Multiple regression model coefficients and performance statistics for previous AST model
| Parameter . | Parameter-wise shrunken estimate . | Standard error of shrunken estimate . | Shrinkage factor . | P value of parameter . |
|---|---|---|---|---|
| Intercept | −116.5 | 14.87 | 0.994 | 1.27 × 10−09 |
| β1 | 251.5 | 15.34 | 0.997 | 2.86 × 10−44 |
| β2 | −49.4 | 7.11 | 0.976 | 6.97 × 10−07 |
| β3 | 714.3 | 48.51 | 0.993 | 1.88 × 10−25 |
| Performance measure | Development dataset | Validation dataset | ||
| R2 | 0.866 (0.836-0.896) | 0.711 (0.492-0.930) | ||
| Root mean squared error | 73.4 (63.5-83.2) | 109.6 (54.3-164.8) | ||
| Calibration slope | 1.00 (0.95-1.06) | 0.92 (0.84-0.99) | ||
| Calibration intercept | −0.9 (−21.1 to 19.2) | 37.5 (7.0-68.1) | ||
| AUC at pass threshold of 430 nmol/L | 0.944 (0.919-0.969) | 0.969 (0.937-0.998) | ||
| Parameter . | Parameter-wise shrunken estimate . | Standard error of shrunken estimate . | Shrinkage factor . | P value of parameter . |
|---|---|---|---|---|
| Intercept | −116.5 | 14.87 | 0.994 | 1.27 × 10−09 |
| β1 | 251.5 | 15.34 | 0.997 | 2.86 × 10−44 |
| β2 | −49.4 | 7.11 | 0.976 | 6.97 × 10−07 |
| β3 | 714.3 | 48.51 | 0.993 | 1.88 × 10−25 |
| Performance measure | Development dataset | Validation dataset | ||
| R2 | 0.866 (0.836-0.896) | 0.711 (0.492-0.930) | ||
| Root mean squared error | 73.4 (63.5-83.2) | 109.6 (54.3-164.8) | ||
| Calibration slope | 1.00 (0.95-1.06) | 0.92 (0.84-0.99) | ||
| Calibration intercept | −0.9 (−21.1 to 19.2) | 37.5 (7.0-68.1) | ||
| AUC at pass threshold of 430 nmol/L | 0.944 (0.919-0.969) | 0.969 (0.937-0.998) | ||
Numbers in brackets correspond to 95% CI. Parameters correspond to previous AST model: 30-minute cortisol = Intercept + β1 × √(new baseline cortisol/100) + β2×(previous baseline cortisol/100) + β3×(previous 30-minute cortisol/1000).
Abbreviation: AUC, area under receiver operated characteristics curve.
Multiple regression model coefficients and performance statistics for previous AST model
| Parameter . | Parameter-wise shrunken estimate . | Standard error of shrunken estimate . | Shrinkage factor . | P value of parameter . |
|---|---|---|---|---|
| Intercept | −116.5 | 14.87 | 0.994 | 1.27 × 10−09 |
| β1 | 251.5 | 15.34 | 0.997 | 2.86 × 10−44 |
| β2 | −49.4 | 7.11 | 0.976 | 6.97 × 10−07 |
| β3 | 714.3 | 48.51 | 0.993 | 1.88 × 10−25 |
| Performance measure | Development dataset | Validation dataset | ||
| R2 | 0.866 (0.836-0.896) | 0.711 (0.492-0.930) | ||
| Root mean squared error | 73.4 (63.5-83.2) | 109.6 (54.3-164.8) | ||
| Calibration slope | 1.00 (0.95-1.06) | 0.92 (0.84-0.99) | ||
| Calibration intercept | −0.9 (−21.1 to 19.2) | 37.5 (7.0-68.1) | ||
| AUC at pass threshold of 430 nmol/L | 0.944 (0.919-0.969) | 0.969 (0.937-0.998) | ||
| Parameter . | Parameter-wise shrunken estimate . | Standard error of shrunken estimate . | Shrinkage factor . | P value of parameter . |
|---|---|---|---|---|
| Intercept | −116.5 | 14.87 | 0.994 | 1.27 × 10−09 |
| β1 | 251.5 | 15.34 | 0.997 | 2.86 × 10−44 |
| β2 | −49.4 | 7.11 | 0.976 | 6.97 × 10−07 |
| β3 | 714.3 | 48.51 | 0.993 | 1.88 × 10−25 |
| Performance measure | Development dataset | Validation dataset | ||
| R2 | 0.866 (0.836-0.896) | 0.711 (0.492-0.930) | ||
| Root mean squared error | 73.4 (63.5-83.2) | 109.6 (54.3-164.8) | ||
| Calibration slope | 1.00 (0.95-1.06) | 0.92 (0.84-0.99) | ||
| Calibration intercept | −0.9 (−21.1 to 19.2) | 37.5 (7.0-68.1) | ||
| AUC at pass threshold of 430 nmol/L | 0.944 (0.919-0.969) | 0.969 (0.937-0.998) | ||
Numbers in brackets correspond to 95% CI. Parameters correspond to previous AST model: 30-minute cortisol = Intercept + β1 × √(new baseline cortisol/100) + β2×(previous baseline cortisol/100) + β3×(previous 30-minute cortisol/1000).
Abbreviation: AUC, area under receiver operated characteristics curve.
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
