Page 63 - Contributed Paper Session (CPS) - Volume 8

CPS2184 M Lutfor Rahman
Diabetes Yes 0.840 0.04 2.316
mellitus 3 (1.028,5.2
17)*
Chronic No 1
renal failure Yes -0.203 0.60 0.816
1 (0.381,1.7
48)
Exposure characteristics
Contact Casual 1
type
1.585 0.00 4.877 1.920 0.00 6.819
Household 0 (2.550,9.3 0 (2.906,16.0
27)** 00)**
Exposure 0.001 0.01 1.001 -0.084 0.03 0.919
duration 6 (1.000,1.0 7 (0.8490.995
(hours) 02)* )*
Sleeping No 1
together Yes 1.681 0.00 5.373
1 (2.053,14.0
64)**
Eat No 1
together Yes 0.540 0.02 1.716
6 (1.067,2.7
60)*
Size of the Large 1
exposure Small 0.105 0.72 1.110
site 8 (0.615,2.00
4)NS
Yes 1
Ventilation No -0.157 0.70 0.855
of the 9 (0.374.1.95
exposure 3)NS
site
Note: * means significant at 5% level, ** means significant at 1% level

The results of fitting the hierarchical and network logistic regression models
to the TB data are shown in Table 5. In Table 5, the results from hierarchical
model show that the covariates age, diabetes mellitus, and contact type are
significant at 1% or 5% level of significance. Whereas, the variables age, and
contact type are found to be significant in network logistic regression analysis
at 5% level. The network logistic regression considers the structural position
of the covariates in a network. Interestingly, household contacts are 38.016
times more likely to have TB infection than the casual contacts. The results in
network logistic regression provide more precise selection of variables that
can be emphasized by the public health community in the real-life practice.

52 | I S I W S C 2 0 1 9