Table 3.
Observed Power and Assessment of Over-sampling (n = 91).
| {Pobs > (Pexp + 2.5%)} (n = 69) |
Well-powered (n = 3) |
Under-powered {Pobs < (Pexp—2.5%)} (n = 19) |
|
|---|---|---|---|
| HRobs better magnitude of effect than HRexp (n = 54) | 52 | 1 | 1 |
| HRobs similar magnitude of effect as HRexp (n = 15) | 12 | 0 | 3 |
| HRobs worse magnitude of effect than HRexp (n = 22) | 5 | 2 | 15 |
Over-sampled endpoints are highlighted in bold. Study endpoints were considered to be over-sampled if the endpoint had Pobs larger that Pexp and HRobs similar or worse magnitude of effect than HRexp OR if the endpoint was well-powered and HRobs worse magnitude of effect than HRexp. Nineteen endpoints were considered oversampled (19/91, 20%).