Table 3.
Univariate and multivariable logistic regression models of receipt of cancer-related follow-up care (within prior 2 years) among childhood cancer survivors (diagnosed in 1996-2010; Los Angeles County)a
| Characteristic | Bivariate analyses |
Multivariable model |
||
|---|---|---|---|---|
| OR (95% CI) | P | Adjusted OR (95% CI) | P | |
| Years since diagnosis | 0.88 (0.85 to 0.90) | <.001 | 0.88 (0.84 to 0.92) | <.001 |
| Age at survey completion, y | ||||
| 18-20 | 1.00 (referent) | — | 1.00 (referent) | — |
| 21-25 | 1.69 (1.31 to 2.19) | <.001 | 0.65 (0.50 to 0.85) | .002 |
| 26-30 | 0.70 (0.54 to 0.92 | .01 | 0.32 (0.22 to 0.48) | <.001 |
| 31-39 | 0.45 (0.33 to 0.61) | <.001 | 0.35 (0.24 to 0.50) | <.001 |
| Female (vs Male) | 1.39 (1.09 to 1.77) | .01 | 1.16 (0.86 to 1.58) | .34 |
| Race and ethnicity (relative to non-Hispanic White) | ||||
| Non-Hispanic White | 1.00 (referent) | — | 1.00 (referent) | — |
| Hispanic | 0.81 (0.59 to 1.11) | .18 | 0.69 (0.51 to 0.95) | .02 |
| Asian | 0.75 (0.43 to 1.29) | .29 | 0.83 (0.52 to 1.31) | .42 |
| Other | 0.89 (0.61 to 1.31) | .54 | 0.69 (0.48 to 0.99) | .04 |
| Socioeconomic status (relative to lowest group) | ||||
| Lowest | 1.00 (referent) | — | 1.00 (referent) | — |
| Low | 0.89 (0.66 to 1.19) | .42 | 0.92 (0.66 to 1.26) | .59 |
| Medium | 1.06 (0.75 to 1.49) | .74 | 1.12 (0.76 to 1.65) | .56 |
| High | 1.59 (1.13 to 2.24) | .01 | 1.01 (0.67 to 1.52) | .97 |
| Highest | 1.09 (0.78 to 1.52) | .60 | 0.93 (0.62 to 1.39) | .73 |
| Health insurance (any vs bone) | 3.05 (1.96 to 4.75) | <.001 | 2.06 (1.28 to 3.32) | .003 |
| High levels of depressive symptoms | 0.93 (0.71 to 1.21) | .57 | (not included) | — |
| No. of late effects (relative to none) | ||||
| 0 | 1.00 (referent) | — | 1.00 (referent) | — |
| 1 | 1.23 (0.89 to 1.69) | .21 | 1.41 (1.08 to 1.83) | .01 |
| ≥2 | 1.51 (1.10 to 2.07) | .01 | 1.54 (1.23 to 1.92) | <.001 |
| Treatment intensity | 1.28 (1.10 to 1.50) | .002 | 1.18 (0.92 to 1.52) | .20 |
| Received written cancer treatment summary | 2.72 (2.10 to 3.52) | <.001 | 1.47 (1.16 to 1.87) | .002 |
| Has doctor for regular (noncancer) health checkups | 2.03 (1.54 to 2.67) | <.001 | 1.47 (1.13 to 1.92) | .005 |
| Discussed cancer-related follow-up care needs with a doctor (in the last 2 years) | 3.28 (2.54 to 4.24) | <.001 | 1.95 (1.49 to 2.55) | <.001 |
| Knowledge of need of lifelong follow-up care | 3.53 (2.71 to 4.60) | <.001 | 3.57 (2.90 to 4.39) | <.001 |
| Health-care self-efficacy | 1.35 (1.23 to 1.48) | <.001 | 1.23 (1.09 to 1.39) | <.001 |
| Family influence health-care decisions | 1.36 (0.96 to 1.93) | .08 | 0.90 (0.59 to 1.38) | .63 |
a
All logistic regression models adjust for clustering at diagnosing hospital. All variables are included, and mutually adjusted for, in the multivariable model except for depressive symptoms (which was not statistically significant in the bivariate analyses). P values are 2-sided. CI = confidence interval; OR = odds ratio; — indicates no P value.