Table 3.

Logistic regression models predicting risk of cure (model 1), abandonment of treatment (model 2) and death (model 3) from tuberculosis among all TB cases in Rio de Janeiro, Brazil, 2010

	Odds ratio	Standard error	95 % confidence interval
Model 1: risk of cure*
On directly observed treatment (DOT)	1.16	0.12	0.95, 1.42
Residence in a slum (AGSN) area	0.77	0.08	0.63, 0.95
Sex (reference female)	0.73	0.06	0.62, 0.85
Age in years	1.01	0.002	1.00, 1.01
Extrapulmonary clinical disease	2.07	0.23	1.66, 2.59
HIV/AIDS	0.23	0.02	0.19, 0.28
DOT x AGSN (interaction term)	1.67	0.31	1.17, 2.40
Model 2: risk of abandonment§
On directly observed treatment (DOT)	1.15	0.13	0.92, 1.43
Residence in a slum (AGSN) area	1.33	0.16	1.05, 1.67
Sex (reference female)	1.41	0.13	1.18, 1.69
Age in years	0.98	0.003	0.97, 0.98
Extrapulmonary clinical disease	0.44	0.06	0.33, 0.57
HIV/AIDS	1.97	0.23	1.56, 2.48
DOT x AGSN (interaction term)	0.61	0.12	0.41, 0.90
Model 3: risk of death†
On directly observed treatment (DOT)	0.39	0.11	0.23, 0.67
Residence in a slum (AGSN) area	1.23	0.23	0.85, 1.78
Age in years	1.03	0.01	1.02, 1.04
Extrapulmonary clinical disease	0.64	0.13	0.43, 0.95
HIV/AIDS	8.79	1.45	6.35, 12.16
DOT x AGSN (interaction term)	0.10	0.10	0.01, 0.77

AGSN aglomerados subnormais

^*N: 3797, R² = 0.061

^§N: 3797, R² = 0.047

^†N: 3797, R² = 0.175