vaccinated (themselves or their children)	16.9%	±2.1	(14.9; 19.2)	15.2%	18.8%	5.06	57,589	±8,400	932
Family or friends don’t support vaccination	15.6%	±2.0	(13.7; 17.7)	14.0%	17.4%	4.71	53,221	±8,288	835
Decided to delay vaccination	7.3%	±2.8	(5.0; 10.6)	5.3%	10.0%	17.82	24,964	±9,460	427
Needed permission to take child for vaccination	6.6%	±2.4	(4.6; 9.5)	4.9%	9.0%	14.60	22,641	±9,644	420
Not available to get vaccinated (busy, away from home, or child was sick)	6.4%	±1.2	(5.4; 7.7)	5.5%	7.5%	3.46	21,906	±4,042	408
Pain or distress from injection	4.5%	±1.0	(3.6; 5.5)	3.7%	5.3%	3.36	15,187	±3,428	288
Didn’t know or forgot when due for vaccination	4.1%	±1.1	(3.2; 5.4)	3.3%	5.1%	4.66	14,035	±3,689	247
Not concerned about the disease(s) that the vaccine prevents	3.2%	±1.2	(2.3; 4.6)	2.4%	4.4%	6.83	11,059	±4,013	215
Health worker didn’t recommend vaccination	2.9%	±1.0	(2.0; 4.0)	2.1%	3.8%	5.31	9,718	±3,535	142
Vaccine not important / needed / effective	2.7%	±0.6	(2.2; 3.4)	2.3%	3.3%	2.09	9,217	±2,074	236
Community leaders don’t support vaccination	2.4%	±0.7	(1.8; 3.3)	1.9%	3.1%	3.51	8,198	±2,536	137
Vaccine not safe / concerned about side effects	2.0%	±0.7	(1.4; 2.8)	1.5%	2.7%	3.68	6,860	±2,361	123
Don’t trust health workers who give vaccines	1.7%	±0.6	(1.2; 2.4)	1.2%	2.2%	3.30	5,699	±2,059	107
Heard rumours or something bad about vaccines	1.4%	±0.7	(0.9; 2.3)	1.0%	2.1%	4.94	4,828	±2,262	78
Health workers do not spend enough time with people	1.2%	±0.5	(0.8; 1.8)	0.9%	1.7%	2.94	4,049	±1,592	68
Prefer traditional or alternative approaches	1.2%	±0.5	(0.7; 1.8)	0.8%	1.7%	3.86	3,938	±1,794	107
Health center has inconvenient opening times	1.0%	±0.5	(0.6; 1.7)	0.6%	1.6%	4.56	3,411	±1,789	57
Could not afford other costs related to vaccination (e.g., transport to health center, taking time away from work, childcare, recovery time)	1.0%	±0.5	(0.6; 1.6)	0.6%	1.5%	4.28	3,264	±1,821	59
Health center is hard to get to	1.0%	±0.4	(0.6; 1.4)	0.7%	1.3%	2.41	3,260	±1,269	85
His Age not eligible to that vaccine	0.9%	±0.4	(0.6; 1.4)	0.6%	1.3%	2.80	3,080	±1,280	79
Not contacted about being due for vaccination	0.8%	±0.3	(0.6; 1.2)	0.6%	1.2%	2.00	2,868	±1,072	97
Didn’t know where to go for vaccination	0.8%	±0.3	(0.6; 1.2)	0.6%	1.1%	1.55	2,823	±956	83
Health center turns people away (vaccines not available, session cancelled)	0.6%	±0.3	(0.4; 1.0)	0.4%	0.9%	2.66	2,056	±1,052	41
Health workers were not respectful	0.6%	±0.3	(0.3; 1.0)	0.4%	0.9%	3.03	1,989	±1,153	31
Unsafe to travel to the vaccination site (may be due to weather, crime, conflict, or disaster)	0.6%	±0.4	(0.3; 1.1)	0.3%	1.0%	3.48	1,951	±1,196	31
Health workers were missing skills	0.4%	±0.3	(0.2; 0.9)	0.2%	0.8%	3.78	1,454	±1,072	12
Could not afford the cost at the health center (e.g., payment for services)	0.4%	±0.3	(0.2; 0.8)	0.2%	0.7%	2.81	1,405	±948	31
Health center has long waiting times	0.2%	±0.2	(0.1; 0.6)	0.1%	0.5%	3.92	769	±800	24
Religious leaders don’t support vaccination	0.1%	±0.1	(0.0; 0.2)	0.0%	0.1%	0.64	263	±186	18
Health center is not clean	0.0%	±0.0	(0.0; 0.1)	0.0%	0.1%	0.24	101	±72	8
Don't know	11.5%	±2.6	(9.2; 14.4)	9.5%	13.9%	10.61	39,264	±11,596	696
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 6,041; Note: Because answers follow a select-multiple format, proportions in this table do not necessarily sum to 100%
Note: Categories are based on existing behavioural and social drivers (BeSD) of vaccination tools and guidance. See WHO (2022).

Interestingly, 11.5% of truly-zero-dose children had caregivers who either did not know (11.5%) or refused to answer the question (4.3%) when asked why their child had received no vaccines (Table 4.78). For many caregivers, the reasons behind non-vaccination may be unclear even to themselves, difficult to articulate, or sensitive to disclose.

This type of answers could reflect genuine uncertainty, especially if decisions around vaccination are shaped by multiple overlapping factors that caregivers find hard to disentangle. They could also point to discomfort in openly sharing reasons that could be seen as socially undesirable (for example, if the respondent assumes the researchers support vaccination).

Although the questionnaire options were based on the WHO BeSD framework, each BeSD domain and subdomain was later mapped onto the Exemplars in Global Health (EGH) Exemplars in Vaccine Delivery Framework (Carter et al. 2020). The EGH Framework, informed by the work of LaFond et al. (2015) and Phillips et al. (2017), was developed to understand how exemplar countries achieve high vaccine coverage (Bednarczyk et al. 2022). It groups barriers into three domains—intent, access, and service delivery—each with related subdomains.

