Coastal Hospital Group
Assisting doctors with Generative AI
The context
Assistance with drafting discharge letters for internal medicine
Following an invitation to an event organised by Synotis in March 2024 in Geneva, the Centre Hospitalier de la Côte (EHC) requested support, as it was looking to optimize medical time, particularly on the drafting of hospital discharge letters, which are mandatory after each stay, and the manual transcription of laboratory results, a time-consuming and error-prone task.
THE OBJECTIVE
Reduce writing time, improve the quality of exit letters
The hospital discharge summary, mandatory after each stay, requires considerable time for drafting and proofreading. The project was initiated with the aim of relieving doctors of some of their administrative burden. After analysis, two priority features were identified:
- a contextual prediction of the investigation text based on the patient's diagnoses and laboratory results, offering superior personalized content compared to generic templates;
- the automation of the retrieval and formatting of laboratory results, eliminating the risk of copying errors while improving the document structure and reducing the workload of writers.
Our approach
Involve doctors and compare models
- Framework for prioritizing features to be developed
- Exchange workshop
- 10 proposed and costed features
- 2 features developed as a priority, according to the client's choice
- Training the model and confronting it with reality
- Nearly 3,000 exit letters were used for training.
- The reality of the historical data differed from certain best practices — this had to be taken into account for the predictions.
- Presentation to > 15 doctors
- Feedback from the business and the suitability of the proposed solution to the needs are paramount.
The result
AI for a concrete impact in real life
The presentation to physicians at the end of the Proof of Concept (PoC) in Live Demo mode was a success. All the physicians present felt that the proposed solution could save them time when writing discharge letters. Their self-estimates ranged from 10 minutes for a simple letter to 30 minutes for a complex one. Regardless of the accuracy of this self-assessment, the crucial element for us is that the physicians perceive a benefit, both in terms of time saved and ease of use . The user-friendly interface that can assist with writing by suggesting missing values in the text was very well received.
Following this demonstration, it was decided to proceed with the industrialization of the solution. This will be done in phases, each validated jointly by the IT department and the physicians, always with the aim of providing a solution that effectively addresses the observed problems.
Transforming the field data of a major healthcare player into actionable insights using generative AI
Leveraging the power of Data and AI to drive Somfy Group’s competitive advantage
Introducing IDEA's business teams and leaders to Generative AI
Acculturation to Generative AI: BDO Teams Train in This New Technology and Define Business Use Cases
How Snowleader became a major player in the French outdoor industry, exceeding 100 million euros in revenue in ten years.