知镜 Cognispect · 医学文献与研究工作台

Identifying and Addressing Problematic Studies in Allergy and Asthma Research: A Commentary.

Chu AWL, Guyatt GH, Busse JW, Brignardello-Petersen R, Chu X, Rayner DG, Oykhman P, Wong V, Roldan Y, Baker DR, Bacharier LB, Beck LA, Ben-Shoshan M, Chipps B, Hoyte F, Israel E, Lang DM, Mathur SK, Nyenhuis SM, O'Byrne PM, Oliver ET, Oppenheimer J, Perry TT, Rivera-Spoljaric K, Saini SS, Sheikh J, Sumino K, Waserman S, Abrams EM, Anagnostou A, Bernstein JA, Ellis AK, Golden DBK, Horner CC, Ledford DK, Lieberman J, Mosnaim G, Rank MA, Shaker M, Wang J, Khan DA, Chu DK.
The journal of allergy and clinical immunology. In practice · 2026-08-25 · PMID 42641984

文献摘要

Randomized controlled trials underpin evidence-based medicine, but a growing proportion of trials contain implausible, inaccurate, or fabricated data. When such studies are incorporated into systematic reviews, they distort effect estimates, inflate evidence certainty, and can mislead the guideline recommendations that the reviews aim to inform. Within 8 recent American Academy of Allergy, Asthma & Immunology and American College of Allergy, Asthma and Immunology Joint Task Force on Practice Parameters (JTFPP) systematic reviews, 17% of trials published between 2021 and 2024 were problematic, and in a network meta-analysis of antihistamines for chronic urticaria, 39% of recent trials were excluded because of implausible data, statistical anomalies, and discrepancies between protocols and the final publication. The proliferation of generative artificial intelligence further facilitates the ability to produce superficially credible but flawed reports. Standardized tools have been developed to offer systematic approaches to detecting problematic trials, but robust safeguards such as verifiable ethics approval, prospective protocol registration, and adherence to structure reporting guidelines remain essential. As part of efforts to ensure quality research, the JTFPP will continue to update our systematic reviews to remove problematic trials as they are identified. We believe that this approach will promote the development of trustworthy guideline recommendations.

在知镜精读原文来源DOI 原文

知镜提供医学文献检索、阅读笔记、人体图谱、统计绘图与研究复现工具。