舌象仪临床应用研究的方法学及报告质量评价

目的 评价舌象仪临床应用研究的随机对照试验（RCT）、病例系列研究的方法学及报告质量。方法 检索PubMed、中国期刊全文数据库（CNKI）、维普数据库（VIP）和万方数据库（Wanfang database）,从建库截至2018年12月开始检索,收集应用舌象仪观察临床疾病的RCT、病例系列研究和病例对照研究,分别参考Cochrane手册5.1.0版的偏倚风险评估工具和CONSORT 2010版的声明,对纳入的RCT研究进行方法学和报告质量进行评价;根据NICE和STROBE声明第四版进行方法学的报告质量评价研究。结果 初检出相关文献567篇,最终纳入59篇文献,包括2个RCT,57个病例系列研究。50%的RCT应用Cochrane偏倚风险工具评价后显示采用的随机分配方法比较合理,没有采用分配隐藏和盲法;均采用了意向性分析,遗憾的是并没有报告失访的情况;研究也缺乏选择性报告。NICE病例系列研究质量评价结果 病例系列研究中仅14.04%符合多中心研究;59.64%的研究纳入排除标准均不明确;数据未达到预期设定目标,未进行分层分析结局。CONSORT声明2010版评价结果显示：纳入的文献均不能从题目辨别出是否是随机临床试验,多项条目报告率为0;应用STROBE声明评价后结果显示：纳入的病例系列研究均未在题目中明确标明是病例系列研究,多项条目报告率为0。结论 舌象仪临床应用的RCT和病例系列研究报告质量不高,建议在设计和报告临床病例对照和随机对照试验时分别参考CONSORT声明、STROBE声明,纠正目前研究存在的方法学问题,进一步提高临床研究质量,增强舌象仪临床应用于疾病效用证据的强度,推进中医的客观化和现代化。

Methodology and Report Quality Evaluation for Clinical Application of Tongue Imager

Objective To evaluate the randomized controlled trial (RCT) of clinical application of tongue imager, the methodology of case series study and the quality of report. Methods PubMed, CNKI, VIP, and WanFang databases were searched. The database was searched from databse establishment to December 2018, and RCT, case series, and case-control studies were performed using tongue images to observe clinical disease. Referring to Cochrane Manual 5.1.0 Version of the Risk Assessment Tool and the CONSORT 2010 version of the statement, the methodological and reporting quality of the included RCT study was evaluated; and the methodological quality evaluation study of the methodology was conducted according to the fourth edition of the NICE and STROBE Statement. Results 567 related literatures were initially detected and 59 were included, including 2 RCTs and 57 case series studies. 50% of the RCTs used Cochrane Bias Risk Tool to evaluate, which showed that the random allocation method was reasonable and did not use allocation concealment and blind method. Intentional Therapy (ITT) analysis was used, unfortunately, no missing interviews were reported, and there was no selective report in the study. The quality evaluation results of NICE case series studies: only 14.04% of the case series studies met the multi-center study; 59.64% of the studies included in the exclusion criteria were not clear; the data did not meet the expected set goals and did not carry out stratified analysis. The evaluation results of the 2010 edition of CONSORT statement showed that the included literature could not distinguish whether it was a randomized clinical trial or not, and the reporting rate of multiple items was 0. The results of evaluation with STROBE statement showed that all the included case series studies were not clearly identified as case series studies in the title, and the reporting rate of multiple items was 0. Conclusion The RCT and case series of clinical reports on tongue imaging are of low quality. It is recommended to refer to the CONSORT statement and the STROBE statement when designing and reporting clinical case-control and randomized controlled trials to correct the current methodological problems and further improve clinical research, enhance the intensity of clinical application of tongue imager in clinical efficacy, and promote the objectification and modernization of Chinese medicine.