African American (AA) Effect from Multivariate Cox Proportional Hazard Regression of Time to Registration with KP.org: Stratified by Condition Cohort1 and Level of Education2
| N | Percent AA | Hazard Ratio (HR) for AA Vs. White | Upper 95% CI | Lower 95% CI | p Value for HR = 0.00 | p Value for Difference in HR∗ | |
| Study population stratified by condition cohort | |||||||
| Diabetes | 505 | 55% | 0.63 | 0.86 | 0.47 | <0.01 | 0.34 |
| Elevated lipids | 621 | 45% | 0.72 | 0.96 | 0.55 | 0.02 | 0.19 |
| Low Risk | 651 | 47% | 0.58 | 0.80 | 0.42 | <0.01 | — |
| Study population stratified by level of education | |||||||
| HS or less | 356 | 53% | 0.56 | 0.88 | 0.36 | 0.01 | 0.27 |
| Some College | 600 | 58% | 0.64 | 0.87 | 0.47 | <0.01 | 0.44 |
| College Graduate | 343 | 49% | 0.66 | 0.98 | 0.44 | 0.04 | 0.73 |
| Postgraduate College | 444 | 35% | 0.67 | 0.93 | 0.49 | 0.02 | — |
-
↵1 Within the condition cohort models, independent variables include race/ethnicity, activation, level of education, age group, gender, and 2005 survey response (Internet vs. mail).
-
↵2 Within the level of education models, independent variables include race/ethnicity, activation, condition cohort, age group, gender, and 2005 survey response (Internet vs. mail).
-
↵∗ Differences in HR by condition cohort or level of education tested by t-statistic (1df).









