Skip to main content
. 2016 Sep 20;16:494. doi: 10.1186/s12879-016-1835-1

Table 2.

Comparison of tuberculosis treatment outcome between patients receiving and not receiving directly observed treatment, separated by census tract of residence: slum (aglomerados subnormais - AGSN) and non-slum (non-AGSN), Rio de Janeiro, Brazil, 2010

AGSN Non-AGSN P-value
DOT (%) RR§ 95 % CI DOT (%) RR§ 95 % CI
Total Yes No Total Yes No
Missing outcome 174 (10.2) 39 (6.1) 135 (12.7) 0.56 0.42-0.74 462 (10.2) 98 (7.9) 364 (11.0) 0.77 0.64-0.92 0.06
Cure 963 (56.7) 430 (67.4) 533 (50.1) 1.66 1.45-1.91 2622 (57.8) 758 (61.4) 1864 (56.5) 1.15 1.04-1.27 <0.01
Abandonment 325 (19.1) 118 (18.5) 207 (19.5) 0.97 0.83-1.14 746 (16.5) 265 (21.5) 481 (14.6) 1.37 1.23-1.53 <0.01
Death TB 68 (4.0) 8 (1.3) 60 (5.6) 0.20 0.08-0.46 199 (4.4) 28 (2.3) 171 (5.2) 0.54 0.38-0.76 0.03
Death other 32 (1.9) 9 (1.4) 23 (2.2) 0.68 0.37-1.24 128 (2.8) 24 (1.9) 104 (3.2) 0.65 0.45-0.95 0.92
Transfer out 140 (8.2) 34 (5.3) 106 (10.0) 0.57 0.41-0.78 378 (8.3) 61 (4.9) 317 (9.6) 0.60 0.47-0.75 0.82

*There were 88 (4.4 %) cases in AGSN and 219 (4.4 %) outside of AGSN where the case’s DOT status was unknown

§Relative risk (RR) estimating crude risk of outcome in patients undergoing DOT compared to patients not undergoing DOT in slum versus non-slum neighborhood

P-values assessing for effect measure modification of AGSN census tracts on the relationship between DOT and respective outcome with the Mantel-Haenszel chi-square test