PRECISE Framework: GPT-based Text For Improved Readability, Reliability, and Understandability of Radiology Reports For Patient-Centered Care

Satvik Tripathi,Liam Mutter,Meghana Muppuri,Suhani Dheer,Emiliano Garza-Frias,Komal Awan,Aakash Jha,Michael Dezube,Azadeh Tabari,Christopher P. Bridge,Dania Daye

2024-02-20

Abstract:This study introduces and evaluates the PRECISE framework, utilizing OpenAI's GPT-4 to enhance patient engagement by providing clearer and more accessible chest X-ray reports at a sixth-grade reading level. The framework was tested on 500 reports, demonstrating significant improvements in readability, reliability, and understandability. Statistical analyses confirmed the effectiveness of the PRECISE approach, highlighting its potential to foster patient-centric care delivery in healthcare decision-making.

Computation and Language,Artificial Intelligence,Human-Computer Interaction,Machine Learning

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the readability of radiology reports and patients' understanding, especially by simplifying medical language to make the reports more understandable. Specifically, the research aims to enhance patients' understanding and participation by introducing and evaluating the **PRECISE** framework driven by OpenAI's GPT - 4, providing clearer and more accessible radiology reports with a sixth - grade level of understanding difficulty. This improvement is of great significance for promoting patient - centered medical decision - making.

### Background problems
1. **Complexity of medical information**: Medical documents, especially radiology reports, usually contain highly technical language and lack patient - centered information, which makes it difficult for patients to understand their own disease states.
2. **21st Century Cures Act**: Since its implementation on April 5, 2021, this act requires that all medical documents be released to patients immediately after completion, which brings new challenges to communication between doctors and patients.
3. **Patient engagement**: If patients can better understand medical information, they will often be more actively involved in their own treatment, thus obtaining better health outcomes and treatment satisfaction.
4. **Doctor burden**: Doctors need to spend a lot of time explaining complex medical information, which increases administrative workload, may lead to doctor fatigue and increased liability, and affects the overall quality of medical care.

### Solutions
- **PRECISE framework**: By using GPT - 4 to generate patient - friendly radiology report summaries, make the language of the reports more concise, clear and effective, while reducing the administrative burden on doctors.
- **Evaluation methods**:
- **Readability**: Use the Flesch Reading Ease, Gunning Fog Index and Automated Readability Index to evaluate the readability of the text.
- **Reliability**: Have clinical volunteers with medical degrees evaluate whether the generated text accurately summarizes all the information in the original report.
- **Understandability**: Have volunteers without a medical background evaluate whether the generated text is easy to understand.

### Results
- **Significantly improved readability**: The Flesch Reading Ease score increased from 38.28 to 80.82, the Gunning Fog Index score decreased from 13.04 to 6.99, and the Automated Readability Index score decreased from 13.33 to 5.86.
- **High reliability**: 95% of the generated summaries were considered reliable by clinical volunteers.
- **Good understandability**: 97% of the summaries generated by PRECISE were considered fully understood by volunteers without a medical background.

### Conclusion
The application of the PRECISE framework significantly improves the readability and understandability of radiology reports. By providing reliable patient - friendly summaries, it helps promote patient participation and understanding of medical decision - making, while reducing the burden on doctors and promoting more inclusive and patient - centered medical services.