Table 4.
Predictive performance measures of the model of Syngelaki et al. [20] and effect of an intervention on preeclampsia incidence for different risk thresholds
| Risk thresholda | High risk, % (n/n) | Sensitivity, % (n/n) | Specificity, % (n/n) | PPV, % (n/n) | NPV, % (n/n) | PE incidence reduction, % (NNP,n) Effect interventionb |
||
|---|---|---|---|---|---|---|---|---|
| RR 0.9 | RR 0.8 | RR 0.6 | ||||||
| 1.0% | 76.2 (1,993/2,614) | 98.7 (75/76) | 24.4 (620/2,538) | 3.8 (75/1,993) | 99.8 (620/621) | 9.9 (266) | 19.7 (133) | 39.5 (66) |
| 2.0% | 55.8 (1,459/2,614) | 94.7 (72/76) | 45.4 (1,151/2,538) | 4.9 (72/1,459) | 99.7 (1,151/1,155) | 9.5 (203) | 18.9(101) | 37.9 (51) |
| 3.0% | 37.3 (974/2,614) | 76.3 (58/76) | 63.9 (1,622/2,538) | 6.0 (58/974) | 98.9 (1,622/1,640) | 7.6 (168) | 15.3 (84) | 30.5 (42) |
| 4.0% | 20.2 (528/2,614) | 53.9 (41/76) | 80.8 (2,051/2,538) | 7.8 (41/528) | 98.3 (2,051/2,086) | 5.4 (129) | 10.8 (64) | 21.6 (32) |
| 5.0% | 12.4 (323/2,614) | 38.2 (29/76) | 88.4 (2,244/2,538) | 9.0 (29/323) | 97.9 (2,244/2,291) | 3.8 (111) | 7.6 (56) | 15.3 (28) |
| 6.0% | 8.6 (224/2,614) | 17.1 (13/76) | 91.7 (2,327/2,538) | 5.8 (13/224) | 97.4 (2,327/2,390) | 1.7(172) | 3.4 (86) | 6.8 (43) |
NNP, number needed to prevent; NPV, negative predictive value; PE, preeclampsia; PPV, positive predictive value; RR, risk reduction.
a
Predicted risk at or above this level was considered as high risk.
b
Intervention in all high-risk women.