Concerns of Using Large Language Models in Health Care Research and Practice: Umbrella Review.

Area of Science:

• Healthcare research
• Artificial Intelligence in Medicine
• Large Language Models (LLMs)

Background:

• Rapid advancements in Large Language Models (LLMs) like ChatGPT are increasing their use in healthcare.
• There's a growing need for automation and AI support in medical research and practice.

Purpose of the Study:

• To examine concerns of healthcare professionals and researchers regarding LLM use in healthcare.
• To identify common issues and their implications for patient care, policy, and practice.

Main Methods:

• An umbrella review methodology was employed, registering the protocol on PROSPERO.
• Searches across 7 databases identified systematic reviews published after January 2017 concerning LLM use in healthcare.
• Included reviews underwent quality appraisal (AMSTAR-2) and certainty of evidence assessment (GRADE), with data synthesized narratively.

Main Results:

• 42 systematic reviews were included, covering 12 distinct populations, including researchers and clinicians.
• Reviews were of very poor quality, with significant overlap potential.
• Key themes identified were technical capability, ethical/legal/societal issues, and costs associated with LLMs.

Conclusions:

• This is the first umbrella review on LLM concerns in healthcare, revealing common narratives but limited by poor study quality.
• Addressing data quality and ethical, legal, and societal implications is vital for responsible AI adoption.
• Healthcare must adapt to accelerating technology, prioritizing equity, diversity, inclusion, and safety.