Table 2. Multiple Logistic Regression Analysis of Hepatitis B Screening Behavior at 6-Month Follow-up Among Previously Unscreened Participants (n = 446), Study of the Effect of a Liver Cancer Education Program on Hepatitis B Screening Among Chinese, Korean, and Vietnamese Americans in the Baltimore–Washington Metropolitan Area, 2009–2010.
| Variable | Total | Intervention group | Control group |
|---|---|---|---|
| Number of clusters | 27 | 15 | 12 |
| Number of participants | 446 | 220 | 226 |
| Percentage screened for hepatitis B | 21.5 | 33.6 | 9.7 |
|
| |||
| Intervention effect |
Adjusted Odds Ratio
(95% Confidence Interval)a |
P Valuea | |
|
| |||
| Model 1b | |||
| Control group | 1 [Reference] | <.001 | |
| Received intervention | 4.63 (2.30–9.32) | ||
| Model 2b | |||
| Control group | 1 [Reference] | <.001 | |
| Received intervention | 5.13 (3.14–8.39) | ||
| Ethnicity | |||
| Korean | 1 [Reference] | <.001 | |
| Chinese (vs Korean) | 3.00 (1.66–5.41) | ||
| Vietnamese (vs Korean) | 3.59 (2.04–6.34) | ||
a
Odds ratios, 95% confidence intervals, and P values were obtained from multiple logistic regression models comparing intervention and control groups adjusted for age and clustering within randomized groups for both models.
b
Calculated by using generalized estimating equation. Model 1 used generalized estimating equation to adjust for the cluster effect and age, and Model 2 further adjusted for ethnicity.