Table 4.
Definitions | Chi square* |
% Sensitivity |
% Specificity |
% False Neg |
% False Pos |
AUC | Kappa | Rank# | Final Score# |
---|---|---|---|---|---|---|---|---|---|
3 of any 6 improved by at least 20%, no more than 1 worsened by more than 30% which cannot be muscle strength |
90.3 | 98 | 87 | 9 | 3 | 92 | 0.86 | 131 | 113 |
3 of any 6 improved by at least 20%, no more than 2 worsened by ≥ 25%, which cannot be muscle strength (IMACS definition P1) (13) |
90.3 | 98 | 87 | 9 | 3 | 92 | 0.86 | 104 | 90 |
3 of any 6 improved by at least 20%, no more than 2 worsened by more than 30% which cannot be muscle strength |
90.3 | 98 | 87 | 9 | 3 | 92 | 0.86 | 81 | 70 |
2 of any 6 improved by at least 40%, no more than 1 worsened by more than 30% which cannot be muscle strength |
85.2 | 97 | 87 | 13 | 3 | 92 | 0.84 | 61 | 51 |
2 of any 6 improved by at least 30%, no more than 1 worsened by more than 30% which cannot be muscle strength |
90.1 | 100 | 78 | 0 | 5 | 89 | 0.85 | 46 | 39 |
3 of any 6 improved by at least 20%, no more than 1 worsened by more than 30% |
84.3 | 98 | 83 | 10 | 4 | 90 | 0.83 | 36 | 30 |
3 of any 6 improved by at least 20%, no more than 2 worsened by more than 30% |
84.3 | 98 | 83 | 10 | 4 | 90 | 0.83 | 17 | 14 |
3 of any 6 improved by at least 20%, no more than 2 worsened by ≥ 25% (IMACS definition P2) (13) |
84.3 | 98 | 83 | 10 | 4 | 90 | 0.83 | 13 | 11 |
2 of any 6 improved by at least 40%, no more than 1 worsened by more than 30% |
79.2 | 97 | 83 | 14 | 4 | 90 | 0.81 | 13 | 11 |
2 of any 6 improved by at least 40%, no more than 2 worsened by more than 30%, which cannot be muscle strength |
79.2 | 97 | 83 | 14 | 4 | 90 | 0.81 | 13 | 11 |
2 of any 6 improved by at least 30%, no more than 2 worsened by more than 30%, which cannot be muscle strength |
84.3 | 100 | 74 | 0 | 6 | 87 | 0.82 | 6 | 5 |
3 of any 6 improved by at least 20% (IMACS definition P3) (13) |
84.3 | 98 | 83 | 10 | 4 | 90 | 0.83 | 3 | 3 |
2 of any 6 improved by at least 30%, no more than 1 worsened by more than 30% |
84.3 | 100 | 74 | 0 | 6 | 87 | 0.82 | 1 | 1 |
All chi-squares correspond to a p value < 0.0001
The ranks were obtained by asking the attendees of the consensus meeting to decide upon which of the definitions of improvement that performed best were easiest to use and most credible (content validity). Than for each definition, the content validity rankings obtained were summed up and the resulting sum was multiplied by the corresponding value of the kappa statistic, to obtain the “final score” that incorporated both statistical criteria and experts’ judgments.”