The mapping results show that most common barriers fall within the “intent” domain, especially in the subdomains of (social)norms and attitudes and perceptions (Figure 4.59). Nearly six in ten (59%) caregivers of truly zero-dose children reported at least one reason in the “intent” domain (Table 4.77).

4.3.3 Access to Healthcare and Service Environment

4.3.3.1 Physical Access

As shown in Figure 4.60, for most children—across all survey areas—the nearest health facility is within 1-5 km of their household. Nassarawa stands out, with more than half of children having a facility within one km, likely reflecting its urban setting. In contrast, in Gaya and Gabasawa, only around 40% of children have a health facility within that distance, and in non-sentinel LGAs, the figure is even lower, at 29%. This figure considers any healthcare facility, regardless of whether immunization services are offered.

Survey Area

Figure 4.60: Distance to nearest health facility, by survey area

Zero-Dose Status

Figure 4.61: Distance to nearest health facility, by zero-dose status

Table 4.79: Distance to nearest health facility, by survey area and zero-dose status of the household

Survey area	Zero-dose status (household)	10 minutes (<1 Km)		10-60 minutes (1-5 Km)		1-2 hours (6-10 Km)		More than 3 hours (>10 Km)
		%	Est. total	%	Est. total	%	Est. total	%	Est. total
Gabasawa	Non-zero-dose household	38.9%	8,061	46.5%	9,642	12.9%	2,674	1.7%	354
	Zero-dose household	37.5%	2,896	48.0%	3,709	11.8%	910	2.7%	207
	All households with children 12-23	38.5%	10,957	46.9%	13,351	12.6%	3,583	2.0%	560
Gaya	Non-zero-dose household	44.7%	8,775	45.4%	8,907	9.2%	1,811	0.6%	127
	Zero-dose household	31.8%	3,061	51.9%	4,995	14.1%	1,354	2.2%	210
	All households with children 12-23	40.5%	11,835	47.5%	13,901	10.8%	3,165	1.2%	337
Nassarawa	Non-zero-dose household	57.6%	13,703	37.1%	8,830	3.5%	844	1.8%	428
	Zero-dose household	47.6%	1,934	47.3%	1,924	3.3%	134	1.8%	74
	All households with children 12-23	56.1%	15,637	38.6%	10,755	3.5%	979	1.8%	502
Non-sentinel LGAs	Non-zero-dose household	29.7%	96,987	60.0%	196,279	10.0%	32,751	0.3%	1,046
	Zero-dose household	23.2%	27,090	64.3%	75,026	12.5%	14,558	0.1%	96
	All households with children 12-23	28.0%	124,077	61.1%	271,305	10.7%	47,309	0.3%	1,142
All LGAs	Non-zero-dose household	32.6%	127,526	57.2%	223,658	9.7%	38,080	0.5%	1,954
	Zero-dose household	25.3%	34,980	62.0%	85,653	12.3%	16,956	0.4%	587
	All households with children 12-23	30.7%	162,506	58.4%	309,311	10.4%	55,036	0.5%	2,541
Note: Sample restricted to caregivers of children aged 12-23 months. The term "zero-dose household" refers to a household where at least one age-eligible child has not received a Penta-1 vaccine. These results should be interpreted with caution, as the original question combined time and distance units, skipped intermediate intervals (e.g., 2-3 hours), and the response wording was changed partway through fieldwork.

This information can also be visualized on a map. There are several ways this could be done, but the plots below model the probability that caregivers self-report being within 10 minutes (or 1km) of a healthcare facility.

Gaya

Figure 4.62: Probability that households in Gaya LGA report being within 10 minutes of a healthcare facility. The visualization uses an Indicator Kriging (IK) method to interpolate and smooth coverage estimates over the entire area of the LGA.

Gabasawa

Figure 4.63: Probability that households in Gabasawa LGA report being within 10 minutes of a healthcare facility. The visualization uses an Indicator Kriging (IK) method to interpolate and smooth coverage estimates over the entire area of the LGA.

Nassarawa

Figure 4.64: Probability that households in Nassarawa LGA report being within 10 minutes of a healthcare facility. The visualization uses an Indicator Kriging (IK) method to interpolate and smooth coverage estimates over the entire area of the LGA.

The majority rely on walking to reach the nearest healthcare facility, with an estimated 57.3% reporting this as their primary mode of transport (Figure 4.65). This is followed by 23.4% who use motorcycles, and smaller shares who rely on public transportation (9%) or three-wheelers (7.2%). Use of bicycles (2%) and private cars (less than 1%) is rare.

These figures suggest that traveling to a health facility for vaccination may involve considerable time or physical effort. In fact, 42% of caregivers that need to travel between 6-10 km and 33% of caregivers that need to travel more than 10 km still resort to walking to reach the health facility. Furthermore, the widespread reliance on walking and motorcycles suggests potential vulnerability to weather, road conditions, or safety concerns, especially for individuals in rural areas and women.

Figure 4.65: Primary modes of transportation to the healthcare facility

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Walking	57.2%	±3.2	(54.0; 60.4)	54.6%	59.9%	24.03	723,468	±57,751	13,499
Motorcycle	23.5%	±2.4	(21.1; 26.0)	21.5%	25.6%	19.06	296,552	±32,918	5,391
Public transportation (e.g., bus, taxi)	9.0%	±1.8	(7.4; 11.0)	7.6%	10.6%	22.86	113,801	±22,106	1,613
Three-wheeler	7.2%	±1.3	(6.0; 8.7)	6.2%	8.5%	15.64	91,531	±16,798	1,372
Bicycle	2.0%	±1.4	(1.0; 4.0)	1.2%	3.6%	55.23	25,791	±18,399	458
Private car	0.8%	±0.6	(0.4; 1.7)	0.4%	1.5%	28.80	10,031	±7,842	101
Other	0.2%	±0.1	(0.1; 0.4)	0.1%	0.3%	3.53	2,791	±1,432	52
Don't know							16,039	±3,673	292
Refuse to answer							8,020	±4,980	91
All	100.0%			100.0%	100.0%		1,288,024	±64,766	22,869
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 22,869

4.3.3.2 Service Availability

As shown in Figure 4.66, zero-dose children have caregivers who are not only more likely to say that vaccination services are not offered at the nearest health facility (14%), but often who do not simply know whether such services exist (15%). In contrast, among children who were vaccinated with Penta-1, only 1% has caregivers who are unaware of local vaccination services. The remaining 10% with caregivers who report no nearby services could reflect those who made the effort to visit more distant facilities because they value immunization.

