From: Blinded sample size re-estimation in a comparative diagnostic accuracy study
 |  | McCray et al. (2017) | Our approach |
---|---|---|---|
General information | Endpoint | \(\frac{\mathrm{S}{\mathrm{e}}_{\mathrm{E}}}{\mathrm{S}{\mathrm{e}}_{\mathrm{C}}}\) and \(\frac{\mathrm{S}{\mathrm{p}}_{\mathrm{E}}}{\mathrm{S}{\mathrm{p}}_{\mathrm{C}}}\) | SeE − SeC and SpE − SpC |
\({\mathbf{H}}_{{\mathbf{0}}_{\mathbf{global}}}\) | \({\mathrm{H}}_{0_{\mathrm{S}\mathrm{e}}}:\frac{\mathrm{S}{\mathrm{e}}_{\mathrm{E}}}{{\mathrm{S}\mathrm{e}}_{\mathrm{C}}}=1\cup\) \({\mathrm{H}}_{0_{\mathrm{S}\mathrm{p}}}:\frac{\mathrm{S}{\mathrm{p}}_{\mathrm{E}}}{{\mathrm{S}\mathrm{p}}_{\mathrm{C}}}=1\) | \({\mathrm{H}}_{0_{\mathrm{Se}}}:\mathrm{S}{\mathrm{e}}_{\mathrm{E}}-{\mathrm{Se}}_{\mathrm{C}}=0\cup\) \({\mathrm{H}}_{0_{\mathrm{Sp}}}:\mathrm{S}{\mathrm{p}}_{\mathrm{E}}-{\mathrm{Sp}}_{\mathrm{C}}=0\) | |
Sample size calculation | Conventional approach α per endpoint: 0.05 (two-sided) Power per endpoint: 0.8 | Optimal approach α per endpoint: 0.05 (two-sided) Overall power: 0.8 | |
Parameter of dependency between both tests | \(\mathrm{TPPR}=\frac{{n_{\mathrm{D}}}_{11}}{n_{\mathrm{D}}}\) \(\mathrm{TNNR}=\frac{{n_{\mathrm{ND}}}_{00}}{n_{\mathrm{ND}}}\) | \({\psi}_{\mathrm{D}}=\frac{{n_{\mathrm{D}}}_{10}+{n_{\mathrm{D}}}_{01}}{n_{\mathrm{D}}}\) \({\psi}_{\mathrm{ND}}=\frac{{n_{\mathrm{ND}}}_{10}+{n_{\mathrm{ND}}}_{01}}{n_{\mathrm{ND}}}\) | |
Initial sample size calculation | Size of internal pilot study | TPPRmax and TNNRmax correspond to \({\psi}_{{\mathrm{D}}_{\mathrm{min}}}\) and \({\psi}_{{\mathrm{ND}}_{\mathrm{min}}}\) | |
Parameter of dependency between both tests for initial sample size calculation | TPPRmax = SeC = 0.81 TNNRmax = SpC = 0.66 | \({\psi}_{{\mathrm{D}}_{\mathrm{min}}}=\left|\mathrm{S}{\mathrm{e}}_{\mathrm{C}}-\mathrm{S}{\mathrm{e}}_{\mathrm{E}}\right|=0.09\)  \({\psi}_{{\mathrm{ND}}_{\mathrm{min}}}=\left|\mathrm{S}{\mathrm{p}}_{\mathrm{C}}-\mathrm{S}{\mathrm{p}}_{\mathrm{E}}\right|=0.14\) | |
Initial sample size, size of internal pilot study | 186 | 133 | |
Sample size re-estimation | Estimation of nuisance parameters | \(\hat{\pi}=0.44\) \(\hat{TPPR}=0.80\mid\) \(\hat{TNNR}=0.66\) | \(\hat{\pi}=0.44\) \({\hat{\psi}}_{\mathrm{D}}=0.11\mid\) \({\hat{\psi}}_{\mathrm{ND}}=0.14\) |
Re-estimated sample size | 242 | 200 |