Patient Readmission Risk Index (SXI)
SCORE + CORRELATE = IMPROVE
Current SXI Score
1.16
Current Readmissions Rate
46.10%
Readmissions Prediction Accuracy
99.99%
Target SXI Score
1.19
Target Readmissions Rate
(20%) 36.88%
Readmissions Prediction Precision
99.99%
Potential Healthcare Improvements
Current Number of Patient Readmissions = 46,902
Improvement
20% or 9380 less
Patient Readmissions
Improvement
35% or 16,416 less
Patient Readmissions
Improvement
57.30% or 26,875 less
Patient Readmissions
| Auto-ML | SXI | SXI++ | |
|---|---|---|---|
| Accuracy (%) | 62.62% | 98.8% | 99.99% |
| Precision (%) | 61.87% | 98% | 99.99% |
| AUC | 0.67 | 0.98 | 1 |
| True Positives = 11003 | False Positives = 2 |
|---|---|
| False Negatives = 0 | True Negatives = 9448 |
| Patient Number | SXI Score | Actual Target | Predicted Target |
|---|---|---|---|
| 59286361 | 1.157 | Not Readmitted | Readmitted |
| 97811968 | 0.345 | Readmitted | Readmitted |
| 22839136 | 0.654 | Readmitted | Readmitted |
| 85048098 | 1.654 | Not Readmitted | Not Readmitted |
| 81725289 | 1.614 | Not Readmitted | Not Readmitted |
| 73834604 | 1.159 | Not Readmitted | Readmitted |
| 13042607 | 1.373 | Not Readmitted | Not Readmitted |
| 63627495 | 1.21 | Not Readmitted | Not Readmitted |
| 54981158 | 0.585 | Readmitted | Readmitted |
| 20783153 | 0.711 | Readmitted | Readmitted |
Current Readmissions Decision Tree
Target Readmissions Decision Tree
Sensitivity Analysis
| FEATURES | DIAG_2 | DIAG_1 | DIAG_3 | NUM_LAB_PROCEDURES | NUM_MEDICATIONS | PREDICTION |
|---|---|---|---|---|---|---|
| Actual Values | 440.0 | 250.7 | 496.0 | 63.0 | 30.0 | 0 (Readmitted) |
| Adjusted Values | 660 | 200.56 | 719.2 | 44.1 | 37.2 | 1 (Not Readmitted) |
| Change in values (%) | 50 | -20 | 45 | -30 | 24 | - |
The target variable changed from Patient Readmitted (Prediction in Actual Data) to Not Readmitted (Prediction in Adjusted Data).
Race
Most Common : Caucasian
Gender
Most Common: Female
Glipizide
Most Common: Glipizide_Down