通过盲法评估提升开放标签临床试验:一项糖尿病足感染研究的探索
本研究以一项糖尿病足感染随机对照试验为案例,从16例患者中随机选取样本,由非盲研究者和盲法中心阅片研究者共同进行感染分级评估,一致性检验显示两者结果不一致(Kappa=0.248,p=0.384)。作者据此提出在开放标签试验中采用各中心独立盲法评估员及标准化操作规程等策略。
要点
- 非盲站点研究者评估可能夸大试验组疗效,深部伤口经中心阅片照片观察存在困难。
- 非盲与盲法中心阅片评估一致性检验结果不一致(Kappa=0.248,p=0.384)。
- 作者建议在开放标签试验中采用各中心独立盲法评估员并配套标准化操作规程。
核心结论
- 非盲站点研究者的评估可能潜在夸大试验组疗效,深部伤口经中心阅片照片观察存在困难。
- 开放标签临床试验中客观评估至关重要,尤其当伤口观察作为主要疗效指标时。
- 作者建议采用各中心独立盲法评估员并配套完整的盲法评估标准化操作规程,以平衡主观偏倚、提升开放标签试验结果的可信度与一致性。
研究设计与样本
- 本研究为探索性研究,以一项研究糖尿病足感染的随机对照试验为代表性案例。
- 该试验以美国感染病学会(IDSA)和国际糖尿病足工作组(IWGDF)指南的糖尿病足感染分级作为主要疗效指标。
- 从该随机对照试验中随机选取16例患者,由非盲研究者和盲法中心阅片研究者共同评估糖尿病足感染分级,并进行一致性检验。
- 研究结束时还收集了参与近期盲法评估试验设计与实施人员的其他观点。
主要发现
- 5例受试者因照片质量或缺乏治疗后访视被排除,最终对11例受试者(试验组6例、对照组5例)的治疗后IDSA/IWGDF分级结果进行比较。
- 非盲评估与中心阅片盲法评估的一致性检验显示结果不一致(Kappa=0.248,p=0.384)。
- 试验组中3例在非盲评估中判为1级、在中心阅片盲法评估中判为2级。
- 对照组中3例在非盲评估中判为2级、在中心阅片盲法评估中判为1级。
- 上述两类情况在22次治疗后判定中合计为27%(6/22)。
临床怎么用
- 作者建议在开放标签临床试验中采用各中心独立盲法评估员,并配套完整的盲法评估标准化操作规程。
- 作者提出的策略涵盖人员配置、培训、受试者管理、试验药物给药、疗效指标收集和安全事件管理等方面。
局限性
- 5例受试者因照片质量或缺乏治疗后访视被排除,最终仅对11例受试者的治疗后分级结果进行比较。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Blinding drugs through simulation techniques is an important means to control the subjective bias of investigators and subjects. However, clinical trials face significant challenges in the placebo production of drugs, and many trials cannot be double-blinded. OBJECTIVE: This study was conducted to ascertain the consistency between non-blind and blind evaluation results in clinical trials and to pioneer strategies to control information bias, particularly in trials where double-blinding is not feasible. METHODS: In this investigation, a randomized controlled trial (RCT) studying diabetic foot infections (DFIs) was utilized as a representative case. In this trial, the grading of DFIs, as per guidelines by the Infectious Disease Society of America (IDSA) and International Working Group on Diabetic Foot (IWGDF), was used as the primary efficacy indicator. A sample of sixteen patients was randomly chosen from the RCT, and DFI grading was assessed jointly by both non-blinded investigators and blinded center-reading investigators. A consistency test was then deployed to compare the evaluation results, forming the basis for our proposed strategies for effective blinded evaluation. In addition, other perspectives were collected at the end of this study, including with those involved in designing and conducting the recent blinded evaluation trial. RESULTS: Five subjects were excluded due to the quality of photos or the lack of post-treatment visits. The post-treatment IDSA/IWGDF grading results were compared in 11 subjects (experimental group=6, control group=5), and the consistency test showed inconsistent results between the non-blinded and center reading blinded evaluations (Kappa=0.248, p=0.384). In the experimental group, three cases were judged as grade 1 in the non-blinded evaluation and grade 2 in the central reading blinded evaluation; in the control group, three cases were judged as grade 2 in the non-blinded evaluation and grade 1 in the central reading blinded evaluation. The sum of these two cases in 22 post-treatment determinations was 27% (6/22). Furthermore, researchers propose several strategies for implementing blinded evaluations in clinical trials after this trial, which encompass aspects such as staff allocation, training, participant management, trial drug administration, efficacy indicator collection, and safety event management. CONCLUSIONS: The study highlighted that evaluations from non-blinded site investigators may potentially exaggerate the efficacy of the experimental group and that deep wounds can present challenges for observation via center-reading photos. These findings underline the vital necessity for objective assessment in open clinical trials, especially those where wound observation serves as the primary efficacy indicator. The study suggests the adoption of independent blinded investigators at each site, complemented by a comprehensive set of standard operating procedures for blinding evaluation. These measures could serve as an effective counterbalance to subjective bias, thereby augmenting the credibility and consistency of results in open clinical trials. The implications of these findings and recommendations could be of great significance for the design and execution of future open clinical trials, potentially bolstering the quality of clinical research in this area. TRIAL REGISTRATION: ChiCTR2000041443. Registered on December 2020.
引用
- Li Q, Xiao M, Liu X, et al. Enhancing open clinical trials through blinded evaluations: an exploration with diabetic foot infections. Trials. 2023. PMID: 37946260. doi:10.1186/s13063-023-07652-y
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。