Proportion of the association between financial hardship and health-related quality of life explained by limiting care due to cost
| ACME | Direct effect | Total effect | Proportion mediated | Sensitivity analysis | |
|---|---|---|---|---|---|
| Mean (95% CI) | Mean (95% CI) | Mean (95% CI) | % (95% CI) | (ρ) | |
| FACT-G | −1.6 (−2.4 to −0.9) | −2.4 (−4.7–0.0) | −4.0 (−6.3 to −1.7) | 40.5 (25.5–92.7) | −0.19 |
| White | −0.8 (−1.8 to −0.1) | −3.6 (−7.4–0.2) | −4.4 (−8.1 to −0.7) | 18.4 (9.7–88.7) | −0.12 |
| African American | −2.1 (−3.4 to −1.1) | −2.1 (−5.2–1.1) | −4.2 (−7.3 to −1.1) | 50.5 (29.1–188.1) | −0.23 |
NOTE: These models control for mean-centered continuous age and for sex, race, marital status, income, education, employment status, comorbid conditions, health insurance, cancer site, stage at diagnosis, and treatments received. Because the mediation models do not allow for dummy-variable adjustment, marital status, employment status, and stage at diagnosis were treated as binary variables (married/cohabitating vs. not; employed full- or part-time vs. not, stage I vs. stages II–IV, respectively), and ordinal income and education were treated as continuous variables.