No abstract

Randomized controlled trials (RCTs) are the best method to determine causal effects for treatments if they are well done and well reported. Good evidence about proposed treatments for obesity is needed, and Hsieh et al. [1] are to be commended for putting moxibustion to the test. However, careful evaluation of the paper, similar to a prior review of another paper on moxibustion [2], reveals inconsistencies and apparent reporting errors, which raise doubts about conclusions from the study.The primary concern relates to the randomization process. Table 1 in Hsieh et al., which describes baseline anthropometrics of both groups in their study, showed that baseline body weight (BW) differed with a high degree of statistical significance (p < 0.001) based on the comparison of groups using a t test [1], which, by definition, is very unlikely to occur under randomization [3][4][5][6]. We also note that the baseline variances for BW were significantly heterogeneous (Bartlett's test; χ 2 = 6.86, df = 1, p = 0.0088 ), which is again, by definition, highly unlikely under randomization. These baseline distribution differences by group raise questions as to whether the randomization was properly performed. Results might be influenced by any problem in the randomization process, which could bias the estimated treatment effects [6]. The details of the randomization process were not clear, and we found no Abstract Randomized Controlled Trials (RCTs) are the best method to determine causal effects for treatments if they are well done and well reported. Good evidence about proposed treatments for obesity is needed, and Hsieh et al. (Biomed Eng Online 17:149, 2018) are to be commended for putting moxibustion to the test. However, careful evaluation of the paper reveals inconsistencies and apparent reporting errors, which raise doubts about conclusions from the study.

No abstract

No abstract

Randomisation can do many thingsbut it cannot "fail" Although randomisation has long been seen as crucial to reaching reliable insights from data, it is still falling victim to some peculiar -and troublesome -misconceptions.