Complementarity comes of age.

The only inherent and inevitable difference between randomized controlled trials and nonrandomized studies (a term that is preferable to "outcomes research") is in the allocation of patients to different interventions. Although simple interventions, such as drugs, can and should always be evaluated using a randomized design, it is often not possible or appropriate to subject more complex interventions or policies to that approach. Rather than viewing randomized and nonrandomized designs as in competition, they should be seen as complementary. One of the major advantages of nonrandomized studies is that they can make use of existing large, high quality databases that have been assembled for other reasons and therefore provide excellent value-for-money.