Abstract: SA-PO0975
Artificial Intelligence (AI)-Assisted Educational Videos for CKRT and Intermittent Hemodialysis (iHD) Improve Trainee Knowledge and Confidence
Session Information
- Educational Research
October 24, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Educational Research
- 1000 Educational Research
Authors
- Fichadiya, Harshil, Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Balakrishnan, Suryanarayanan, Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Abdelgadir, Yasir, Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Alam, Sreyoshi Fatima, Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Bizer, Benjamin, Mayo Clinic Minnesota, Rochester, Minnesota, United States
- Cheungpasitporn, Wisit, Mayo Clinic Minnesota, Rochester, Minnesota, United States
Background
AI can assist educators in presenting complex concepts in simplified and visually engaging formats. iHD and CRRT are among the more complex topics in nephrology and may benefit from animation-based learning. We developed a 10-video AI-assisted CRRT/iHD educational series for IM residents, APP & 1st nephrology fellows
Methods
Educational content provided by senior nephrology fellows was converted into 10 videos using cinematic and explainer modes of NotebookLM. Videos covered key CRRT/iHD concepts. Videos were graded using a 5-point Likert scale. Learners completed pre- and post-intervention questionnaires assessing core CRRT knowledge and application. AI-generated songs (Suno AI) were also used to reinforce key concepts
Results
Videos demonstrated high educational quality with mean scores (out of 5) for 4.84 (accuracy), 4.92 (completeness), 4.95 (ease of comprehension), with minimal hallucinations (4.68) and good graphics (4.56). Learner-reported comfort with ordering CRRT improved significantly, from 2/5 to 4/5, alongside improved confidence in identifying indications of CRRT over iHD (mean 2.8/5 vs 4.45/5). Knowledge improved across domains (Table 1). Recognition of CRRT indications increased from 78% to 100% (hemodynamic instability). Understanding of key mechanisms improved, including determinant of clearance (26% vs 91%) and regional citrate anticoagulation (74% vs 91%). Circuit design knowledge improved from 74% to 82%. CRRT prescribing knowledge improved from 47% to 100%. Knowledge of citrate adjustment in cirrhosis (37% to 100%) & institutional knowledge of CRRT solutions improved from 63% to 82%. AI-generated songs enhanced engagement and retention. AI-generated songs enhanced engagement and retention
Conclusion
AI-assisted educational tools significantly improved CRRT concept and prescribing knowledge. While not a replacement for traditional teaching, AI-based visual and audio tools provide an effective & engaging adjunct, offering a novel approach to nephrology education
Pre and post video results
| Category | Pre | Post |
| Indications – Hemodynamic instability & Florid volume overload | 68% | 100% |
| Clearance determinant (RFR) | 26% | 91% |
| Circuit design (citrate/calcium sequence) | 74% | 82% |
| RCA mechanism | 74% | 91% |
| CRRT ordering (RFR/Ca/citrate) | 47% | 100% |
| Citrate adjustment – Cirrhosis & Severe shock | 37% | 100% |
| Institutional CRRT solutions knowledge | 63% | 82% |