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

Abstract

Medical professionals have been burdened by clerical work, and artificial intelligence may efficiently support physicians by generating clinical summaries. However, whether hospital discharge summaries can be generated automatically from inpatient records stored in electronic health records remains unclear. Therefore, this study investigated the sources of information in discharge summaries. First, the discharge summaries were automatically split into fine-grained segments, such as those representing medical expressions, using a machine learning model from a previous study. Second, these segments in the discharge summaries that did not originate from inpatient records were filtered out. This was performed by calculating the n-gram overlap between inpatient records and discharge summaries. The final source origin decision was made manually. Finally, to reveal the specific sources (e.g., referral documents, prescriptions, and physician’s memory) from which the segments originated, they were manually classified by consulting medical professionals. For further and deeper analysis, this study designed and annotated clinical role labels that represent the subjectivity of the expressions and builds a machine learning model to assign them automatically. The analysis results revealed the following: First, 39% of the information in the discharge summary originated from external sources other than inpatient records. Second, patient’s past clinical records constituted 43%, and patient referral documents constituted 18% of the expressions derived from external sources. Third, 11% of the missing information was not derived from any documents. These are possibly derived from physicians’ memories or reasoning. According to these results, end-to-end summarization using machine learning is considered infeasible. Machine summarization with an assisted post-editing process is the best fit for this problem domain.

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

“…When examining whether AI can generate discharge summaries from inpatient records, it was shown that the origins of information contained in a discharge summary is varied and 39% of the information comes from the inpatient record, with 11% deriving from the physician’s memory. 20 This raises concerns with the feasibility of AI involvement in clinical communication, which aligns with the data governance issues raised by Patel and Lam. 16 They discussed the potential of using ChatGPT to generate discharge summaries and, while they recognised the benefits of automation, standardisation, and efficiency, they also highlighted that errors are frequent, and a minor error could have catastrophic consequences in the clinical environment.…”

“… 16 They, like Ando et al , have recommended that any future AI involvement will always need human oversight to minimise patient safety risks as well as providing more personalised information where needed. 16 , 20 …”

“…This study adds to the increasing evidence of AI being of benefit in the clinical environment. Several studies have explored the use of AI in generating patient clinic letters, including discharge summaries 17,20,21 . The studies are generally positive and, like this one, suggest that AI has a future role, but there are still logistical or even ethical issues that need to be overcome before widescale integration of AI into clinical communication processes is possible.…”

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

“…Examining ways to standardize how teach-back is implemented and delivered to patients may achieve better results. Artificial intelligence may have a future role in a patient’s discharge process (Ando et al, 2022; Patel & Lam, 2023). Using artificial intelligence models may communicate with patients in a style or level that is better understood.…”

ChatGPT in Plastic and Reconstructive Surgery

“…The authors, therefore, concluded that AI's can only be a tool used as part of the discharge summary writing process. 26 …”