Comparative effectiveness of school-based caries prevention: a prospective cohort study

BMC Oral Health

Table 3 Model results comparing caries prevention programs over time

	TOCE				Untreated decay (permanent teeth)
3a: Coefficients (GEE and multilevel models)	IRR	p	95% CI		OR	p	95% CI
Visit (trend)	1.20	< .001	1.18	1.23	0.77	< .001	0.72	0.82
Exposed (vs unexposed)	0.95	0.274	0.88	1.04	0.77	0.034	0.60	0.98
Baseline TOCE	1.45	< .001	1.43	1.46	1.12	< .001	1.09	1.15
Age at exam	1.01	0.038	1.00	1.03	1.27	< .001	1.21	1.33
Sex (Male)	0.99	0.517	0.95	1.03	1.06	0.468	0.90	1.25
Previous dental care	1.42	< .001	1.35	1.49	0.93	0.457	0.78	1.12
Water fluoridation	1.15	< .001	1.09	1.20	1.07	0.486	0.88	1.29
Treatment*Visit	0.96	0.001	0.94	0.98	–	–	–	–
3b: Coefficients (smoothed GAM)^a	IRR	p	95% CI		OR	p	95% CI
Baseline TOCE	1.26	< .001	1.25	1.26	1.15	< .001	1.12	1.19
Sex (Male)	0.99	0.330	0.98	1.01	1.09	0.980	0.92	1.23
Previous dental care	1.43	< .001	1.40	1.45	0.79	0.020	0.65	0.96
Water fluoridation	1.12	< .001	1.11	1.14	0.94	0.510	0.79	1.13

^aNotes: Exposed group includes participants receiving primary and secondary prevention. Unexposed group includes participants receiving primary prevention only. Parametric effects for age, exposure (treatment), and visit (non-smoothed) are shown for standard GEE (adult decay) and multilevel (TOCE) models for comparison purposes (model 3a). Age, visit, exposure (treatment), and the visit-treatment interaction were included in model 3b as smoothed terms in GAMs and do not have estimated parametric coefficients. Nonlinear effects for these terms in the generalized additive model are shown in Table 4
TOCE total observed caries experience, GEE generalized estimating equations, GAM generalized additive models, IRR incident rate ratio, OR odds ratio

ISSN: 1472-6831