|
Levels of evidence for treatment effect |
|
| Level I |
Randomized trials with low false positive (α) and low false |
| negative (β) error (ie high power) |
|
| Level II |
Randomized trials with high α error or low power |
| Level III |
Non-randomized concurrent cohort studies |
| Level IV |
Non-randomized historic cohort studies |
| Level V |
Case series |
|
Adapted from Cook et al [2]. | |
Kellum Critical Care 1997 1:53 doi:10.1186/cc103 |
|