Vaccination services are reported as slightly more available in Nassarawa (88%) than in the other survey areas (between 81–83%), as seen in Figure 4.67.

Zero-Dose Status

Figure 4.66: Availability of vaccines in the nearest healthcare facility, by zero-dose status

Survey Area

Figure 4.67: Availability of vaccines in the nearest healthcare facility, by survey area

Vaccination service availability, as reported by caregivers, is heavily concentrated on weekdays, especially Monday through Thursday, with limited or no provision on Fridays and weekends (Table 4.80). Additionally, almost one in ten children has caregivers who do know on which days vaccination is available.

Table 4.80: Weekdays in which vaccination is available at the nearest health facility according to caregivers

	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Monday	33.2%	±4.1	(29.2; 37.4)	29.8%	36.7%	31.75	302,756	±43,109	5,521
Tuesday	34.6%	±4.3	(30.5; 39.0)	31.1%	38.3%	33.99	315,711	±46,004	5,828
Wednesday	29.5%	±3.4	(26.2; 33.1)	26.7%	32.5%	23.64	269,472	±35,045	5,657
Thursday	34.5%	±4.2	(30.4; 38.9)	31.0%	38.1%	33.20	314,843	±45,367	5,033
Friday	11.2%	±1.7	(9.5; 13.0)	9.8%	12.7%	12.81	101,904	±15,132	1,893
Saturday	1.3%	±0.4	(0.9; 1.8)	1.0%	1.7%	6.05	11,931	±4,100	212
Sunday	0.8%	±0.3	(0.5; 1.2)	0.6%	1.1%	5.25	7,303	±2,961	107
Don't know	8.8%	±1.1	(7.7; 10.0)	7.8%	9.8%	6.87	79,952	±11,980	1,162
Refuse to answer	0.1%	±0.1	(0.0; 0.3)	0.0%	0.3%	6.26	842	±1,075	4
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 16,097; Note: Because answers follow a select-multiple format, proportions in this table do not necessarily sum to 100%

Vaccinated children are more likely to have caregivers who report both that vaccines were unavailable at some point in the past six months (25%) and that they were always available (68%) (Figure 4.68). Intuitively, caregivers who value and engage with routine immunization services are more connected to the local health system, and therefore more likely to notice both its successes and its shortcomings.

In contrast, only 9% of zero-dose children have caregivers who report stockouts, and 54% who report no stockouts. Strikingly, 37% of these children has caregivers who do not know whether stockouts occurred at all. This suggests that a large proportion of caregivers of zero-dose children have so little interaction with the health system that they are not in a position to observe supply issues. Therefore, while stockouts are a genuine barrier for many, they may not be the main reason zero-dose children are missed: limited caregiver engagement with the health system appears to be a bigger challenge.

Zero-Dose Status

Figure 4.68: Stockout events in the nearest healthcare facility (past six months prior to the survey), by zero-dose status

Survey Area

Figure 4.69: Stockout events in the nearest healthcare facility (past six months prior to the survey), by survey area

4.3.3.3 Information Environment

The analysis reveals important gaps in caregivers’ exposure to vaccination information. As shown in Figure 4.71, just over half of children (57%) have caregivers who report having ever heard or seen information about vaccines or vaccination. The trend varies by location: children in Nassarawa are the most likely to have a caregiver with reported exposure (62%), while Gaya and Gabasawa display lower figures, at 49% and 50% respectively.

A clearer divide also emerges when examining vaccination status (Figure 4.70). Among Penta-1 vaccinated children, 64% have caregivers who report having received such information, compared to just 45% among zero-dose children.

Zero-Dose Status

Figure 4.70: Exposure to information about vaccines or vaccinations, by zero-dose status

Survey Area

Figure 4.71: Exposure to information about vaccines or vaccinations, by survey area

When caregivers do receive information, health workers are by far the most common source (Figure 4.72). However, vaccinated children are much more likely to have caregivers who receive information from health workers and community leaders. By contrast, zero-dose children are more likely to have caregivers who rely on family or friends, or the radio/TV, as their main sources of health information.

Community health campaigns as well as schools or educational institutions play smaller roles overall, though schools are more commonly mentioned by caregivers of vaccinated children.

These patterns suggest that zero-dose children are more often in environments where informal or less authoritative information channels dominate, while vaccinated children are more likely to benefit from caregivers who are better connected to formal health communication sources.

Figure 4.72: Caregivers’ source of information about vaccines, by zero-dose status

Awareness of vaccination campaigns appears limited overall, with slight variation between groups. As shown in Figure 4.74, across all LGAs, only about 21% of children have caregivers who say they are aware of vaccination campaigns in their area. Gabasawa shows the highest level at 25%, while Gaya is at the lower end with 18%.

Intuitively, vaccinated children are more likely to have caregivers who report awareness of campaigns: 24% compared to 16% among zero-dose children (Figure 4.73).

Zero-Dose Status

Figure 4.73: Awareness of local vaccination campaigns, by zero-dose status

Survey Area

Figure 4.74: Awareness of local vaccination campaigns, by survey area

Community leaders and local organizations play an important role in promoting vaccination in Kano State, but their impact and visibility varies (Figure 4.75). Overall, only about 41% of children have caregivers who report that community leaders or organizations have actively promoted vaccination campaigns. This figure is highest in Gabasawa (51%) and Gaya (46%), while Nassarawa and non-sentinel LGAs are slightly lower (at 39% and 40% respectively).

Unsurprisingly, vaccinated children are more likely to have caregivers who report active promotion by community leaders and local organizations (45%) compared to zero-dose children (33%) (Figure 4.76).

Survey Area

Figure 4.75: Community leaders’ and local organizations’ promotion of vaccination campaigns, by survey area

Zero-Dose Status

Figure 4.76: Community leaders’ and local organizations’ promotion of vaccination campaigns, by zero-dose status

The patterns observed about vaccination campaigns and the role of community leaders align with broader findings on the information environment: vaccinated children are generally in households where caregivers are more exposed to or aware of formal vaccination efforts and leadership figures, while zero-dose children are more likely to come from households that are disconnected from such information.

