Table 4.

Hypothetical scenarios described in 57 methods publications addressing allocation bias directionality and mechanisms

Favouring treatment:	Favouring control:	Direction unpredictable:
“An investigator for a pharmeceutical company, very anxious to see the company’s latest pharmaceutical product succeed, guesses the randomization sequence and randomizes patients he or she deems more likely to respond positively to the new therapy when he believes the new therapy to be next in the sequence” [21]a “Or older patients might receive the traditional therapy and youngsters the new one” [41] “Because of the vested interest the sponsor may have in the outcome of the trial (Hogel and Gaus, 1999), there may be a temptation to attempt to bias trials towards more favorable outcomes.“ [27] “…when clinicians know ahead of time which treatment their next eligible and consenting patient will receive, they may (consciously or unconsciously) enter patients with lower risk and/or higher responsiveness into the experimental treatment group. As a result, the trial can become biased in favor of experimental therapy from the start.” [29]	“The clinician might, for example, allocate the patients with the worst prognosis to the, in his/her opinion, ‘promising’ new therapy and the better ones to the older treatment, no doubt with the best possible intention in respect of his/her patients” [41] “On the other hand, in some unblinded trials, one of the treatments may be systematically assigned to the sickest patients. Randomization appears to have been systematically slanted to make these patients benefit from the test medication which would appear to be more effective.” [31]	“A sympathetic nurse coordinator tries to assign a favorite patient to the new therapy rather than placebo” [21] “If the referring health care provider knows the next subject will be allocated to Slimmenow b , he/she may be inclined to try to help a certain patient he/she thinks may benefit more. Or perhaps knowing the next subject is to be allocated to placebo, he/she refers someone who really does not need to lose much weight.” [61] “They perhaps “know” the more effective treatment, so they may want certain patients to benefit or may want the results of a study to reveal what they believe to be valid.” [20] “For instance, if the referring clinician thinks that treatment A is less effective than treatment B, and he/she knows that the next subjects will be allocated to treatment A and B, respectively, he/she may be inclined to chose a patient with mild symptoms for treatment A, and a patient with more severe symptoms for treatment B.” [43]

^aReferences indicate from which publication scenario is taken; ^bHypothetical experimental treatment