Table 4.

Cluster analysis of secondary outcomes (cascada) among parents of children aged 9-60 months (t-test of difference between 18 intervention and 14 control sites, cluster analysis as in Tables 2 and 3).

	Intervention clusters	Control clusters	Outcome in primary (cluster) analysis (t-test)	GEE adjusting for baseline differences
Respondent could mention an illness preventable by mean 0.74	2368/3153 mean 0.74	1437/2431 mean 0.58	Difference 0.17, 95% 0.067-0.272 t = 3.369, 30 df p = 0.002	Difference 0.121 95% 0.055-0.189 (1)
Do you think it's worthwhile to vaccinate children? (attitude)	3006/3161 mean 0.95	2116/2475 mean 0.84	Difference 0.11, 95% 0.021-0.197 t = 2.543, 30 df p = 0.016	Difference 0.054 95% 0.013-0.105 (1)
Do your neighbours think it's worthwhile to vaccinate children? (subjective norm)	2842/3166 mean 0.89	1884/2475 mean 0.74	Difference 0.15, 95% 0.039-0.260 t = 2.755, 30 df p = 0.010	Difference 0.095 95% 0.011-0.182 (1)
How much time are you prepared to spend to take a child from your household to be vaccinated? Willing to take some time (intention)	2954/3088 mean 0.95	2037/2317 mean 0.84	Difference 0.11, 95% 0.002-0.227 t = 2.086, 30 df p = 0.046	Difference 0.073 95% 0.015-0.156 (2)
Mother included in decisions about vaccination (agency)	1834/3131 mean 0.59	1345/2434 mean 0.54	Difference 0.04, 95% -0.024-0.108 t = 1.299, 30 df p = 0.204	Difference 0.043 95% -0.009-0.097 (NS) (3)
Have you discussed vaccination for children in the family? (discussion)	1584/3155 mean 0.49	826/2459 mean 0.31	Difference 0.19, 95% 0.054-0.318 t = 2.868, 30 df p = 0.007	Difference 0.155 95% 0.032-0.270 (2)

Initial model: intervention, willingness to travel, access, knowledge of preventable disease.

(1) Final model: intervention, willingness to travel, access.

(2) Final model: intervention, knowledge of preventable disease.

(3) Final model: intervention, knowledge of preventable disease, access.