4.3.4 Expenses Related to Vaccination

Figure 4.77 and Figure 4.78 highlight how caregivers perceive the burden of paying for vaccinations, considering clinic fees, travel costs, and lost income from taking time away from work.

While financial barriers may exist for a few, the majority of caregivers do not see paying for vaccination as a burden. Most children, across both groups and survey areas, have caregivers who reported that paying for vaccination was “very easy.”

This perception was strongest among caregivers of children aged 12–23 months who had received the Penta-1 vaccine (68%), compared to 40% of those whose children had not. In fact, a sizable proportion of caregivers whose children did not receive Penta-1 is unable to report about the easiness to pay for the vaccination.

Similarly, when looking at survey areas, the proportion reporting it as “very easy” hovered between 55% and 59%, with only minor differences between LGAs.

Across all groups, only a small minority (5–8%) found the costs “not at all easy” to manage.

Zero-Dose Status

Figure 4.77: Perceived financial burden of vaccination, by zero-dose status

Survey Area

Figure 4.78: Financial burden of vaccination, by survey area

While most children had caregivers who described the cost of vaccination as easy to manage, Table 4.82 reveals that a significant portion is noticing higher costs of vaccination relative to the previous visit. In fact, an estimated 41% of children had caregivers who felt that the cost of their child’s most recent vaccination was at least “slightly” if not “much more” than what they had paid previously, compared to only 20.8% who perceived a decrease.

This suggests that although caregivers generally find it easy to cover vaccination costs, a sizable portion is finding it more expensive than the previous vaccinations. If sustained over time, this trend could start to erode the prevailing sense that vaccination is affordable, as shown in Figure 4.77 and Figure 4.78. Nevertheless, perceptions may vary and are subject to change; higher costs may also be viewed within the broader economic context rather than necessarily reflecting the true underlying cost of vaccination.

Table 4.81: Financial burden of vaccination, by survey area and zero-dose status

Survey area	Zero-dose status	Not at all easy		A little easy		Moderately easy		Very easy		Refuse to answer		Don't know
		%	Est. total	%	Est. total	%	Est. total	%	Est. total	%	Est. total	%	Est. total
Gabasawa	Child without Penta-1	5.5% (4.1; 7.3)	624	13.2% (10.6; 16.5)	1,515	11.1% (8.8; 13.8)	1,265	39.4% (35.8; 43.2)	4,513	4.5% (2.7; 7.5)	515	26.3% (23.3; 29.6)	3,013
	Child with Penta-1	4.9% (4.0; 6.1)	1,138	12.4% (10.6; 14.5)	2,859	17.1% (15.0; 19.3)	3,935	63.1% (60.4; 65.7)	14,554	0.2% (0.1; 0.4)	36	2.3% (1.6; 3.4)	540
	All caregivers of children 12-23	5.1% (4.3; 6.1)	1,761	12.7% (11.2; 14.4)	4,374	15.1% (13.4; 16.9)	5,200	55.3% (52.9; 57.6)	19,067	1.6% (1.0; 2.6)	551	10.3% (9.0; 11.7)	3,553
Gaya	Child without Penta-1	9.3% (7.0; 12.3)	1,267	6.6% (5.0; 8.5)	892	9.0% (7.2; 11.1)	1,221	38.5% (35.1; 41.9)	5,220	1.9% (1.2; 3.1)	264	34.7% (31.2; 38.4)	4,711
	Child with Penta-1	5.2% (4.1; 6.6)	1,160	11.3% (9.5; 13.2)	2,511	13.8% (12.0; 15.8)	3,066	68.0% (65.2; 70.6)	15,150	0.2% (0.0; 3.0)	45	1.6% (1.1; 2.3)	358
	All caregivers of children 12-23	6.8% (5.6; 8.1)	2,427	9.5% (8.2; 11.0)	3,403	12.0% (10.6; 13.5)	4,287	56.8% (54.6; 59.0)	20,371	0.9% (0.5; 1.4)	309	14.1% (12.7; 15.7)	5,069
Nassarawa	Child without Penta-1	5.4% (3.2; 9.1)	437	15.0% (11.1; 19.9)	1,207	12.8% (9.3; 17.4)	1,032	44.8% (39.0; 50.8)	3,616	2.0% (0.5; 7.5)	164	19.9% (15.8; 24.9)	1,609
	Child with Penta-1	5.7% (4.1; 7.8)	1,350	9.9% (8.2; 12.0)	2,370	17.1% (14.4; 20.1)	4,072	64.6% (61.1; 67.9)	15,401	0.3% (0.0; 95.8)	70	2.4% (1.5; 3.9)	578
	All caregivers of children 12-23	5.6% (4.3; 7.3)	1,787	11.2% (9.6; 13.1)	3,577	16.0% (13.7; 18.6)	5,104	59.6% (56.6; 62.5)	19,018	0.7% (0.3; 2.0)	235	6.9% (5.5; 8.5)	2,187
Non-sentinel LGAs	Child without Penta-1	7.7% (5.4; 10.8)	15,046	8.6% (6.7; 11.0)	16,838	9.6% (5.7; 15.7)	18,831	40.2% (35.1; 45.6)	78,823	2.0% (0.9; 4.3)	3,883	31.9% (28.1; 35.9)	62,422
	Child with Penta-1	4.5% (3.3; 6.1)	14,217	9.7% (7.2; 12.9)	30,707	15.6% (13.1; 18.5)	49,173	68.2% (63.0; 73.0)	215,244	0.2% (0.0; 1.5)	569	1.8% (1.0; 2.9)	5,560
	All caregivers of children 12-23	5.7% (4.4; 7.3)	29,263	9.3% (7.6; 11.4)	47,545	13.3% (10.4; 16.8)	68,004	57.5% (52.5; 62.4)	294,067	0.9% (0.4; 1.8)	4,452	13.3% (11.4; 15.5)	67,982
All LGAs	Child without Penta-1	7.6% (5.6; 10.2)	17,373	8.9% (7.3; 11.0)	20,452	9.8% (6.3; 14.8)	22,349	40.3% (35.8; 44.9)	92,173	2.1% (1.2; 3.8)	4,826	31.3% (28.1; 34.8)	71,755
	Child with Penta-1	4.6% (3.6; 5.9)	17,865	10.0% (7.9; 12.6)	38,446	15.7% (13.6; 18.0)	60,246	67.7% (63.5; 71.6)	260,350	0.2% (0.1; 0.4)	720	1.8% (1.2; 2.7)	7,036
	All caregivers of children 12-23	5.7% (4.7; 7.1)	35,238	9.6% (8.1; 11.3)	58,898	13.5% (11.0; 16.3)	82,595	57.5% (53.3; 61.5)	352,522	0.9% (0.5; 1.6)	5,546	12.8% (11.2; 14.6)	78,791
Note: Sample restricted to caregivers of children aged 12-23 months. Table entries refer to the estimated percentage and number of children with caregivers who responded to the following survey question: "How easy is it to pay for vaccination? When you think about the cost, please consider any payments to the clinic, the cost of getting there, plus the cost of taking time away from work." CIs are reported in parentheses.

