Background
Methods
Data
Model selection, outcome metrics, cross-validation and variable importance
Performance comparison
PROMs
Minimal important differences (MIDs)
Results
Hip replacement surgery | 2015/2016 | 2016/2017 |
Observations | 30,524 | 31,905 |
Female | 18,224 (59.7%) | 19,009 (59.58%) |
Age band (years) | ||
20 to 29 | 5 (0.02%) | 0 |
30 to 39 | 22 (0.07%) | 0 |
40 to 49 | 576 (1.89%) | 457 (1.43%) |
50 to 59 | 3819 (12.51%) | 4204 (13.18%) |
60 to 69 | 10,633 (34.83%) | 10,898 (34.16%) |
70 to 79 | 11,607 (38.03%) | 12,179 (38.17%) |
80 to 89 | 3844 (12.59%) | 4130 (12.94%) |
90 to 120 | 18 (0.06%) | 37 (0.12%) |
Previous hip-replacement surgery | 2481 (8.13%) | 1587 (4.97%) |
Disability | 16,654 (54.56%) | 16,899 (52.97%) |
Mean preoperative VAS score | 64.85 (±21.94) | 64.35 (±22.30) |
Mean postoperative VAS score | 76.91 (±18.17) | 77.61 (±17.66) |
Mean preoperative Q score | 18.47 (±8.34) | 18.19 (±8.31) |
Mean postoperative Q score | 39.66 (±8.62) | 39.74 (±8.65) |
Patients with improvement (VAS) | 13,321 (43.64%) | 14,512 (45.49%) |
Patients with improvement (Q score) | 27,636 (90.54%) | 29,026 (90.98%) |
Knee replacement surgery | 2015/2016 | 2016/2017 |
Observations | 34,110 | 34,406 |
Female | 19,253 (56.44%) | 19,483 (56.63%) |
Age band (years) | ||
40 to 49 | 43 (0.13%) | 19 (0.05%) |
50 to 59 | 3368 (9.87%) | 3552 (10.32%) |
60 to 69 | 13,025 (38.19%) | 12,716 (36.96%) |
70 to 79 | 13,849 (40.60%) | 13,974 (40.62%) |
80 to 89 | 3825 (11.21%) | 4145 (12.05%) |
90 to 120 | 0 | 0 |
Previous knee-replacement surgery | 2348 (6.88%) | 1194 (3.47%) |
Disability | 17,964 (52.66%) | 17,576 (51.08%) |
Mean preoperative VAS score | 68.18 (±20.24) | 67.86 (±20.36) |
Mean postoperative VAS score | 74.27 (±18.61) | 74.81 (±18.40) |
Mean preoperative Q score | 19.34 (±7.79) | 19.32 (±7.71) |
Mean postoperative Q score | 35.60 (±9.51) | 35.83 (±9.37) |
Patients with improvement (VAS) | 11,037 (32.08%) | 11,679 (33.94%) |
Patients with improvement (Q score) | 28,657 (84.01%) | 29,096 (84.57%) |
Hip replacement surgery | Knee replacement surgery | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
VAS | Q score | VAS | Q score | |||||||||
Model
| xgbTree | msaenet | neural net | xgbTree | msaenet | glm | xgbTree | msaenet | glm | xgbTree | msaenet | glm |
Training
| ||||||||||||
AUC | 0.87 | 0.87 | 0.86 | 0.78 | 0.78 | 0.78 | 0.87 | 0.86 | 0.86 | 0.71 | 0.71 | 0.71 |
Best threshold | 0.5 | 0.45 | 0.55 | 0.5 | 0.5 | 0.5 | 0.5 | 0.45 | 0.45 | 0.5 | 0.5 | 0.5 |
Sensitivity | 0.81 | 0.79 | 0.81 | 0.78 | 0.76 | 0.76 | 0.82 | 0.79 | 0.79 | 0.70 | 0.69 | 0.69 |
Specificity | 0.76 | 0.77 | 0.76 | 0.64 | 0.67 | 0.67 | 0.73 | 0.76 | 0.76 | 0.59 | 0.61 | 0.61 |
J-statistic | 0.57 | 0.57 | 0.57 | 0.42 | 0.43 | 0.43 | 0.56 | 0.56 | 0.56 | 0.29 | 0.30 | 0.30 |
Testing
| ||||||||||||
Sensitivity | 0.82 | 0.72 | 0.84 | 0.79 | 0.78 | 0.77 | 0.83 | 0.70 | 0.71 | 0.70 | 0.70 | 0.70 |
Specificity | 0.77 | 0.85 | 0.73 | 0.63 | 0.64 | 0.65 | 0.73 | 0.83 | 0.83 | 0.61 | 0.61 | 0.62 |
Pos Pred Value | 0.75 | 0.79 | 0.72 | 0.96 | 0.96 | 0.96 | 0.62 | 0.69 | 0.68 | 0.91 | 0.91 | 0.91 |
Neg Pred Value | 0.84 | 0.78 | 0.85 | 0.23 | 0.22 | 0.22 | 0.89 | 0.85 | 0.85 | 0.27 | 0.27 | 0.27 |
F1 | 0.78 | 0.75 | 0.78 | 0.86 | 0.86 | 0.85 | 0.71 | 0.69 | 0.69 | 0.79 | 0.79 | 0.79 |
Balanced Accuracy | 0.79 | 0.78 | 0.79 | 0.71 | 0.71 | 0.71 | 0.78 | 0.77 | 0.77 | 0.66 | 0.66 | 0.66 |
J-statistic | 0.59 | 0.56 | 0.58 | 0.42 | 0.42 | 0.42 | 0.57 | 0.54 | 0.54 | 0.31 | 0.31 | 0.31 |