Unlocking Trauma Data’s Potential
Algorithms for Early Prediction
In Challenge Event 1, teams were provided real-world trauma center datasets containing thousands of cases with >500 variables. This data was collected from pre-hospital (medevac, ambulance) until 4 hours after hospital admission. Teams developed algorithms to predict the need for life-saving interventions during this time.
In Challenge Event 2, the data size will significantly increase in size. Additional data types will add to the training data’s complexity. To win a prize, teams’ performance must exceed the specified baseline algorithm as well as specificity, sensitivity and lead time thresholds
Leaderboard
Scoring criteria: Accuracy of LSI prediction (for any LSI and for more specific classifications, such as hemorrhage or airway interventions)
| Teams | Score | 1 | 2 |
|---|---|---|---|
| Coordinated Robotics* | 625 | 434 | 191 |
| MSAI* | 555 | 364 | 191 |
| CAMA* | 532 | 368 | 164 |
| LENS | 310 | 129 | 181 |
| AI TEMPO | 268 | 91 | 177 |
| CRITIC | 189 | 21 | 168 |
| TrueFit.AI* | 179 | 37 | 142 |
| Robotika* | 143 | 27 | 116 |
| ALICE | 109 | 18 | 91 |
| AUSTERE | 61 | 49 | 12 |
| MGB-Harvard* | 44 | 13 | 31 |
| CNA* | 35 | 8 | 27 |
* Self-funded
Overview
