Background: Many randomized controlled trials (RCTs) and reviews concluded that acupuncture is not an effective treatment for low back pain (LBP) and there is no difference between real acupuncture and sham acupuncture in the treatment of LBP.
Objective: This study aims to evaluate the most recently published RCTs and reviews from the clinical protocol, which is used by professional acupuncturists.
Data sources and synthesis: English-language studies were identified through searches of The Cochrane Library, PubMed/MEDLINE and EMBASE, limited to those published between January 2007 and January 2010. Eighteen trials studying conditions related to LBP conditions treated by acupuncture or acupuncture as one of the co-interventions were eligible. STUDY APPRAISAL METHODS: A critical clinical qualitative evaluation is the core methodology of this study. The study design employs daily used clinical skills and knowledge with supplementary statistical concepts to evaluate the quality and reliability of the selected RCTs and reviews. Five evaluation criteria were established for the assessment of the selected RCTs and reviews. By implementing the evaluation criteria, which are similar to the inclusion and exclusion criteria used in RCTs, the weaknesses, limitations or errors of RCTs and reviews can be identified, despite the strength of internal validity produced by statistical calculations.
Results: Various terms are used when describing LBP conditions. However, 16 RCTs or 88% of the trials did not establish a medical diagnosis. No trials had objective measurements as a clinical standard for assessing outcomes. Furthermore, significant variation existed due to the numerous treatment methods or therapists used in the RCTs included in the study. Various co-interventions were used in 10 or 55% of the trials, among them, 6 or 33% of the trials used non-steroidal anti-inflammatory drugs (NSAID) or analgesics. Due to the lack of diagnostic status, the accuracy of the external validity was put into question. No logistic regression models were used in any of the selected RCTs and reviews to resolve the degree of heterogeneity.
Conclusion: All selected RCTs either partially met the evaluation criteria or did not fulfill the evaluation criteria while being compared. The major problem existing in all RCTs was the lack of accurate medical diagnosis and a lack of objective measurements to judge the clinical outcomes, which in turn, created incorrect eligibility criteria, improper matches and inaccuracies in data recording before entry into statistical calculations. As a consequence of the subjective nature of measurements, conclusions of RCTs and RCT-based reviews were biased due to overgeneralized or cross-generalized estimations, which infer that alternative explanations cannot be excluded. In addition, the range of variables created in the treatment procedure was difficult to control or estimate, in turn, threatening the reliability of RCTs' estimations. For these reasons, creating appropriate diagnostic criteria before randomization and constructing a related objective outcome measurement, which are more relevant to clinical practice, should be considered in future RCTs and systematic review studies.