The impact of education programs on smoking prevention: a randomized controlled trial among 11 to 14 year olds in Aceh, Indonesia

Table 3 Impact of the health and Islamic-based interventions and their interactions on knowledge and attitude

Outcomes	Main effect of health				Main effect of Islam				Health × Islam Interaction effect		ICC⁵
	Health (n = 239)	Non health (n = 237)	β ± SE	p value¹	Islam (n = 226)	Non Islam (n = 250)	β ± SE	p value²	β ± SE³	p value⁴
Health Knowledge (mean ± SD)
Visit 1	6.7 ± 2.6	8.2 ± 2.7			7.3 ± 2.7	7.6 ± 2.7
Visit 2	11.9 ± 2.8	9.8 ± 2.9	3.9 ± 0.6	< 0.001	11.9 ± 2.4	9.8 ± 3.1	3.8 ± 0.6	< 0.001	−3.4 ± 0.9	< 0.01	0.10
Islamic Knowledge (mean ± SD)
Visit 1	10.0 ± 3.0	11.7 ± 2.5			11.0 ± 2.7	10.7 ± 3.1
Visit 2	11.5 ± 2.7	12.9 ± 3.0	0.1 ± 0.6	0.88	13.6 ± 2.6	11.0 ± 2.8	3.5 ± 0.5	< 0.001	−2.0 ± 0.8	0.02	0.08
Attitude (mean ± SD)
Visit 1	48.7 ± 12.7	41.8 ± 10.0			43.9 ± 10.6	46.4 ± 13.0
Visit 2	42.9 ± 11.0	40.3 ± 10.6	−3.0 ± 1.9	0.14	38.6 ± 10.3	44.3 ± 10.7	−7.1 ± 1.5	< 0.001	5.2 ± 3.0	0.11	0.08

Note. Results for the main effects were obtained using a generalized linear model for the outcome with adjustment for baseline values and use of robust standard errors to account for the clustering within both classrooms and schools. In all analyses, the 4 groups were recoded into two factors: (1) Health, and (2) Islamic. Codes of one or zero were used for each factor depending on whether the subjects’ group was one that contained the respective health or Islamic factor. That is, control group (Islamic = 0, Health = 0), Islamic (Islamic = 1, Health = 0), Health (Health = 1, Islamic = 0), combined (Health = 1, Islamic = 1).
¹ Baseline adjusted effect of Health.
² Baseline adjusted effect of Islam.
³The interaction effect represents the additional effect of being in the combined group beyond the separate main effects presented for Health and Islam.
⁴P value for interaction between Health × Islam.
⁵ICC = Intra-class correlations coefficients, from mixed effects random intercept model.

ISSN: 1471-2458