language model
Conditions
Interventions
Group 1: Intraoperative data collection (using audiovisual recording equipment) following the patient’s prior consent.
Subsequently, a preliminary surgical report is generated by AI. Comparison with a
Sponsors
Herz- und Diabeteszentrum NRW, Klinik für Thorax- und Kardiovaskularchirurgie Universitätsklinik der Ruhr-Universität Bochum Medizinische Fakultät OWL (Universität Bielefeld)
Eligibility
Sex/Gender
All
Age
18 Years to No maximum
Inclusion criteria
Inclusion criteria: Written informed consent available; Patients aged 18 and older; Routine procedures (e.g. isloated aortic valve surgery, mitral valve surgery)
Exclusion criteria
Exclusion criteria: Pregnancy; patients with psychiatric or neurological disorders associated with impaired legal capacity;
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Overall quality of the generated surgical reports (measured using automated metrics (ROUGE) and determined manually through annotation). | — |
Secondary
| Measure | Time frame |
|---|---|
| 1. Effort required for manual correction (measured in time) 2. Potential savings from using a language model (measured in time and monetary savings) 3. Types and severity of model errors 4. Added value of multimodal input data (e.g., audio or video) These values are reviewed after the surgical report is generated. | — |
Countries
Germany
Contacts
Public ContactHeike Windhagen
Herz- und Diabeteszentrum NRW, Klinik für Thorax- und Kardiovaskularchirurgie Universitätsklinik der Ruhr-Universität Bochum Medizinische Fakultät OWL (Universität Bielefeld)
Outcome results
None listed