In investigating all the back articles of the Japan Society of Acupuncture Journal (vol. 32 no. 1 1982) until the present (vol. 41 no. 2 1991) it was found that out of 169 articles using statistical analysis 164 of them (97%) contained some kind of problem. Of these, 99 out of 105 (94%) stimulus experimentations by animal and human, and 62 out 64 (97%) clinical studys were found to contain misuses of statistics approximately the same percentage for each group.
A high incidence of problem was of multiplicity, with 73 out of 164 problem articles (45%) containing misuse of statistics.
Beside this, 19% (31/164) of the articles failed to list the methods of analysis or test of significance used, 13% (22/164) published their conclusions based only on comparisons within a group, 14% (23/164) published only the rate of effectiveness, without reporting the rate of uncertain results, for example, confidence interval, and 12% (19/164) failed to give assurances against randomness. Overall there are 18 checkpoints, but there is a less than 10% rate of statistical misuse in each of the others.
Because each report has multiple problem the total percentage is greater than 100%. Even excluding the 19% of the articles failing to list methods of analysis or test of significance and the 14% listing only the rates of effectiveness, neither of which ultimately affect the article results, there ara still statistical misuse in 62% of the articles using statistical analysis.
Looking at the authors it can be seen that 11 writers account for 30% (49 articles) of the problem. In short, the same people are repeating the same mistakes in from 3 to 7 articles each.
Of these, 91% contain problems of multiplicity, and 55% contain comparisons conducted within a group instead of between groups as required. These are their most glaring faults.

In most cases, when papers reported that no significant difference was found in randomized clinical trials this is a result of the sample size being too small. For this reason it is important to design the sample size before any trials or tests are carried out. To this end, we would like to introduce some reference material concerning determination of sample sizes for clinical trials and basic research, as well as some software for determining sample size designed for personal computers.

We looked into the statistical methods used in “The Journal of the Japan Society of Acupuncture” over the last ten years. Our results showed that the t test was most common, used 72.5% (127) of the time; the χ² test was next, used 25.7% (45) of the time; and the third most popular was the analysis of variance, at 5.7%. None of the remaining 17 methods were used more than 5% of the time. In all the instances where the t test was used, this was only clearly indicated in 36.5% of the cases. Similarly, where the χ² was used, it was only clearly indicated in 13.7% of the cases. In a similar percentage of the cases where two statistical analysis was used, the name of the method used was not listed in the article. Up until now, in 14% of the instances where the t test was referred to, it was written as the T test. This was surprisingly often. And in only 1.1% of the cases are the names of the software used given. This is surprisingly small. Considering these omissions, together with the fact that these methods have been misused 97% of the time, we can only conclude that what is needed is more education in statistical methods, as well as stricter editorial policies governing the submission of papers coupled with a system of peer review.

There is a serial correlation to sequential measurement data, where data is taken over a period of time. Analysis of each of these points in times as an independent unit with its own value will lead to inacurate results. The sequential measurement model is used quite frequency in acupuncture reaearch. Because misuse of statistics occurs quite commonly in acupuncture reaearch, almost all cases we studied of analysis of this sequential measurement model were in error. Because of this, we undertook to review literature on sequential measurement data and computer software packages (statistical software), in order to report on proper methods of analysis to the society members.

The relative importance of soft, or qualitative, data over hard, or metric, data is now beginning to be recognized in clinical trials. This tendency can be seen in the recent importance attached to evaluating quality of life. It is said that acupuncture and moxibustion are used to cure the patient, not the illness. This makes it likely that the simple and convenient Visual Analogue Scale (VAS) will come into wide use henceforth in evaluating research in acupuncture and moxibustion. Therefore, we would like to take this opportunity to discuss how to make a VAS, how to analyze it, points to look for in implementing it, and the relation between VAS and similar methods.

Zen Nippon Shinkyu Gakkai Zasshi(Journal of the Japan Society of Acupuncture and Moxibustion:JJSAM).
2010; 60(2): 244-251. Received 19 Mar, 2009 Accepted 27 Mar, 2010

Treatment by acupuncture and moxibustion and by orthopedic medicine have both been found to be effective against lower back pain. The time required for the treatment to become effective has been found to correspond with the Weibull distribution. Using the parameters arrived at by the distributon a sequential comparison of both treatments was conducted using the Monte Carlo simulation, using randomized numbers from the distribution. Although there are problems involving the estimated value of the distribution based on complete data, and the difference in quality of the subject groups, it has been found in 100 trials (inolving 100 pairs of subjects) that acupuncture and moxibustion tend to take effect significantly faster (p=0.07) than orthopedic medicine. I present these results as a hypothesis to be used in randomized clinical trials. Furthermore, I propose a new method of clinical evaluation, involing simulation derived from the investigation of models, in cases where it is difficult to form control groups for comparison.

Introduction
We tried to prove the effectiveness of acupuncture against unidentified syndrome.
Method
Subjects were women complaining of unidentified syndrome and divided into the control group and test group. Galenicals extact was given to all of them and acupuncture was administered to the test group. The examination was a sequential medical trial in a restricted design (2α=0.05, 1-β=0.95, θ₁=0.90 and N=19). Acupuncture was applied twice a week for two weeks. We evaluated their subjective symptoms, difficulties in daily life and so on with MPI-test and MV-test.
Conclusion
Among the two groups, patients were randomly paired off into ten couples. Thus, a subjective comparison was possible. Global improvement rating by patients was effective for the symptoms of 60% of the test group, and 10% of the control group (p<0.05). As to the difficulties in daily life, it was effective for 70% of the test group, whereas only 20% of the control group. About subjective symptoms, there was a significant difference in the stiffness in shoulder (p<0.10). On MV-test, the test group of five pairs showed an increase in theta waves (p<0.05) and a decrease in beta waves (p<0.10).

To identify indications for acupuncture treatment, we examined certain evidence-based medicine (EBM) practices by acupuncture/moxibustion clinicians who are novices in this field and identified problems and solutions from the perspective of an acupuncture/moxibustion clinician.
We collected cases of lumbago, shoulder disorders and strokes from the literature, and critically assessed these reports. We also investigated Cochrane Library's acupuncture/moxibustion reviews and evaluated the quality of domestic studies using randomized controlled trials and controlled clinical trials (RCT/CCT). Simultaneously, one of the authors evaluated her own clinical reports in light of EBM.
As a result, acupuncture/moxibustion treatments showed promise for treating lumbago and shoulder disorders, but the higher the quality of studies, the lower the effect of treatment after stroke. According to research by the Cochrane Library (as of issue 1, 2003), the effects of acupuncture/moxibustion was measured only for cases of idiopathic headache but those for other disorders were measured more severely. In Japan, RCT started as early as the 1960 s, but the studies were sluggish and high quality studies were not produced until the late 1990 s.
We presented problems of RCT assessment score, the gap between RCT designs and actual clinical scenes, and the difficulty of masking at the bed-side. To contribute to the accumulation of more evidence-based data, it would be desirable for clinicians to acquire EBM methods, consider clinical problems and collaborate positively with investigators in the field.