Predicting Publication of Clinical Trials Using Structured and Unstructured Data: Model Development and Validation Study

Rates of discontinuation and non-publication of endovascular stroke clinical studies: A cross-sectional analysis

BACKGROUND

The discontinuation and non-publication of clinical studies in various medical fields undermine research efforts and may bias the medical evidence base. This study investigates the prevalence and factors associated with these issues in endovascular stroke studies.

METHODS

Clinical Trials.gov was searched for all studies registered from inception up till May 2022 and included patients with endovascular stroke. Publications from these studies were identified by extensive online searching using the NCT identifier number and other related keywords, and multiple logistic regression analysis was performed to identify characteristics associated with study discontinuation and non-publication.

RESULTS

Our search yielded 88 endovascular stroke studies, all including both genders. Among these, 63 (71.6 %) were completed, and 25 (28.4 %) were discontinued. Of the completed studies, 40 (63.5 %) were published. The majority of trials were single-centric (57 %), had a large sample size of >=100 (53.4 %), involved only adults (97.6 %), and were primarily funded by non-industrial sources (70 %). Discontinuation was more frequent in smaller-sized, single-center trials, but the associations diminished after adjustment. The sample size was a significant predictor of non-publication even after adjustment.

CONCLUSION

There is evidence of non-dissemination bias in clinical studies of endovascular stroke. These biases distort the therapeutic information available to inform clinical practice and raise ethical concerns regarding exposing volunteering participants to potential risks without furthering practice.

Zero- and few-shot prompting of generative large language models provides weak assessment of risk of bias in clinical trials

Existing systems for automating the assessment of risk-of-bias (RoB) in medical studies are supervised approaches that require substantial training data to work well. However, recent revisions to RoB guidelines have resulted in a scarcity of available training data. In this study, we investigate the effectiveness of generative large language models (LLMs) for assessing RoB. Their application requires little or no training data and, if successful, could serve as a valuable tool to assist human experts during the construction of systematic reviews. Following Cochrane's latest guidelines (RoB2) designed for human reviewers, we prepare instructions that are fed as input to LLMs, which then infer the risk associated with a trial publication. We distinguish between two modelling tasks: directly predicting RoB2 from text; and employing decomposition, in which a RoB2 decision is made after the LLM responds to a series of signalling questions. We curate new testing data sets and evaluate the performance of four general- and medical-domain LLMs. The results fall short of expectations, with LLMs seldom surpassing trivial baselines. On the direct RoB2 prediction test set (n = 5993), LLMs perform akin to the baselines (F1: 0.1-0.2). In the decomposition task setup (n = 28,150), similar F1 scores are observed. Our additional comparative evaluation on RoB1 data also reveals results substantially below those of a supervised system. This testifies to the difficulty of solving this task based on (complex) instructions alone. Using LLMs as an assisting technology for assessing RoB2 thus currently seems beyond their reach.

The Current Landscape of COPD-Related Clinical Trials Registered on the World Health Organization's International Clinical Trials Registry Platform: A Comprehensive Analysis of Study Characteristics and Publication Status

Background

Despite studies investigating the publication rates and factors influencing publication outcomes of clinical trials in some disease fields, there is a notable lack of research focusing on chronic obstructive pulmonary disease (COPD) clinical trials. This study aims to explore the characteristics of COPD-related clinical trials and identify factors associated with publication status and publication time.

Methods

A systematic search was conducted on the World Health Organization International Clinical Trials Registry Platform on April 28, 2022, to identify completed interventional clinical trials related to COPD. Various trial features were analyzed, and factors influencing publication status and time were examined.

Results

A total of 2577 completed interventional clinical trials focusing on COPD were identified. A total of 42.76% of trials enrolled ≤50 participants. The majority of trials were randomized (81.72%), blind (57.39%), parallel-assignment (59.14%), single-center (51.30%), multi-arm (83.86%), nonindustry funded (52.00%), and conducted for therapeutic purposes (73.11%). The 2-year cumulative publication rate was found to be 27.9%. The median time of study duration, dissemination lag, and publication lag were 17.27, 21.07, and 24.70 months, respectively. Multivariate analysis revealed that sample size, blind design, and study phase significantly influenced the likelihood of publication, while intervention model, primary purpose, study phase, funder, and study duration were significant factors affecting publication time.

Conclusions

The findings highlight the inadequacy of large multi-center interventional clinical trials for COPD and indicate a low 2-year cumulative publication rate. Strengthening collaboration among investigators and adopting scientifically robust designs for larger phase 3 clinical trials are crucial to advancing COPD research and enhancing publication outcomes.

Transforming clinical trials: the emerging roles of large language models

Clinical trials are essential for medical research, but they often face challenges in matching patients to trials and planning. Large language models (LLMs) offer a promising solution, signaling a transformative shift in the field of clinical trials. This review explores the multifaceted applications of LLMs within clinical trials, focusing on five main areas expected to be implemented in the near future: enhancing patient-trial matching, streamlining clinical trial planning, analyzing free text narratives for coding and classification, assisting in technical writing tasks, and providing cognizant consent via LLM-powered chatbots. While the application of LLMs is promising, it poses challenges such as accuracy validation and legal concerns. The convergence of LLMs with clinical trials has the potential to revolutionize the efficiency of clinical trials, paving the way for innovative methodologies and enhancing patient engagement. However, this development requires careful consideration and investment to overcome potential hurdles.