Table 2.
Group Differences on Support, Participation, and Quality-of-life Variables at Two-month Follow-up — Results from Regression Models (N = 246)*
| Means (Covariate Adjusted) | Difference Between Means | 95% CI† of Difference | |||||
|---|---|---|---|---|---|---|---|
| Outcome Variables | CHESS | Control | CHESS-Control | Lower | Upper | Main Effect P Value‡ | Interaction |
| Social/information support | |||||||
| 1. Social support | 80.8 | 78.4 | 2.4 | −1.2 | 5.9 | ||
| 2. Information competence | 70.4 | 65.6 | 4.8 | 1.5 | 8.1 | <.01 | |
| 3. Unmet information need¶ | 70.0 | 67.2 | 2.8 | −2.7 | 8.4 | Race,§ Education§ | |
| Participation in health care | |||||||
| 1. Participation: behavioral involvement | 75.6 | 73.1 | 2.5 | −1.1 | 6.1 | Race,§ Education§ | |
| 2. Participation: level of comfort | 80.7 | 74.3 | 6.4 | 2.1 | 10.7 | <.01 | |
| 3. Confidence in doctor(s) | 83.0 | 77.3 | 5.7 | 1.0 | 11.3 | <.05 | |
| Quality of Life | |||||||
| 1. Social/family well-being | 79.3 | 78.2 | 1.1 | −3.0 | 5.4 | ||
| 2. Emotional well-being | 73.9 | 72.8 | 1.1 | −3.1 | 5.2 | ||
| 3. Functional well-being | 62.2 | 63.0 | −0.8 | −5.3 | 3.6 | Insurance∥ | |
| 4. Breast cancer concerns¶ | 65.1 | 63.3 | 1.8 | −1.9 | 5.5 | Insurance¶ | |
Follow-up scores were statistically adjusted for pretest scores, days since diagnosis, stage of breast cancer, race, education, insurance status, chemotherapy status, and Karnofsky's performance status using the method of recycled predictions.32 All scores were linearly transformed to a 0–100 scale.
CI, confidence interval.
Only P values smaller than .05 are shown. For behavioral involvement and level of comfort, Holm's method requires the smallest P value to less than .025 and the next to exceed .05. For the four quality-of-life measures, Holm's method requires the smallest P value to less than .0125, the next smallest to less than .0167, the third .025, and the fourth .05.
Indicates a significant interaction (using same multiple comparison corrections as for main effects) between experimental condition and patient characteristics. See Figures 2 and 3 for details.
Indicates an interaction whose P value was less than .05, but was not significant because of the correction for multiple comparisons. The trend in these cases was always in the same direction as that in effects reported as significant: greater CHESS benefit. for the underserved.
High scores indicate less breast cancer–related concerns and less unmet information need.