Research Article
First Use of Multiple Imputation with the National Tuberculosis Surveillance System
Table 1
Predictor variables included in the imputation model, with percent of nonmissing observations and
for association with the outcome.
| Variable | No. evaluated (%)
| |
| Isoniazid resistance | 1614 (100) | 0.02 | Age categories | 1614 (100) | <0.01 | Race | 1614 (100) | <0.01 | Occupation | 1614 (100) | <0.01 | Gender | 1614 (100) | <0.01 | HIV positive | 849 (53) | <0.01 | Injecting drug use within the previous year | 1367 (85) | <0.01 | Drug use (noninjecting) within the previous year | 1351 (84) | <0.01 | Alcohol use within the previous year | 1340 (83) | <0.01 | Homeless within the previous year | 1500 (93) | 0.18 | Abnormal chest radiograph | 1525 (94) | <0.01 | Resident of a correctional facility at diagnosis | 1604 (99) | 0.03 | Foreign birth | 1600 (99) | <0.01 | Positive tuberculin skin test | 953 (59) | <0.01 | Positive AFB smear from a nonsputum specimen | 1419 (88) | 0.12 |
|
|