Table 4 Simulated scenarios in the unpaired and paired study design testing for superiority in both endpoints. The proportion of discordant test results is only relevant in the paired design
From: Blinded sample size re-estimation in a comparative diagnostic accuracy study
| Â | 10,000 simulation runs per scenario | |
|---|---|---|
Nominal significance level α per endpoint | 0.05 (two-sided) | |
Nominal overall target power | 0.8 | |
| Â | Initial scenario | Variation of initial scenario |
Sensitivity comparator test SeC | 0.8 | 0.6, 0.7 |
Specificity comparator test SpC | 0.7 | 0.6, 0.8 |
True prevalence π true | 0.2 | 0.4, 0.6, 0.8 |
Assumed prevalence π ass. | πtrue +  0.1 | πtrue - 0.1 πtrue +  0.2 πtrue +  0.3 |
True discordant results diseased population \({\psi}_{{\mathrm{D}}_{\mathrm{true}}}\) | 0.11 (0.15, if: SeE − SeC = 0.15) | 0.18, 0.26 |
Assumed discordant results diseased population \({\psi}_{{\mathrm{D}}_{\mathrm{ass.}}}\) | 0.18 | Â |
True discordant results non-diseased population \({\psi}_{{\mathrm{ND}}_{\mathrm{true}}}\) | 0.14 (0.15, if: SpE − SpC = 0.15) | 0.24, 0.38 |
Assumed discordant results non-diseased population \({\psi}_{{\mathrm{ND}}_{\mathrm{ass.}}}\) | 0.24 | Â |
Sensitivity experimental test SeE | \(\hat{=}\mathrm{S}{\mathrm{e}}_{\mathrm{C}}\) | Â |
Specificity experimental test SpE | \(\hat{=}\mathrm{S}{\mathrm{p}}_{\mathrm{C}}\) | Â |
Sensitivity experimental test SeE | SeC +  0.1 | SeC +  0.05 SeC +  0.15 |
Specificity experimental test SpE | SpC +  0.1 | SpC +  0.05 SpC +  0.15 |