Table 4.82: Perceived change in most recent vaccination service cost relative to prior cost

	Cumulative Proportion	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
Much more than the previous cost	25.0%	25.0%	±5.1	(20.2; 30.4)	21.0%	29.5%	21.07	127,112	±325,122	1,592
Slightly more than the previous cost	36.5%	11.5%	±3.0	(8.9; 14.8)	9.2%	14.2%	13.09	58,507	±123,904	828
About the same as the previous cost	77.9%	41.5%	±3.0	(38.5; 44.5)	39.0%	44.0%	5.72	210,951	±473,281	2,121
Slightly less than the previous cost	88.4%	10.4%	±3.7	(7.3; 14.7)	7.7%	13.9%	21.82	52,990	±95,550	640
Much less than the previous cost	100.0%	11.6%	±1.5	(10.2; 13.2)	10.5%	12.9%	3.25	59,270	±132,448	705
All		100.0%			99.9%	100.0%		508,829	±1,136,602	5,886
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 5,886
Note: Sample restricted to caregivers of children aged 12-23 months.

4.3.4.1 Direct Out-of-Pocket Costs

Figure 4.79 displays estimated percentage of households reporting any vaccination-related expenses, by type of expense. Nearly 75% of caregivers across LGAs reported spending some amount to get their children vaccinated. Medications (e.g., to treat vaccine side effects) were a common expense, affecting around half of households across survey areas. Other important costs included food and transportation (particularly in urban Nassarawa), each reported by a substantial share of respondents.

Expenses for registration, vaccination cards or supplies were instead rare, affecting less than one family in ten in all LGAs.

While health facilities rarely charge fees for registration and cards, families still face other costs that can add up.

Figure 4.79: Percentage of households reporting vaccination-related expenses, by line item

On average, households spent ₦946 ($0.62) per visit across LGAs, with costs ranging from ₦681 ($0.44) in Gaya to ₦1,824 ($1,19) in Nassarawa.²¹ The median cost was lower at ₦500 ($0.33), suggesting that a few outlier households are pulling up the average. This pattern is especially clear in Nassarawa, where the average cost reached ₦1,824 but the median was just ₦500. Gaya displays the lowest median at ₦200 ($0,13), alongside the lowest average.

Table 4.83: Average and median amount spent during the most recent vaccination visit, by survey area

Survey area	Average cost		Median cost
	Estimate	95% CI	Estimate	95% CI
Gabasawa	₦ 970	(₦ 888; ₦ 1,051)	₦ 500	(₦ 500; ₦ 600)
Gaya	₦ 690	(₦ 631; ₦ 748)	₦ 200	(₦ 200; ₦ 300)
Nassarawa	₦ 2,062	(₦ 1,023; ₦ 3,101)	₦ 500	(₦ 500; ₦ 700)
Non-sentinel LGAs	₦ 817	(₦ 684; ₦ 951)	₦ 500	(₦ 500; ₦ 800)
All LGAs	₦ 865	(₦ 725; ₦ 1,006)	₦ 500	(₦ 500; ₦ 800)
Note: Sample restricted to caregivers of children aged 12-23 months. Estimates refer to the total amount households spent, including transportation, fees, and other expenses. Amounts are expressed in Nigerian naira (₦).

Table 4.84: Average and median amount spent during the most recent vaccination visit, by survey area and zero-dose status of the household

Survey area	Zero-dose status (household)	Average cost		Median cost
		Estimate	95% CI	Estimate	95% CI
Gabasawa	Non-zero-dose household	₦ 985	(₦ 898; ₦ 1,072)	₦ 500	(₦ 500; ₦ 600)
	Zero-dose household	₦ 794	(₦ 585; ₦ 1,002)	₦ 150	(₦ 0; ₦ 500)
	All households with children 12-23	₦ 970	(₦ 888; ₦ 1,051)	₦ 500	(₦ 500; ₦ 600)
Gaya	Non-zero-dose household	₦ 679	(₦ 616; ₦ 742)	₦ 200	(₦ 200; ₦ 300)
	Zero-dose household	₦ 808	(₦ 637; ₦ 979)	₦ 300	(₦ 0; ₦ 600)
	All households with children 12-23	₦ 690	(₦ 631; ₦ 748)	₦ 200	(₦ 200; ₦ 300)
Nassarawa	Non-zero-dose household	₦ 2,101	(₦ 1,029; ₦ 3,174)	₦ 500	(₦ 500; ₦ 700)
	Zero-dose household	₦ 839	(₦ 519; ₦ 1,159)	₦ 700	(₦ 300; ₦ 1,200)
	All households with children 12-23	₦ 2,062	(₦ 1,023; ₦ 3,101)	₦ 500	(₦ 500; ₦ 700)
Non-sentinel LGAs	Non-zero-dose household	₦ 799	(₦ 675; ₦ 924)	₦ 500	(₦ 500; ₦ 800)
	Zero-dose household	₦ 1,232	(₦ 644; ₦ 1,820)	₦ 500	(₦ 100; ₦ 1,300)
	All households with children 12-23	₦ 817	(₦ 684; ₦ 951)	₦ 500	(₦ 500; ₦ 800)
All LGAs	Non-zero-dose household	₦ 852	(₦ 719; ₦ 984)	₦ 500	(₦ 500; ₦ 800)
	Zero-dose household	₦ 1,167	(₦ 667; ₦ 1,668)	₦ 500	(₦ 200; ₦ 1,300)
	All households with children 12-23	₦ 865	(₦ 725; ₦ 1,006)	₦ 500	(₦ 500; ₦ 800)
Note: Sample restricted to caregivers of children aged 12-23 months. Estimates refer to the total amount households spent, including transportation, fees, and other expenses. Amounts are expressed in Nigerian naira (₦). The term "zero-dose household" refers to a household where at least one age-eligible child has not received a Penta-1 vaccine.

