Reliability of the NICMAN Scale : an instrument to assess the quality of acupuncture administered in clinical trials

Background. The aim of this study was to examine the reliability of a scale to assess the methodological quality of acupuncture administered in clinical research. Methods. We invited 36 acupuncture researchers and postgraduate students to participate in the study. Firstly, participants rated two articles using the scale. Following this initial stage, modifications were made to scale items and the exercise was repeated. Interrater reliability was assessed for individual items using the Fleiss kappa statistic, whilst the overall scale used the intraclass correlation coefficient statistic. A threshold agreement of ≥0.61 was acceptable. Results. We received 26 responses and a 72% response rate. The first phase of testing found moderate reliability with intraclass correlation coefficients of 0.46 and 0.55 for the articles. The interrater reliability of the scales varied between and within the researchers (0.35, 0.60) and was more consistent with the postgraduate students (0.54, 0.54). Five items on the scale scored below the threshold and were revised for further testing. In this phase the intraclass correlation coefficient demonstrated variability between articles but improved to achieve reliability above the agreed threshold. Conclusion. This study provides evidence of the reliability of the NICMAN scale although improvements to a small number of items remain.