The quality and safety of using generative AI to produce patient-centred discharge instructions

Kristian Stanceski^{1

2

3}, Sharleen Zhong⁴, Xumou Zhang⁴, Sam Khadra², Marguerite Tracy^{5

6

7}, Linda Koria⁸, Sarita Lo⁸, Vasi Naganathan^{8

9}, Jinman Kim⁴, Adam G Dunn¹⁰, Julie Ayre⁵

Affiliations

¹ Biomedical Informatics and Digital Health, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia.
² Royal Prince Alfred Hospital, Sydney Local Health District, Sydney, NSW, Australia.
³ School of Clinical Medicine, UNSW Medicine & Health, St George and Sutherland Clinical Campus, Sydney, NSW, Australia.
⁴ School of Computer Science, Faculty of Engineering, The University of Sydney, Sydney, NSW, Australia.
⁵ Sydney Health Literacy Lab, Sydney School of Public Health, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia.
⁶ General Practice Clinical School, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia.
⁷ Drug Health Services, Blacktown and Mount Druitt Hospitals, Western Sydney Local Health District, Sydney, NSW, Australia.
⁸ Centre for Education and Research on Ageing, Department of Geriatric Medicine, Concord Repatriation Hospital, Sydney, NSW, Australia.
⁹ Concord Clinical School, Faculty of Medicine and Health, University of Sydney, Sydney, NSW, Australia.
¹⁰ Biomedical Informatics and Digital Health, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia. adam.dunn@sydney.edu.au.

PMID: 39567722
PMCID: PMC11579500
DOI: 10.1038/s41746-024-01336-w

The quality and safety of using generative AI to produce patient-centred discharge instructions

Kristian Stanceski et al. NPJ Digit Med. 2024.

. 2024 Nov 20;7(1):329.

doi: 10.1038/s41746-024-01336-w.

Authors

Affiliations

¹ Biomedical Informatics and Digital Health, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia.
² Royal Prince Alfred Hospital, Sydney Local Health District, Sydney, NSW, Australia.
³ School of Clinical Medicine, UNSW Medicine & Health, St George and Sutherland Clinical Campus, Sydney, NSW, Australia.
⁴ School of Computer Science, Faculty of Engineering, The University of Sydney, Sydney, NSW, Australia.
⁵ Sydney Health Literacy Lab, Sydney School of Public Health, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia.
⁶ General Practice Clinical School, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia.
⁷ Drug Health Services, Blacktown and Mount Druitt Hospitals, Western Sydney Local Health District, Sydney, NSW, Australia.
⁸ Centre for Education and Research on Ageing, Department of Geriatric Medicine, Concord Repatriation Hospital, Sydney, NSW, Australia.
⁹ Concord Clinical School, Faculty of Medicine and Health, University of Sydney, Sydney, NSW, Australia.
¹⁰ Biomedical Informatics and Digital Health, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW, Australia. adam.dunn@sydney.edu.au.

PMID: 39567722
PMCID: PMC11579500
DOI: 10.1038/s41746-024-01336-w

Abstract

Patient-centred instructions on discharge can improve adherence and outcomes. Using GPT-3.5 to generate patient-centred discharge instructions, we evaluated responses for safety, accuracy and language simplification. When tested on 100 discharge summaries from MIMIC-IV, potentially harmful safety issues attributable to the AI tool were found in 18%, including 6% with hallucinations and 3% with new medications. AI tools can generate patient-centred discharge instructions, but careful implementation is needed to avoid harms.

PubMed Disclaimer

Conflict of interest statement

Competing interests: J.A. is a director of a health literacy consultancy, which provides health literacy advice to health services and organisations, but no personal income is received. A.D. is Deputy Editor for npj Digital Medicine. All other authors declare no competing interests.

Figures

**Fig. 1. A review of safety issues identified in the AI-generated patient-centred discharge instructions.**
The safety issues were identified in 100 patient-centred discharge instructions, and descriptions include the severity and provenance of the safety issues. Four examples are identified on the figure based on their error category and source.

See this image and copyright information in PMC

References

1. Wimsett, J., Harper, A. & Jones, P. Review article: components of a good quality discharge summary: a systematic review. Emerg. Med. Australas.26, 430–438 (2014). - PubMed
1. Scarfo, N. L. et al. General practitioners’ perspectives on discharge summaries from a health network of three hospitals in South Australia. Aust. Health Rev.47, 433–440 (2023). - PubMed
1. Weetman, K., Spencer, R., Dale, J., Scott, E. & Schnurr, S. What makes a “successful” or “unsuccessful” discharge letter? Hospital clinician and General Practitioner assessments of the quality of discharge letters. BMC Health Serv. Res.21, 349 (2021). - PMC - PubMed
1. Mixon, A. S. et al. Characteristics associated with postdischarge medication errors. Mayo Clin. Proc.89, 1042–1051 (2014). - PMC - PubMed
1. Hoek, A. E. et al. Patient discharge instructions in the emergency department and their effects on comprehension and recall of discharge instructions: a systematic review and meta-analysis. Ann. Emerg. Med.75, 435–444 (2020). - PubMed

LinkOut - more resources

Full Text Sources
- Nature Publishing Group
- PubMed Central

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

The quality and safety of using generative AI to produce patient-centred discharge instructions

Affiliations

The quality and safety of using generative AI to produce patient-centred discharge instructions

Authors

Affiliations

Abstract

Conflict of interest statement

Figures

References

LinkOut - more resources

Full Text Sources