Figure 4.80: Distribution of household expenditure on most recent vaccination visit, by survey area

Table 4.85 shows what categories drive household costs for vaccination. Across all LGAs, the category for food or other services²² costs (₦498, or $0.33) accounts for the largest share of household spending during vaccination visits (though the average in this category vary greatly across areas, and are especially high in Nassarawa). The food/other category is followed by medication (₦339, or $0.22) and transport (₦335, or $0.22).

Registration, supplies, and vaccination cards each added only small amounts, on average under ₦50.

These figures reinforce the pattern seen earlier: the bulk of vaccination-related spending comes from other costs rather than fees charged by health facilities. Such costs are not only more common, as shown in Table 4.83, but also higher on average.

Table 4.85: Average expenditure on vaccination visits, by LGA and line item

Survey area	Registration	Medication	Food	Supplies	Vaccination card	Transport	Total
Gabasawa	₦26 (20; 32)	₦360 (331; 389)	₦388 (328; 447)	₦27 (17; 36)	₦22 (11; 33)	₦359 (328; 390)	₦970 (888; 1051)
Gaya	₦34 (22; 45)	₦264 (240; 288)	₦270 (228; 313)	₦28 (22; 35)	₦35 (26; 43)	₦323 (289; 356)	₦690 (631; 748)
Nassarawa	₦47 (34; 61)	₦522 (440; 604)	₦5184 (-907; 11274)	₦43 (19; 66)	₦33 (22; 44)	₦485 (409; 561)	₦2062 (1023; 3101)
Non-sentinel LGAs	₦23 (10; 35)	₦283 (225; 341)	₦295 (240; 350)	₦4 (2; 5)	₦7 (3; 11)	₦293 (231; 354)	₦817 (684; 951)
All LGAs	₦24 (12; 36)	₦294 (242; 346)	₦482 (250; 713)	₦7 (5; 9)	₦10 (6; 14)	₦303 (246; 360)	₦865 (725; 1006)
Note: Sample restricted to caregivers of children aged 12-23 months. Estimates refer to the total amount households spent, including transportation, fees, and other expenses. Amounts are expressed in Nigerian naira (₦).

4.3.4.2 Indirect Costs

This section sheds light on the more indirect costs households face for vaccination.

For more than eight visits out of ten, only one person made the trip (likely the primary caregiver and child), while about 11% involved two people (Table 4.86). While most households limit the number of travelers, the fact that one in ten visits involved at least two people means transport and food costs can double or more for those families.

Table 4.86: Number of household members traveling for the most recent vaccination visit

	Cumulative Proportion	Proportion	MOE (Prop)	95% CI	LCB	UCB	DEFF	Estimated Total	MOE (Total)	Unweighted Cases
1	87.2%	87.2%	±3.5	(83.2; 90.3)	83.9%	89.9%	22.82	591,539	±1,389,850	6,545
2	96.8%	9.6%	±2.1	(7.6; 11.9)	7.9%	11.5%	10.69	64,961	±133,780	1,011
3	99.1%	**	**	**	**	**	51.10	15,540	±63,771	158
4	100.0%	0.9%	±1.2	(0.3; 3.3)	0.3%	2.7%	31.30	6,386	±27,711	93
All		100.0%			100.0%	100.0%		678,426	±1,599,202	7,807
Abbreviations: CI = Confidence Interval; MOE = Margin of Error; LCB = Lower Confidence Bound; UCB = Upper Confidence Bound; DEFF = Design Effect; Domain sample size = 7,807; Potentially unreliable values have been suppressed and marked with asterisks (reason(s): The absolute CI width is ≥5% and the relative CI width is ≥130%)
Note: Sample restricted to caregivers of children aged 12-23 months.

Vaccination visits require a considerable time investment, as shown in Table 4.87. This includes travel time, waiting time, time spent at the health facility after receiving the vaccine, and the time to get back home. For caregivers who work, this can involve a significant tradeoff with time that could have been spent on other income-generating activities, adding to the indirect cost of vaccination. On average, households spent between 5.7 and 6.3 hours for a vaccination visit, with Gabasawa and non-sentinel LGAs showing the longest average times (at 6.3 and 6.2 hours, respectively), and Nassarawa showing the lowest (at 5.7 hours). The differences in average time between survey areas is statically significant in a regression-based Wald test \((F_{(3, 496)}=43.2\), \(p<.001)\).

Table 4.87: Average time spent for vaccination, by survey area and activity type

Survey area	Travel to the health facility	Wait to receive vaccine	Wait after receiving vaccine	Return home	Total
Gabasawa	1.4 hours (1.3; 1.4)	2.1 hours (2.1; 2.2)	1.4 hours (1.4; 1.5)	1.4 hours (1.3; 1.4)	6.3 hours (6.2; 6.4)
Gaya	1.3 hours (1.3; 1.3)	2 hours (2; 2.1)	1.4 hours (1.3; 1.4)	1.3 hours (1.3; 1.4)	6 hours (5.9; 6.1)
Nassarawa	1.1 hours (1.1; 1.2)	2 hours (1.9; 2.1)	1.4 hours (1.4; 1.4)	1.2 hours (1.2; 1.2)	5.7 hours (5.6; 5.8)
Non-sentinel LGAs	1.2 hours (1.2; 1.3)	2.3 hours (2.2; 2.4)	1.5 hours (1.4; 1.5)	1.3 hours (1.2; 1.3)	6.2 hours (6.1; 6.4)
All LGAs	1.2 hours (1.2; 1.3)	2.3 hours (2.2; 2.3)	1.5 hours (1.4; 1.5)	1.3 hours (1.2; 1.3)	6.2 hours (6.1; 6.3)
Note: Sample restricted to caregivers of children aged 12-23 months. The calculations include the time spent traveling to the health facility (q511) for the most recent vaccination, waiting (q512), as well as the additional time they spent at the site after receiving it (q513) and traveling back home (q514). CIs are reported in parentheses.

