Artificial intelligence and the future of patient-centered outcomes

Area of Science:

• Health outcomes research
• Artificial intelligence in healthcare
• Patient-reported outcome measures

Background:

• Large language models (LLMs) offer a new vision for patient-reported outcome measures (PROMs).
• Generative artificial intelligence (GenAI) presents opportunities to advance patient-centered outcome assessments.

Purpose of the Study:

• To support the vision of LLM-enabled PROMs.
• To explore the potential of GenAI in assessing patient-centered outcomes.
• To highlight GenAI's role in improving PROM development and data collection.

Main Methods:

• Leveraging GenAI for scalable, in-depth patient interviews.
• Training GenAI interviewers in interview techniques and concept intent.
• Utilizing GenAI for consistent coding of patient responses based on conversational data.

Main Results:

• GenAI can improve the quality and efficiency of traditional PROMs.
• GenAI enables tailored questioning and clarification of meaning, similar to human interviewers.
• GenAI overcomes the scalability limitations of human interviewers in data collection.

Conclusions:

• The health outcomes field should investigate GenAI's capabilities for collecting patient experience data.
• Rigorous evaluation of the quality of GenAI-derived patient assessments is crucial.
• GenAI has the potential to revolutionize patient-centered outcome research.