bims-arines 2025-04-06 papers

PLoS One. 2025 ;20(4): e0320151

Using artificial intelligence tools to automate data extraction for living evidence syntheses.

Evan Mitchell, Elisha B Are, Caroline Colijn, David J D Earn.

Living evidence synthesis (LES) involves repeatedly updating a systematic review or meta-analysis at regular intervals to incorporate new evidence into the summary results. It requires a considerable amount of human time investment in the article search, collection, and data extraction phases. Tools exist to automate the retrieval of relevant journal articles, but pulling data out of those articles is currently still a manual process. In this article, we present a proof-of-concept Python program that leverages artificial intelligence (AI) tools (specifically, ChatGPT) to parse a batch of journal articles and extract relevant results, greatly reducing the human time investment in this action without compromising on accuracy. Our program is tested on a set of journal articles that estimate the mean incubation period for COVID-19, an epidemiological parameter of importance for mathematical modelling. We also discuss important limitations related to the total amount of information and rate at which that information can be sent to the AI engine. This work contributes to the ongoing discussion about the use of AI and the role such tools can have in scientific research.

DOI: https://doi.org/10.1371/journal.pone.0320151

J Med Econ. 2025 Apr 01. 1-19

Artificial Intelligence in key pricing, reimbursement and market access (PRMA) processes: better, faster, cheaper - can you really pick two?

Eva Susanne Dietrich.

The rapid evolution of large language models (LLMs) and machine learning (ML) presents both significant opportunities and challenges for market access processes. These sophisticated AI systems, built on transformer architectures and extensive datasets, offer potential to forecast claims and decisions of health technology assessment (HTA) agencies and streamline processes such as systematic literature reviews and HTA submissions. Furthermore, the analysis of real-world data - also for deriving causal relationships - is being discussed intensively. Despite notable advancements, their adoption in key PRMA processes is still limited at present, with only a small fraction of submissions to HTA bodies incorporating AI. Key barriers include stringent transparency requirements, the necessity of explainability and human oversight in data analyses, and the highly sensitive nature of text drafting - especially in cases where reimbursement decisions or pricing negotiations balance on a knife's edge. These requirements are often not met due to the immaturity of many AI applications, which still lack the necessary precision, reliability, and contextual understanding. Moreover, AI-generated evidence has yet to prove its validity before it can supplement or replace traditional study designs, such as randomized controlled trials (RCTs), which are critical for HTA decisions. Additionally, the environmental and financial costs of training LLMs require careful assessment. This paper explores various current AI applications, their limitations, and future prospects in key PRMA processes from a German perspective while also considering the broader implications of the EU Health Technology Assessment Regulation (HTAR). It concludes that while AI hold transformative potential, its integration into workflows must be approached cautiously, with incremental adoption, and close collaboration between industry, HTA agencies, and academia. Demonstrating robust, unbiased comparative evidence-showcasing superior performance and cost savings over traditional methods-could accelerate the adoption process.

Keywords: AI; AMNOG; H51; HTA; I18; L65; LLM; PRMA; artificial intelligence; drugs; health technology assessment; large language models; market access; pricing; reimbursement; systematic reviews

DOI: https://doi.org/10.1080/13696998.2025.2488154

Conserv Biol. 2025 Apr;39(2): e14464

New opportunities and challenges for conservation evidence synthesis from advances in natural language processing.

Charlotte H Chang, Susan C Cook-Patton, James T Erbaugh, Luci Lu, Yuta J Masuda, István Molnár, Dávid Papp, Brian E Robinson.

Addressing global environmental conservation problems requires rapidly translating natural and conservation social science evidence to policy-relevant information. Yet, exponential increases in scientific production combined with disciplinary differences in reporting research make interdisciplinary evidence syntheses especially challenging. Ongoing developments in natural language processing (NLP), such as large language models, machine learning (ML), and data mining, hold the promise of accelerating cross-disciplinary evidence syntheses and primary research. The evolution of ML, NLP, and artificial intelligence (AI) systems in computational science research provides new approaches to accelerate all stages of evidence synthesis in conservation social science. To show how ML, language processing, and AI can help automate and scale evidence syntheses in conservation social science, we describe methods that can automate querying the literature, process large and unstructured bodies of textual evidence, and extract parameters of interest from scientific studies. Automation can translate to other research agendas in conservation social science by categorizing and labeling data at scale, yet there are major unanswered questions about how to use hybrid AI-expert systems ethically and effectively in conservation.

Keywords: aprendizaje automático; ciencias sociales de la conservación; conservation social science; evidence synthesis; language models; machine learning; modelos lingüísticos; natural language processing; procesamiento lingüístico natural; síntesis de evidencias

DOI: https://doi.org/10.1111/cobi.14464

JMIR Med Educ. 2025 Apr 02. 11 e72998

Citation Accuracy Challenges Posed by Large Language Models.

Manlin Zhang, Tianyu Zhao.

Keywords: AI; AI-based tools; LLM; Saudi Arabia; artificial intelligence; chatGPT; chatbot; faculty; knowledge; large language model; learning; medical education; medical students; perceptions; qualitative study; satisfaction; thematic analysis; universities

DOI: https://doi.org/10.2196/72998

Conserv Sci Pract. 2025 Jan 23. 7(2): e13278

Web-based ecological evidence entry form enables consistent, accessible extraction and visualization for synthesis applications.

Caroline E Ridley, Casey Hansen, Peter Byrley, Tara Greaver, S Douglas Kaylor, R Byron Rice, Kate A Schofield, Andrew Shapiro.

Applying scientific evidence to conservation, environmental management, and policy-making improves outcomes. When synthesizing existing evidence, substantial resources are required to access and read scientific publications and extract and analyze decision-relevant information. To improve this process, we developed a free, publicly available, web-based evidence entry form tailored to extract information about cause-effect relationships from ecological publications. The form enables storage, retrieval, reuse, and visualization of qualitative and quantitative ecological and environmental evidence extracted from publications. Evidence can be analyzed for a wide range of synthesis purposes (e.g., causal assessments, hypothesis testing) and approaches (e.g., rapid reviews, meta-analyses). The database schema underlying the form logically relates information about (a) a publication, (b) its experimental design(s), and (c) reported cause-effect relationships. An ontology of controlled terminology enables consistent extraction and characterization of causes and effects across users, facilitating evidence reuse. Future capabilities include customization of terminology and incorporation of study quality information.

Keywords: assessment; conservation; controlled vocabulary; data visualization; ecology; environmental management; evidence; synthesis; systematic review; web-based content management

DOI: https://doi.org/10.1111/csp2.13278