For a confidence level \(c\),
\(\alpha = 1-c\), \(z = \Phi^{-1}(1-\alpha/2)\), and
$$y = \log(\mathrm{HR}),\qquad
\mathrm{SE} = \{\log(U)-\log(L)\}/(2z).$$
Usage
log_hr_from_ci(hr, ci_lower, ci_upper, ci_level)
Arguments
- hr
Hazard ratio.
- ci_lower, ci_upper
Confidence limits on the HR scale.
- ci_level
Confidence level in \((0,1)\), recycled to hr.
Value
A data frame with yi, sei, vi, and z.
Details
Do not substitute \(z=1.96\) when \(c \ne 0.95\). RESPIRE 1
subgroup intervals use \(c=0.975\); RESPIRE 2 uses \(c=0.951\).
Examples
log_hr_from_ci(hr = 0.61, ci_lower = 0.32, ci_upper = 1.16, ci_level = 0.975)
#> yi sei vi z
#> 1 -0.4942963 0.2872876 0.08253415 2.241403