Bednarczyk, Robert A, Kyra A Hester, Sameer M Dixit, Anna S Ellis, Cam Escoffery, William Kilembe, Katie Micek, Zoë M Sakas, Moussa Sarr, and Matthew C Freeman. 2022. “Exemplars in Vaccine Delivery Protocol: A Case-Study-Based Identification and Evaluation of Critical Factors in Achieving High and Sustained Childhood Immunisation Coverage in Selected Low-Income and Lower-Middle-Income Countries.” BMJ Open 12 (4): e058321. https://doi.org/10.1136/bmjopen-2021-058321.

Carter, Austin, Nadia Akseer, Kevin Ho, Oliver Rothschild, Niranjan Bose, Agnes Binagwaho, Lisa R Hirschhorn, et al. 2020. “A Framework for Identifying and Learning from Countries That Demonstrated Exemplary Performance in Improving Health Outcomes and Systems.” BMJ Global Health 5 (12): e002938. https://doi.org/10.1136/bmjgh-2020-002938.

Center for International Earth Science Information Network (CIESIN), Columbia University. 2024. “GRID3 NGA - Settlement Extents v3.1.” New York: GRID3. https://doi.org/10.7916/x9xg-e262.

Florczyk, Aneta J., Christina Corbane, Daniele Ehrlich, Sergio Freire, Thomas Kemper, Lorenzo Maffenini, Michele Melchiorri, et al. 2019. “GHS Settlement Model grid (GHS‑SMOD), Global Human Settlement Layer Data Package 2019 (GHS P2019).” EUR 29788 EN (JRC 117104). European Commission, Joint Research Centre (JRC). https://doi.org/10.2760/290498.

Kazi, Muhammad Ahmed, Sajid Bashir Soofi, Imtiaz Hussain, Masawar Hussain, Muhammad Umer, Ahmad Khan, Khalid Feroz, et al. 2024. 2023 Supplementary Immunization Coverage Survey in Super High Risk Union Councils of Pakistan (TPVICS-SHRUCs Rounds 1–3). Aga Khan University; Survey Report. Book 3. https://ecommons.aku.edu/pakistan_coe-wch_survey_report/3.

LaFond, Anne, Natasha Kanagat, Robert Steinglass, Rebecca Fields, Jenny Sequeira, and Sangeeta Mookherji. 2015. “Drivers of Routine Immunization Coverage Improvement in Africa: Findings from District-Level Case Studies.” Health Policy and Planning 30 (3): 298–308. https://doi.org/10.1093/heapol/czu011.

Mindset. 2026. “Methods for Estimating Vaccination Coverage and Zero-Dose Prevalence: Head-to-Head Comparison Report.” Mindset. https://zd-head2head.netlify.app/.

Pathways. n.d. “Segmentation.” n.d. https://www.projectpathways.org/segmentation.

Phillips, David E, Joseph L Dieleman, Stephen S Lim, and Jessica Shearer. 2017. “Determinants of Effective Vaccine Coverage in Low and Middle-Income Countries: A Systematic Review and Interpretive Synthesis.” BMC Health Services Research 17 (1): 681. https://doi.org/10.1186/s12913-017-2626-0.

WHO, and UNICEF. 2022. “WUENIC—World Health Organization/UNICEF Estimates of National Immunization Coverage.” 2022. https://immunizationdata.who.int/.

---

1. Unlike for other vaccines, the survey instrument did not ask caregivers to recall meningitis, Vitamin A or YF on a named vaccine-by-vaccine basis. Instead, for these three vaccines, caregivers were asked to list any other vaccines that they could recall after enumerators attempted to review HBRs and after enumerators asked about other explicitly named vaccines. Because of this difference in instrumentation, coverage rates for meningitis, Vitamin A and YF may be slightly lower than if recall questions for these vaccines had followed the same named format as the other antigens.

2. The degree of urbanization in this analysis is based on the open-source Global Human Settlement Layer Settlement Model (SMOD) geospatial layer from the Global Human Settlement Layer (GHSL) project (Florczyk et al. 2019).

3. MCV-2 is excluded from the analysis because not all children aged 12–23 months are yet eligible. Although it can be included, doing so requires care to avoid unintentionally restricting the analysis domain for other vaccine pairs.

4. The authors have chosen to display the probability of receiving a Penta-1 vaccine prior to six weeks even though, in theory, most children should not receive this vaccine prior to this point. The reason for this is threefold. First, it serves as a visual validation check. Second, it showcases that a small proportion of caregivers report that their child received the vaccine prior to six weeks (whether true or not). Third, this approach aligns with the standard approach of reporting survival curves used by VCQI.

5. The prevalence maps in this section use unweighted indicator kriging models to estimate the prevalence of vaccine coverage. GRID3 settlement extents are used to mask uninhabited areas from the map (Center for International Earth Science Information Network (CIESIN), Columbia University 2024).

6. These cutoffs were arbitrarily chosen by the research team but can easily be changed. Although overlap is possible, the plot shows that these two approaches tend to highlight different geographic areas.

7. In the footnotes of its 2023-24 report, the NDHS notes that it considers children who have received a combination of different polio vaccines to be fully immunized if the other basic antigens have been fully administered. This less orthodox approach to encoding polio differs from the conventional approach advocated by the World Health Organization (WHO) (which requires two doses of IPV and three doses of OPV), but for the present study, does not make a large difference in the findings. A trivially small number of children would be classified differently using this more complex polio encoding rule, and for this reason, only the conventional basic antigen definition is presented here.

8. Selection and inference are both design-weighted: at each step a candidate’s contribution is measured by its design-based Wald \(\chi^2\) statistic, and a term is admitted (or kept) iff its Wald \(\chi^2\) exceeds twice its degrees of freedom — the AIC-equivalent criterion under the \(\chi^2\) approximation \(W \approx -2\log\Lambda\).

