Table 3.
Performance of the new regression equations and ANN models in the external validation data set
| Bias (ml/min/1.73 m2) | Precision (ml/min/1.73 m2) | Accuracy (%) | |
|---|---|---|---|
| CKD-EPI | −6.00 | 24.20 | 80.0 |
| Japanese equation 1 | −20.48* | 22.43* | 54.3* |
| Japanese equation 2 | −30.67* | 23.55* | 29.5* |
| New equation 1 | −2.49 | 22.50* | 84.8 |
| ANN1 | −2.70 | 21.37* | 87.6 |
| New equation 2 | −2.88 | 22.07* | 84.8 |
| ANN2 | −4.87 | 20.70* | 83.8 |
| New equation 3 | −3.97 | 21.22* | 80.0 |
| ANN3 | −5.97 | 20.49* | 88.6‡ |
| New equation 4 | −2.23 | 21.11* | 84.8 |
| ANN4 | −3.08 | 19.96* | 84.8 |
| New equation 5 | −3.97 | 21.46* | 80.0 |
| ANN5 | −6.81 | 22.93* | 85.7 |
| New equation 6 | −3.19 | 21.64* | 83.8 |
| ANN6 | −5.31 | 25.87* | 88.6 |
| New equation 7 | −2.91 | 21.35* | 81.9 |
| ANN7 | −5.97 | 22.51* | 87.6 |
| New equation 8 | −4.48 | 21.89* | 81.9 |
| ANN8 | −5.73 | 22.30* | 87.6 |
CKD-EPI equation Chronic Kidney Disease Epidemiology Collaboration, ANN Artificial neural network
* P < 0.001, ‡P < 0.05 comparing with the CKD-EPI equation