Figure 1. Final CABG Algorithm
Figure 2. Correlation and Cross-validation error for CABG Algorithm
Figure 3. Final Algorithm for Herniorrhaphy
Figure 4. Correlation and Cross Validation Error for Herniorrhaphy
Figure 5. Unpruned Algorithm for TKA
Figure 6. Correlation and Cross Validation Error for TKA
Figure 7. Final Algorithm for TKA
Figure 8. Final Algorithm for THA
Figure 9. Correlation and Cross Validation Error for THA
Algorithms That Include sSSI
Rules using similar protocols as that above that included superficial SSI for training did not perform as well.
Table 1. Preliminary Performance on Training Set (includes sSSI)
CABG | HIP | KNEE | HERNIA | |
---|---|---|---|---|
Total | 4525 | 19692 | 7467 | 3867 |
Algorithm Positive | 747 | 2238 | 1980 | 311 |
True Positive | 116 | 220 | 103 | 78 |
False Negative | 28 | 19 | 6 | 17 |
Algorithm Sensitivity | 80.6 | 92.1 | 94.4 | 82.1 |
% Excluded | 83.5 | 88.6 | 73.5 | 91.2 |
NNR Algorithm | 6.4 | 10.2 | 19.2 | 4 |
NNR sans Algorithm | 39 | 89.5 | 72.5 | 49.6 |
CABG - Coronary Artery Bypass Grafting, HERNIA - herniorraphy, KNEE - Total Knee Arthroplasty, HIP - Total Hip Arthroplasty, NNR - number needed to read
Figure 10. Algorithm for CABG (includes sSSI)
Figure 11. Algorithm for Herniorrhaphy (Includes sSSI)
Figure 12. Algorithm for TKA (Includes sSSI)
Figure 13. Algorithm for THA (Includes sSSI)