Table 2. The common flaws in the experimental design of the acupunctural studies. Modified from the chapter of (Asakawa and Xia, 2012).

Weaknesses in the experimental design	Comments	Solutions
The undefined sample size	Sample size is the most crucial aspect of formulating an experimental design.	Estimated according to the statistical principles (Asakawa and Xia, 2012). The method for establishing the sample size should be clearly put down.
Lack of “blinding”	Basically, double-blinding is needed; however, single blinding with objective indices is acceptable.	Methods of Allocation concealment should be included in the paper.
Insufficient randomization or pseudorandom	Only a small number of these studies used a robust randomization procedure while many of them used insufficient randomization, even pseudo-random in some cases.	Simple randomization. Block randomization. Stratified randomization. Covariate adaptive randomization.
Inappropriate control group setting	There are several flaws in using the popular protocol like C + A vs. C (C = classical therapy; A = acupuncture).	A protocol of A + C vs C + P, (P = placebo acupuncture) is recommended. The methodology of setting placebo acupuncture.
Lack of objective evaluation standard	Lack of an objective standard for acupoint-selection. No standardized method of performing manual acupuncture. The evaluation criteria are also subjective.	The ideas of evidence-based medicine should be set up. Change the traditional experience-based ideas.
Weakness in statistical analysis	Fail to describe the data analysis methods. List only raw data in tables, do not use any figure with statistical information.	Consult a statistical expert. Using statistical figures.
Neglect in recording adverse events and withdrawals	Without adequate reporting of adverse events and withdrawals, the results from a clinical trial can not be accepted.	Adverse events and withdrawals should be recorded clearly.