AMPEL explores and applies different models, starting with the analysis of past data (retrospective). Theoretical models are tested in practice, and a randomized controlled study design ensures a rigorous comparison between patients with and without "AI support." If ongoing data shows an advantage with intervention (prospective), the models are integrated into routine care so that all patients can benefit from them.
Overview of the different models within AMPEL:
| Model | Description | Status |
|
Refeeding Syndrome | Detection of refeeding syndrome and initiation of consequences. | Routine Care |
| Acute Kidney Injury | Detection of acute kidney injury in general and under vancomycin therapy. | Randomized Study |
| Anemia | Detection of transfusion-worthy anemias and monitoring of the subsequent transfusion | Randomized Study |
|
Hypercalcemia | Detection and ensuring a consequence (follow-up measurement) for hypercalcemia. | Randomized Study |
| Hypokalemia | Detection and ensuring a consequence (follow-up measurement) for hypokalemia. | Randomized Study |
|
Hypocalcemia | Detection and ensuring a consequence (follow-up measurement) for hypocalcemia. | Randomized Study |
| Hyponatremia | Detection and ensuring a consequence (follow-up measurement) for hyponatremia. | Randomized Study |
| Lactic Acidosis | Detection of an infectious event and ensuring a consequence (follow-up measurement). | Randomized Study |
| Diabetes mellitus | Detection of previously uncoded diabetes mellitus. | Development |
| Fibrinogen | Detection of unknown fibrinogen deficiency. | Development |
| Hyperglycemia | Detection of persistent hyperglycemia or those without consequence (follow-up measurements). | Development |
| Procalcitonin | Detection of an inadequate procalcitonin decrease. | Development |
|
Sepsis | Detection of a potential sepsis constellation. | Development |
>Refeeding Syndrome Model
