Table 1.
Fixed and random effects estimates for models of depression, health-related quality of life, and life satisfaction.
| Parameter | Depression | General | Cognitive | Pain | Energy/vitality | QOLI |
|---|---|---|---|---|---|---|
|
|
|
|
|
|
|
|
| γ(SE) | γ (SE) | γ (SE) | γ (SE) | γ (SE) | γ (SE) | |
| Fixed effects | ||||||
| Intercept | 18.15 (1.24)*** | 50.74 (3.46)*** | 62.75 (3.61)*** | 65.40 (3.74)*** | 43.62 (3.12)*** | 18.74 (5.12)*** |
| Level 1 | ||||||
| Time | −1.96 (0.22)*** | 2.63 (0.50)*** | 2.09 (0.49)*** | 1.41 (0.55)* | 3.34 (0.55)*** | −1.80 (1.41) |
| Sleep problems | 0.60 (0.32) | −2.90 (0.88)** | −1.46 (0.41)*** | −0.65 (0.46) | −1.45 (0.44)*** | −5.08 (1.40)*** |
| Time×Sleep problems | 0.25 (0.18) | 1.56 (0.56)** | ||||
| Level 2 | ||||||
| Treatment group—CBT-AD | −2.73 (1.21)* | 0.06 (3.37) | 1.82 (3.47) | 0.15 (3.61) | 3.48 (3.01) | −9.81 (5.17) |
| Treatment group—ISP-AD | −0.99 (1.21) | 3.09 (3.37) | −0.89 (3.46) | 3.63 (3.61) | 1.80 (3.01) | −2.62 (5.20) |
| Between-person sleep problems | 1.51 (0.32)*** | −2.90 (0.86)** | −2.98 (0.92)** | −4.03 (0.95)*** | −2.17 (0.80)** | −2.25 (1.29) |
| Cross-level interactions | ||||||
| Sleep problems×CBT-AD | 2.93 (1.41)* | |||||
| Sleep problems×ISP-AD | 1.04 (1.42) | |||||
| Time×CBT-AD | 6.36 (1.76)*** | |||||
| Time×ISP-AD | 6.37 (1.76)*** | |||||
| Random effects | ||||||
| Intercept | 26.98 (5.20) | 285.97 (16.91) | 374.55 (19.35) | 348.44 (18.67) | 210.48 (14.51) | 603.02 (24.56) |
| Time | 2.27 (1.51) | 13.51 (3.68) | 7.00 (2.64) | 10.05 (3.17) | 18.58 (4.31) | 22.13 (4.70) |
| Residual | 39.26 (6.27) | 195.91 (14.00) | 218.38 (14.78) | 273.89 (16.55) | 235.47 (15.35) | 321.55 (17.93) |
| ICC for random effects model | ||||||
| 2* log likelihood | 6107.5 | 7571.0 | 7622.2 | 7832.5 | 7643.5 | 7675.7 |
SE: standard error; QOLI: Quality of Life Inventory; CBT-AD: cognitive behavioral therapy for adherence and depression treatment; ISP-AD: information with supportive psychotherapy for adherence and depression; ICC: intraclass correlation coefficient.
Parameter estimates and standard errors are reported for each variable. Bolded entities are predictors of interest. General, Cognitive, Pain, and Energy/Vitality are all subscales of the ACTG QOL 601-602 SF-21. Sleep problems are within-person sleep problems (sleep problems centered within each person) and between-person sleep problems are labeled in the table. Procedures for model building as described in text were followed. However, only the final model results are presented for each outcome and reported in text.
Reduced-form equation
p <0.001,
p <0.01,
p <0.05.