AME Clin Trials Rev. 2026 Apr 25. 4
Kaiping Zhang,
Savvas Lampridis,
Charles B Simone,
Ionut Negoi,
Yan Peng,
Binghan Shang,
Yao Lin,
Yaling Cheng,
Fanghui Yang,
Vishal G Shelat,
Kamyar Kalantar-Zadeh.
Background and Objective: Clinical trials are crucial for evidence-based medicine; however, substantial challenges remain. This review aims to provide an overview of recurring challenges in clinical trials and offer potential strategies to overcome practical barriers for stakeholders.
Methods: A search of PubMed for English-language papers, published from January 1, 2021 to July 31, 2025 was conducted, focusing on reflections on trial registration, protocols, statistical analysis plans (SAP), sample size, risk of bias, transparency, reporting, peer review, dissemination, and emerging areas such as artificial intelligence (AI).
Key Content and Findings: The clinical trial landscape has expanded dramatically, surpassing one million trials in total. However, it is marked by significant redundancy, waste, and lack of reproducibility. Transparency remains hampered by inadequate prospective registration, unreported results, limited protocol and SAP availability, substandard registration quality, and a lack of core requirements on registration platforms, despite improvements under a series of policy initiatives. Regarding quality and integrity, many trials have a high risk of bias, design flaws, underpowered sample sizes, or uncertain findings. Data fabrication and retractions due to dishonesty contribute to the complexity of this landscape, along with a peer-review workflow that underutilizes appraisal in the pre-submission, preprint, and post-publication stages. Regarding reporting and dissemination, challenges include poor adherence to the CONSORT and SPIRIT guidelines, ambiguous adoption of reporting guidelines by journals, heavy burdens on researchers using reporting guidelines, severe spin reporting of results, and inaccurate public dissemination. In emerging areas, AI is rapidly developing in almost every aspect of clinical trials, including its use in assisting with reviewing the risk of bias and integrity. The most problematic issues are the insufficient disclosure of AI use and inadequate human verification. This review proposes 18 suggestions and 19 strategies to address these concerns, such as requiring registration prior to ethical approval by ethical committees, founding journals dedicated to statistically negative trials, and developing an integrated trial quality feedback and fixing mechanism.
Conclusions: There are pressing challenges with uncontrolled trial expansion, insufficient transparency, poor quality, dishonest or wrong practices, biased reporting and dissemination, and insufficient disclosure and verification of AI. The proposed suggestions and strategies may contribute to a healthier clinical trial ecosystem if implemented.
Keywords: Clinical trials; artificial intelligence (AI); quality; registry; reporting