Innovating the Detection of Care Priorities in Heart Failure Using Large Language Models
2024

Using AI to Improve Care Priorities in Heart Failure

Sample size: 7984 publication Evidence: high

Author Information

Author(s): Albashayreh Alaa, Zeinali Nahid, White Stephanie

Primary Institution: University of Iowa

Hypothesis

Can large language models improve the detection of care priorities in electronic health records for older adults with heart failure?

Conclusion

The study demonstrates that Care-BERT can significantly enhance the documentation of care priorities in electronic health records.

Supporting Evidence

  • Care-BERT outperformed existing models in predicting care priorities.
  • Only 2.8% of patients had comfort measures documented.
  • 17.3% of patients had life-sustaining treatments documented.

Takeaway

This study shows that a special AI model can help doctors understand what older patients with heart failure want for their care.

Methodology

The study retrained a large language model using electronic health record data to predict care priorities.

Potential Biases

Potential biases in the electronic health records used for training the model.

Limitations

The study may not generalize to all patient populations due to its specific focus on older adults with heart failure.

Participant Demographics

Older adults with heart failure, mean age 76.9 years.

Statistical Information

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1093/geroni/igae098.4272

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication