Is artificial intelligence capable of generating hospital discharge summaries from inpatient records?

Transforming healthcare documentation: harnessing the potential of AI to generate discharge summaries

BACKGROUND

Hospital discharge summaries play an essential role in informing GPs of recent admissions to ensure excellent continuity of care and prevent adverse events; however, they are notoriously poorly written, time-consuming, and can result in delayed discharge.

AIM

To evaluate the potential of artificial intelligence (AI) to produce high-quality discharge summaries equivalent to the level of a doctor who has completed the UK Foundation Programme.

DESIGN & SETTING

Feasibility study using 25 mock patient vignettes.

METHOD

Twenty-five mock patient vignettes were written by the authors. Five junior doctors wrote discharge summaries from the case vignettes (five each). The same case vignettes were input into ChatGPT. In total, 50 discharge summaries were generated; 25 by Al and 25 by junior doctors. Quality and suitability were determined through both independent GP evaluators and adherence to a minimum dataset.

RESULTS

Of the 25 AI-written discharge summaries 100% were deemed by GPs to be of an acceptable quality compared with 92% of the junior doctor summaries. They both showed a mean compliance of 97% with the minimum dataset. In addition, the ability of GPs to determine if the summary was written by ChatGPT was poor, with only a 60% accuracy of detection. Similarly, when run through an AI-detection tool all were recognised as being very unlikely to be written by AI.

CONCLUSION

AI has proven to produce discharge summaries of equivalent quality to a junior doctor who has completed the UK Foundation Programme; however, larger studies with real-world patient data with NHS-approved AI tools will need to be conducted.

The Effect of Teach-Back on Readmission Rates in Rehabilitation Patients

PURPOSE

Thirty-day readmissions often occur in rehabilitation patients and can happen for many reasons. One of those reasons is when patients do not fully understand how to effectively manage their health after discharge. The purpose of this evidence-based quality improvement project was to determine if implementing the teach-back intervention from the Agency for Healthcare Research and Quality's (AHRQ) Health Literacy Universal Precautions Toolkit would impact 30-day readmission rates among adult rehabilitation patients.

METHODS

Data were collected from the electronic health record of rehabilitation patients. The comparative group included all rehabilitation admissions for 8 weeks prior to the intervention. The implementation group was composed of the rehabilitation admissions for 8 weeks post-implementation. All patients were then followed for 30 days postdischarge to capture readmissions.

RESULTS

The total sample size was 79 ( n = 43 in the comparative group, n = 36 in the implementation group). There was a 45% decrease in the mean percentage of the 30-day readmission rate in the implementation group as compared with the comparative group.

CONCLUSION

Based on the results, using the teach-back intervention from AHRQ's Health Literacy Universal Precautions Toolkit may impact 30-day readmission rates.

ChatGPT in Plastic and Reconstructive Surgery

Background  Chat Generative Pre-Trained Transformer (ChatGPT) is a versatile large language model-based generative artificial intelligence. It is proficient in a variety of tasks from drafting emails to coding to composing music to passing medical licensing exams. While the potential role of ChatGPT in plastic surgery is promising, evidence-based research is needed to guide its implementation in practice. Methods  This review aims to summarize the literature surrounding ChatGPT's use in plastic surgery. Results  A literature search revealed several applications for ChatGPT in the field of plastic surgery, including the ability to create academic literature and to aid the production of research. However, the ethical implications of using such chatbots in scientific writing requires careful consideration. ChatGPT can also generate high-quality patient discharge summaries and operation notes within seconds, freeing up busy junior doctors to complete other tasks. However, currently clinical information must still be manually inputted, and clinicians must consider data privacy implications. Its use in aiding patient communication and education and training is also widely documented in the literature. However, questions have been raised over the accuracy of answers generated given that current versions of ChatGPT cannot access the most up-to-date sources. Conclusions  While one must be aware of its shortcomings, ChatGPT is a useful tool for plastic surgeons to improve productivity for a range of tasks from manuscript preparation to healthcare communication generation to drafting teaching sessions to studying and learning. As access improves and technology becomes more refined, surely more uses for ChatGPT in plastic surgery will become apparent.