

Ep157: AI models for discharge letters and clinical coding
In part 1 of this series we talked about the efficiency gains from use of ambient scribes observed in some hospital settings. In this episode we discuss other opportunities that arise from granting large language models access to all the notes and test results in the patient record. The potential comes most clearly into perspective in regards to discharge summaries. Discharge from hospital is a notorious example of discontinuity in care, with up to half of discharged patients experiencing a related medical error over a year. This results from delays in the release of discharge letters to community practitioners, and gaps or misunderstandings in the communications contained within them. Inconsistent standards of clinical documentation also lead to inaccurate coding of clinical activity that fails to reflect the true complexity of care. This means that hospitals are inadequately remunerated through the activity-based funding model for Australian public health services. For private practitioners, poor documentation of clinical encounters also exposes them to risks should they ever be investigated by the Professional Services Review about their billing activity. We also question the light touch that regulators take towards ambient scribes and large language models. Chapters 1:20 Reliable discharge summaries 11:53 Improved clinical coding 21:55 Safety regulation 30:51 Patient satisfaction Guests Dr Andrew Vanlint FRACP AFRACMA (North Adelaide Local Health Network; Adelaide University; Genesis Care) Professor Ian Scott FRACP MHA MEd (Princess Alexandra Hospital, Metro South Health; University of Queensland) Dr Adam Brand FACEM MPH (Gold Coast University Hospital) Production Production by Mic Cavazzini DPhil. Music licenced from Epidemic Sound includes ‘Midnight sun serenade’ and ‘End of the Ocean’ by Tellsonic and ‘Multicolor’ by Chill Cole. Feedback on this episode kindly provided by Dr Aidan Tan, Dr Simeon Wong, Dr Joseph Lee, Assoc Prof Paul Cooper PhD and Loryn Einstein. Image by da-kuk purchased from Getty Images. Visit the Pomegranate Health web page for a transcript and supporting references. Add educational activity to MyCPD. Subscribe through any podcasting app or our email alerts list.
- Transcript
- Chapters


