9. A Brant test of the proportional-odds assumption is significant (p <0.001), confirming that effects differ across thresholds. The test is computed on an unweighted MASS::polr fit because no design-based Brant test exists; the violation magnitude is large enough that the unweighted test remains informative. A design-based partial-proportional-odds variant that would relax the assumption does not converge on this sample (the unweighted version, Table 4.63, does), so the proportional-odds specification is retained for inference.

10. For the designation of urban/rural households, the team relied on the official classification approved by the Statistician General of the Kano Bureau of Statistics (Appendix E).

11. The Pathways decision tree was applied with several adaptations to align with the design of the household survey. First, because men were also respondents, the fertility question was phrased as “Have you, or a spouse of yours, ever given birth?” rather than the original “Have you ever given birth?” Second, for sanitation, the original split grouped all non-flush toilets together, but Mindset revised it to distinguish flush toilets from pit latrines, bucket toilets, and no facility/bush/field. Third, fuel use was collected as a “select multiple” item rather than “select one.” This created ambiguity for households reporting more than one fuel type (e.g., wood and gas). To resolve this, Mindset classified households based on the more advanced fuel type present, so a household reporting both wood and natural gas would be coded as using natural gas when the tree split on fuel type.

12. These figures are reported for illustrative purposes and come from a back-of-the-envelope calculation of weighted dropout rates (for instance, \((62.7-59.3)/62.7\approx0.05\) for DPT1–3).

13. Note: The Intraclass Correlation Coefficient (ICC) displayed in the Vaccination Coverage and Timeliness Chart (VCTC) plots currently has a different interpretation for sentinel and non-sentinel areas. For sentinel LGAs, the ICC represents the expected degree of homogeneity within a building. For non-sentinel LGAs, it reflects the expected degree of homogeneity within an EA. More meaningful and design-aligned analyses of ICC can be found in (Mindset 2026)

14. Note: The ICC displayed in the VCTC plots currently has a different interpretation for sentinel and non-sentinel areas. For sentinel LGAs, the ICC represents the expected degree of homogeneity within a building. For non-sentinel LGAs, it reflects the expected degree of homogeneity within an EA. More meaningful and design-aligned analyses of ICC can be found in (Mindset 2026)

15. Children aged 6–23 months are used rather than the more conventional 12–23 months to maximize the available training data and thereby increase the model’s statistical power to detect predictors of zero-dose status. Zero-dose classification is relatively stable between 6 and 12 months, and there is no strong a priori reason to expect the underlying drivers of non-vaccination to differ meaningfully between these sub-groups.

16. The regression was restricted to sentinel LGAs because survey weights differ substantially between sentinel and non-sentinel areas: the non-sentinel LGAs are each represented by far fewer sampled households relative to their actual population weight, which would give those observations disproportionate leverage over the model fit. Restricting to the sentinel areas yields a more balanced and reliable basis for estimation. This parallels the same rationale used elsewhere in this report to focus spatial hotspot analyses on sentinel areas.

17. Candidate predictors were drawn from the full household and caregiver questionnaire. Variables were included only if at least 90% of values were present; for the remaining gaps, missing values were filled in by drawing on responses from the ten most similar children, where similarity was assessed across the categorical survey items. Variables with near-zero variance, high cardinality, or a trivially redundant relationship to zero-dose status (such as possession of a vaccine card, administrative bookkeeping fields, and interview timing variables) were excluded before the selection procedure. After these screens, candidate predictors entered the variable-selection procedure. A pseudo-R-squared circuit-breaker was applied as a stopping rule: variable admission was governed by improvement in the design-based AIC, while the loop terminated as soon as a step’s gain in Nagelkerke pseudo-R-squared fell below 0.01. The variable that triggered termination was retained, since it was selected on AIC grounds; subsequent steps were not attempted. These screening and gap-filling steps were computed once across the full survey sample for efficiency, before restricting the regression to sentinel children. As a sensitivity check we re-ran the entire procedure restricted to sentinel children only: the missingness screen retained 185 candidate variables in the full-sample computation but 187 when restricted to sentinel children, but the variable-selection routine chose the same predictors out of either pool, and the resulting odds-ratio estimates were numerically identical to those reported here.

18. Forward stepwise selection is a useful exploratory tool when many candidate variables are available: it heuristically identifies variables most strongly associated with the outcome, filtering out predictors whose apparent association becomes negligible once other variables are controlled for. Its limitations are worth noting: depending on implementation, forward selection can overfit, picking up spurious associations that may not generalise beyond the training data. It also optimises for predictive fit, which is not always equivalent to what is most useful for explanation or policy — a model that predicts well may include proxy variables and miss clinically or programmatically relevant predictors. Some human review of the selected terms is therefore warranted alongside any statistical interpretation.

19. The cross-validation procedure randomly partitioned the 32 sentinel survey strata into five folds, with roughly 6 to 7 strata per fold. In each fold, one grouped holdout fold was used as the test set while the model formula was refit on the remaining 25 to 26 strata; predictions were then generated for the held-out fold. The “All areas” row reports each metric computed once over the held-out predictions concatenated across all five folds, rather than averaged across the per-fold values. This provides an honest estimate of how well the model generalizes to unseen geographic areas within Kano State, and guards against over-optimistic in-sample performance estimates.

20. This squared correlation is computed between predicted probabilities and observed zero-dose indicators across all held-out strata, and is typically somewhat lower than the Nagelkerke pseudo-R-squared reported in Table 4.74. The Nagelkerke pseudo-R-squared is a rescaled in-sample goodness-of-fit measure normalized to the [0, 1] range, which tends to produce higher values than the cross-validated equivalent. The two metrics are therefore not directly comparable; the cross-validated squared correlation is the more conservative and externally valid estimate of how tightly the model’s predictions track the observed outcome.

21. Exchange rate of 1531.6 Naira to the US Dollar as of July 7th, 2025.

22. In the survey questionnaire, response option q505 was coded as “Food or other services,” which prevents distinguishing expenditures on food from those in the residual “Other” category. A more informative design would have separated these items.