Randomization in clinical trials: can we eliminate bias?

Randomization plays a fundamental role in clinical trials. While many m­odern­ clinical­ trials­ employ­ restricted,­ stratified­or­ covariate-adaptive­ randomization designs that pursue balance in treatment assignments and­ balance­ across­ important­ covariates,­ some­ clinical­ trials­ call­ for­ response-adaptive­ or­ covariate-adjusted­ response-adaptive­ (CARA)­ randomization­ designs­ to­ address­ multiple­ experimental­ objectives­ primarily­related­to­statistical­efficiency­and­ethical­considerations.­In­this­ paper,­we­elicit­key­principles­of­the­well-conducted­randomized­clinical­ trial and explore the role of randomization and other important design tools­ in­ achieving­ valid­ and­ credible­ results.­We­ give­ special­ attention­ to­ response-adaptive­ and­ CARA­ randomization­ designs,­ which­ have­ a­firm­ theoretical­ basis,­ but­ are­more­ complex­ and­more­ vulnerable­ to­ operational biases than traditional randomization designs. We conclude that­ modern­ advances­ in­ information­ technology,­ rigorous­ planning,­ and­ adherence­ to­ the­ key­principles­ of­ the­well-conducted­ clinical­ trial­ should­enable­successful­implementation­of­response-adaptive­and­CARA­ randomization designs in the near future.