TABLE 3: Simple and adjusted models to evaluate the adolescent risk of die by acquired immunodeficiency syndrome in Rio de Janeiro City.

	Simple models		Adjusted model
	OR	95% CI	OR	95% CI
Fixed effects
Male	2.05	1.32-3.22	2.12	1.36-3.32
Age	0.88	0.78-1.01
Nonwhite race	0.65	0.39-1.10
Year	0.86	0.84-0.88
Educational level
< Elementary school	1.00		1.00
Elementary school	0.41	0.24-0.67	0.42	0.25-0.69
Middle/High school	0.27	0.09-0.65	0.37	0.13-0.94
Transmission
Blood vs. others
Transfusion	1.16	0.07-19.88
Hemophilia	1.56	0.46-5.25
Injectable drugs	2.90	1.04-7.83	3.15	1.29-7.58
Sexual vs. others	0.12	0.06-0.22	0.08	0.04-0.13
Vertical vs. others	0.25	0.07-0.75	0.11	0.03-0.33
Random effects			Mean
Intercept (neighborhood)			18,629	1,266-67,205
Slope (year)			18,527	1,242-67,031
Slope (HDI)			18,630	1,268-67,203
Deviance			578

OR: odds ratio; CI: confidence interval; HDI: Human Development Index; Year: diagnostic year.

Bayesian multilevel binomial models were used to estimate the mortality AIDS odds in adolescents. Both for simple and adjusted models, the response variable was the deaths. Fix effects estimated in the adjusted model were age, sex, educational level, and types of transmission (sexually transmitted, transmitted by the administration of injectable drugs, and transmitted by vertical and blood routes). In both models, we used random effects for neighborhood, diagnostic year, and HDI.