Table 2.
sPTB | No. sPTB (%) |
AUC [95% CI] |
Cervical length | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0–15 mm (n = 14; 0.22%) | 0–20 mm (n = 71; 1.1%) | 0–25 mm (n = 274; 4.4%) | Best cut‐off* | ||||||||||||
Sensitivity [95% CI] |
Specificity [95% CI] |
FP/TP; NNS |
Sensitivity [95% CI] |
Specificity [95% CI] |
FP/TP; NNS |
Sensitivity [95% CI] |
Specificity [95% CI] |
FP/TP; NNS |
Sensitivity [95% CI] |
Specificity [95% CI] |
FP/TP; NNS |
mm | |||
<28 GW |
3 (0.05) |
0.96 [0.92–1.00] |
1/3 (33.3%) [0.8–90.6] |
6272/6285 (99.8%) [99.7–99.9] |
13; 6288 |
1/3 (33.3%) [0.8–90.6] |
6215/6285 (98.9%) [98.6–99.1] |
70; 6288 |
1/3 (33.3%) [0.8–90.6] |
6012/6285 (95.7%) [95.1–96.2] |
273; 6288 |
3/3 (100%) [29.2–100] |
5778/6285 (91.9%) [91.2–92.6] |
169; 2096 |
27 (n = 510; 8.1%) |
<29 GW |
5 (0.08) |
0.98 [0.95–1.00] |
3/5 (60.0%) [14.7–94.7] |
6272/6283 (99.8%) [99.7–99.9] |
4; 2096 |
3/5 (60.0%) [14.7–94.7] |
6215/6283 (98.9%) [98.6–99.2] |
23; 2096 |
3/5 (60.0%) [14.7–94.7] |
6012/6283 (95.7%) [95.2–96.2] |
90; 2096 |
5/5 (100%) [47.8–100] |
5778/6283 (92.0%) [91.3–92.6] |
101; 1258 |
27 (n = 510; 8.1%) |
<30 GW |
10 (0.16) |
0.86 [0.69–1.00] |
5/10 (50.0%) [18.7–81.3] |
6269/6278 (99.9%) [99.7–99.9] |
2; 1257 |
6/10 (60.0%) [26.2–87.8] |
6213/6278 (99.0%) [98.7–99.2] |
11; 1048 |
6/10 (60.0%) [26.3–87.8] |
6010/6278 (95.7%) [95.2–96.2] |
45; 1048 |
8/10 (80.0%) [44.4–97.5] |
5776/6278 (92.0%) [91.3–92.7] |
63; 786 |
27 (n = 510; 8.1%) |
<31 GW |
15 (0.24) |
0.85 [0.72–0.99] |
6/15 (40.0%) [16.3–67.7] |
6265/6273 (99.9%) [99.8–99.9] |
1; 1048 |
8/15 (53.3%) [26.6–78.7] |
6210/6273 (99.0%) [98.7–99.2] |
8; 786 |
8/15 (53.3%) [26.6–78.7] |
6007/6273 (95.8%) [95.2–96.2] |
33; 786 |
11/15 (73.3%) [44.9–92.2] |
5774/6273 (92.0%) [91.4–92.7] |
45; 572 |
27 (n = 510; 8.1%) |
<32 GW |
18 (0.29) |
0.81 [0.68–0.93] |
6/18 (33.3%) [13.3–59.0] |
6262/6270 (99.9%) [99.8–99.9] |
1; 1048 |
8/18 (44.4%) [21.5–69.2] |
6207/6270 (99.0%) [98.7–99.2] |
8; 786 |
8/18 (44.4%) [21.5–69.2] |
6004/6270 (95.8%) [95.2–96.2] |
33; 786 |
11/18 (61.1%) [35.8–82.7] |
5771/6270 (92.0%) [91.3–92.7] |
45; 572 |
27 (n = 510; 8.1%) |
<33 GW |
26 (0.41) |
0.76 [0.65–0.87] |
6/26 (23.1%) [9.0–43.7] |
6254/6262 (99.9%) [99.8–99.9] |
1; 1048 |
9/26 (34.6%) [17.2–55.7] |
6200/6262 (99.0%) [98.7–99.2] |
7; 699 |
10/26 (38.5%) [20.3–59.4] |
5998/6262 (95.8%) [95.3–96.3] |
26; 629 |
14/26 (53.8%) [33.4–73.4] |
5766/6262 (92.1%) [91.4–92.7] |
35; 449 |
27 (n = 510; 8.1%) |
<34 GW |
41 (0.65) |
0.71 [0.63–0.80] |
6/41 (14.6%) [5.6–29.2] |
6239/6247 (99.9%) [99.8–99.9] |
1; 1048 |
10/41 (24.4%) [12.4–40.3] |
6186/6247 (99.0%) [98.8–99.3] |
6; 629 |
12/41 (29.3%) [16.1–45.5] |
5985/6247 (95.8%) [95.3–96.3] |
22; 524 |
16/41 (39.0%) [24.2–55.5] |
5753/6247 (92.1%) [91.4–92.8] |
31; 393 |
27 (n = 510; 8.1%) |
<35 GW |
69 (1.10) |
0.71 [0.65–0.77] |
7/69 (10.1%) [4.2–19.8] |
6212/6219 (99.9%) [99.8–100.0] |
1; 898 |
13/69 (18.8%) [10.4–30.1] |
6161/6219 (99.1%) [98.8–99.3] |
5; 484 |
17/69 (24.6%) [15.1–36.5] |
5962/6219 (95.9%) [95.3–96.4] |
15; 370 |
53/69 (76.8%) [65.1–86.1] |
3346/6219 (53.8%) [52.6–55.1] |
54; 119 |
35 (n = 2926; 46.5%) |
<36 GW | 114 (1.81) |
0.67 [0.61–0.72] |
7/114 (6.1%) [2.5–12.2] |
6167/6174 (99.9%) [99.8–100.0] |
1; 898 |
15/114 (13.2%) [7.6–20.8] |
6118/6174 (99.1%) [98.8–99.3] |
4; 419 |
23/114 (20.2%) [13.2–28.7] |
5923/6174 (95.9%) [95.4–96.4] |
11; 273 |
81/114 (71.1%) [61.8–79.2] |
3329/6174 (53.9%) [52.7–55.2] |
35; 78 |
35 (n = 2926; 46.5%) |
<37 GW | 225 (3.58) |
0.63 [0.59–0.67] |
7/225 (3.1%) [1.3–6.3] |
6056/6063 (99.9%) [99.8–100.0] |
1; 898 |
18/225 (8.0%) [4.8–12.4] |
6010/6063 (99.1%) [98.9–99.3] |
3; 349 |
34/225 (15.1%) [10.7–20.5] |
5823/6063 (96.0%) [95.5–96.5] |
7; 185 |
144/225 (64.0%) [57.4–70.3] |
3281/6063 (54.1%) [52.9–55.4] |
19; 44 |
35 (n = 2926; 46.5%) |
Positive and negative predictive values and positive and negative likelihood ratios are shown in Table S9.
GW, gestational weeks; sPTB, spontaneous preterm birth; AUC, area under receiver operating characteristic curve; CI, confidence interval; FP, false positive; TP, true positive; NNS, number needed to screen, i.e. number of women needed to screen to detect one true positive test result.
Best cut‐off is the cut‐off associated with the largest number of correctly classified cases, calculated using Youden’s index. 21