Table 4. Sensitivity analysis of the effects of unmeasured confounder on HRs of OS and PFS.
| Prevalence of UC | UC HR | OS | PFS | |||
|---|---|---|---|---|---|---|
| DP group | PF group | HR | 95% CI | HR | 95% CI | |
| 0.4 | 0.2 | 1.3 | 0.678 | 0.490-0.938 | 0.701 | 0.501-0980 |
| 0.4 | 0.2 | 1.5 | 0.700 | 0.506-0.868 | 0.724 | 0.518-0.988 |
| 0.6 | 0.2 | 1.3 | 0.714 | 0.517-0.988 | 0.739* | 0.528-1.033 |
| 0.6 | 0.2 | 1.5 | 0.758 | 0.548-0.949 | 0.784* | 0.561-1.096 |
| 0.8 | 0.2 | 1.3 | 0.751* | 0.543-1.038 | 0.776* | 0.555-1.085 |
| 0.8 | 0.2 | 1.5 | 0.817* | 0.591-1.129 | 0.844* | 0.604-1.181 |
Abbreviations: UC: unmeasured confounder; HR: hazard ratio; OS, overall survival; PFS: progression-free survival.
*
Scenarios in which the effects of unmeasured confounder would render the association of treatment group with survival no longer statistically significant.