Using Generative Artificial Intelligence in Health Economics and Outcomes Research: A Primer on Techniques and Breakthroughs DOI Creative Commons

T Reason,

S. Klijn, William D. Rawlinson

et al.

PharmacoEconomics - Open, Journal Year: 2025, Volume and Issue: unknown

Published: April 29, 2025

The emergence of generative artificial intelligence (GenAI) offers the potential to enhance health economics and outcomes research (HEOR) by streamlining traditionally time-consuming labour-intensive tasks, such as literature reviews, data extraction, economic modelling. To effectively navigate this evolving landscape, economists need a foundational understanding how GenAI can complement their work. This primer aims introduce essentials using tools, particularly large language models (LLMs), in HEOR projects. For new technologies, chatbot interfaces like ChatGPT offer an accessible way explore LLMs. more complex projects, knowledge application programming (APIs), which provide scalability integration capabilities, prompt engineering strategies, few-shot chain-of-thought prompting, is necessary ensure accurate efficient analysis, model performance, tailor outputs specific needs. Retrieval-augmented generation (RAG) further improve LLM performance incorporating current external information. LLMs have significant many common summarising medical literature, extracting structured data, drafting report sections, generating statistical code, answering questions, reviewing materials quality. However, must also be aware ongoing limitations challenges, propensity produce inaccurate information ('hallucinate'), security concerns, issues with reproducibility, risk bias. Implementing requires robust protocols handle sensitive compliance European Union's General Data Protection Regulation (GDPR) United States' Health Insurance Portability Accountability Act (HIPAA). Deployment options local hosting, secure API use, or cloud-hosted open-source varying levels control cost, each unique trade-offs security, accessibility, technical demands. Reproducibility transparency pose challenges. credibility LLM-generated content, explicit declarations version, prompting techniques, benchmarks against established standards are recommended. Given 'black box' nature LLMs, clear reporting structure essential maintain validate outputs, enabling stakeholders assess reliability accuracy analyses. ethical implications (AI) HEOR, including multifaceted, requiring careful assessment use case determine level scrutiny transparency. balance benefits AI adoption risks maintaining practices, while considering accountability, bias, intellectual property, broader impact on healthcare system. As technologies advance, role will become increasingly evident. Key areas promise include creating dynamic, continuously updated materials, providing patients information, enhancing analytics for faster access medicines. maximise these benefits, understand address challenges ownership coming years critical establishing best practices HEOR. encourages adopt responsibly, balancing innovation scientific rigor integrity insights decision-making.

Language: Английский

Using Generative Artificial Intelligence in Health Economics and Outcomes Research: A Primer on Techniques and Breakthroughs DOI Creative Commons

T Reason,

S. Klijn, William D. Rawlinson

et al.

PharmacoEconomics - Open, Journal Year: 2025, Volume and Issue: unknown

Published: April 29, 2025

The emergence of generative artificial intelligence (GenAI) offers the potential to enhance health economics and outcomes research (HEOR) by streamlining traditionally time-consuming labour-intensive tasks, such as literature reviews, data extraction, economic modelling. To effectively navigate this evolving landscape, economists need a foundational understanding how GenAI can complement their work. This primer aims introduce essentials using tools, particularly large language models (LLMs), in HEOR projects. For new technologies, chatbot interfaces like ChatGPT offer an accessible way explore LLMs. more complex projects, knowledge application programming (APIs), which provide scalability integration capabilities, prompt engineering strategies, few-shot chain-of-thought prompting, is necessary ensure accurate efficient analysis, model performance, tailor outputs specific needs. Retrieval-augmented generation (RAG) further improve LLM performance incorporating current external information. LLMs have significant many common summarising medical literature, extracting structured data, drafting report sections, generating statistical code, answering questions, reviewing materials quality. However, must also be aware ongoing limitations challenges, propensity produce inaccurate information ('hallucinate'), security concerns, issues with reproducibility, risk bias. Implementing requires robust protocols handle sensitive compliance European Union's General Data Protection Regulation (GDPR) United States' Health Insurance Portability Accountability Act (HIPAA). Deployment options local hosting, secure API use, or cloud-hosted open-source varying levels control cost, each unique trade-offs security, accessibility, technical demands. Reproducibility transparency pose challenges. credibility LLM-generated content, explicit declarations version, prompting techniques, benchmarks against established standards are recommended. Given 'black box' nature LLMs, clear reporting structure essential maintain validate outputs, enabling stakeholders assess reliability accuracy analyses. ethical implications (AI) HEOR, including multifaceted, requiring careful assessment use case determine level scrutiny transparency. balance benefits AI adoption risks maintaining practices, while considering accountability, bias, intellectual property, broader impact on healthcare system. As technologies advance, role will become increasingly evident. Key areas promise include creating dynamic, continuously updated materials, providing patients information, enhancing analytics for faster access medicines. maximise these benefits, understand address challenges ownership coming years critical establishing best practices HEOR. encourages adopt responsibly, balancing innovation scientific rigor integrity insights decision-making.

Language: Английский

Citations

0