Table 3. Optimal waist-height cut-off values and measures of performance in the derivation sample.
WHtR thresholds | Methods | Sensitivity | Specificity | PPV | NPV | Youden index |
Overall | ||||||
0.465 | Youden index | 0.419 | 0.772 | 0.142 | 0.937 | 0.191 |
0.435 | CTL | 0.562 | 0.595 | 0.111 | 0.938 | 0.157 |
0.500 | 0.305 | 0.858 | 0.162 | 0.932 | 0.163 | |
Boys | ||||||
0.455 | Youden index | 0.429 | 0.884 | 0.351 | 0.914 | 0.313 |
0.425 | CTL | 0.603 | 0.677 | 0.213 | 0.921 | 0.280 |
0.500 | 0.254 | 0.940 | 0.381 | 0.896 | 0.194 | |
Girls | ||||||
0.465 | Youden index | 0.500 | 0.695 | 0.086 | 0.96 | 0.195 |
0.465 | CTL | 0.500 | 0.695 | 0.086 | 0.96 | 0.195 |
0.500 | 0.381 | 0.809 | 0.103 | 0.958 | 0.190 | |
Rural | ||||||
0.415 | Youden index | 0.818 | 0.396 | 0.129 | 0.952 | 0.214 |
0.415 | CTL | 0.818 | 0.396 | 0.129 | 0.952 | 0.214 |
0.500 | 0.182 | 0.851 | 0.118 | 0.905 | 0.033 | |
Semi-rural | ||||||
0.505 | Youden index | 0.316 | 0.871 | 0.222 | 0.916 | 0.187 |
0.455 | CTL | 0.447 | 0.711 | 0.153 | 0.917 | 0.158 |
0.500 | 0.316 | 0.849 | 0.197 | 0.914 | 0.165 | |
Urban | ||||||
0.475 | Youden index | 0.451 | 0.808 | 0.146 | 0.952 | 0.259 |
0.435 | CTL | 0.647 | 0.593 | 0.105 | 0.958 | 0.240 |
0.500 | 0.333 | 0.862 | 0.152 | 0.946 | 0.195 | |
Black | ||||||
0.465 | Youden index | 0.500 | 0.754 | 0.120 | 0.957 | 0.254 |
0455 | CTL | 0.536 | 0.699 | 0.107 | 0.957 | 0.235 |
0.500 | 0.357 | 0.832 | 0.125 | 0.951 | 0.189 | |
Coloured | ||||||
0.425 | Youden index | 0.644 | 0.563 | 0.121 | 0.944 | 0.207 |
0.425 | CTL | 0.644 | 0.563 | 0.121 | 0.944 | 0.207 |
0.500 | 0.288 | 0.877 | 0.179 | 0.930 | 0.165 | |
White | ||||||
0.475 | Youden index | 0.444 | 0.759 | 0.229 | 0.895 | 0.203 |
0.445 | CTL | 0.556 | 0.634 | 0.196 | 0.899 | 0.190 |
0.500 | 0.278 | 0.839 | 0.217 | 0.878 | 0.117 |
CTL, closest top left; NPV, negative predictive value; PPV, positive predictive value; WHtR, waist-to-height ratio.