| Pitfalls
to Randomization |
|
|
|
|
|
|
|
|
|
|
|
| • | Timing
(before vs. after patient enters trial) |
|||||||||
| • | Keeping
track of drop-outs (non-random?) |
|||||||||
| – | “intent
to treat” analysis |
|||||||||
| – | “last
observation carried forward” |
|||||||||
| • | Blinding |
|||||||||
| • | Randomization
need not generate “median |
|||||||||
| person” |
||||||||||