On 2016 May 13, Richard E Harris commented:

The sample size in this study was based on a power calculation using results from the Deluze et al. 1992 study as an estimate of effect size. This was the only publication using acupuncture in fibromyalgia at that time. Our power analysis indicated that we would need 30 participants (per group) to detect a 30% difference between groups at a significance level of 0.05. We enrolled from 27 to 30 per arm suggesting we had significant power to test our hypotheses.

Regarding the drop out rate of 33%, we performed an Intention To Treat analysis which controls for some effects of drop outs. Also there was no significant difference in drop outs (or treatments received) between study arms. These make it unlikely that our results were influenced by dropouts.

On 2016 May 04, Arthur Yin Fan commented:

This study has a significant flaw--the sample size is over small, and could not make sure the acupuncture's efficacy--the design of four groups, should have 800 patients.

The second, the drop out rate 33%, makes the result is not trust-able.

On 2016 May 04, Arthur Yin Fan commented:

This study has a significant flaw--the sample size is over small, and could not make sure the acupuncture's efficacy--the design of four groups, should have 800 patients.

The second, the drop out rate 33%, makes the result is not trust-able.

On 2016 May 13, Richard E Harris commented:

The sample size in this study was based on a power calculation using results from the Deluze et al. 1992 study as an estimate of effect size. This was the only publication using acupuncture in fibromyalgia at that time. Our power analysis indicated that we would need 30 participants (per group) to detect a 30% difference between groups at a significance level of 0.05. We enrolled from 27 to 30 per arm suggesting we had significant power to test our hypotheses.

Regarding the drop out rate of 33%, we performed an Intention To Treat analysis which controls for some effects of drop outs. Also there was no significant difference in drop outs (or treatments received) between study arms. These make it unlikely that our results were influenced by dropouts.