Skip to main content

Table 3 Characteristics of patients included in model derivation and external validation

From: External validation of inpatient neonatal mortality prediction models in high-mortality settings

Indicator

Levels

Derivation

Temporal validationa

Model updating (recalibration)b

External validation

SENSSc, n = 5427

NETSd, n = 4840

SENSSc, n = 1627

NETSd, n = 1443

SENSSc, n = 8848

NETSd, n = 6610

SENSSc, n = 53,909

NETSd, n = 45,090

n (%)

n (%)

n (%)

n (%)

n (%)

n (%)

n (%)

n (%)

Male

Yes

2937 (54.12)

2605 (53.82)

961 (59.07)

850 (58.91)

4963 (56.09)

3713 (56.17)

29,384 (54.51)

24,719 (54.82)

Missing

13 (0.24)

12 (0.25)

2 (0.12)

2 (0.14)

27 (0.31)

25 (0.38)

642 (1.19)

468 (1.04)

Weight

1000 g and below (ELBW)

32 (0.59)

31 (0.64)

10 (0.61)

10 (0.69)

104 (1.18)

96 (1.45)

1243 (2.31)

1083 (2.4)

1001–1499 g (VLBW)

136 (2.51)

115 (2.38)

45 (2.77)

40 (2.77)

379 (4.28)

366 (5.54)

3844 (7.13)

3494 (7.75)

1500–2499 g (LBW)

1180 (21.74)

1043 (21.55)

361 (22.19)

316 (21.9)

2123 (23.99)

1717 (25.98)

13,207 (24.5)

11,320 (25.11)

2500–4000 g (NBW)

3841 (70.78)

3438 (71.03)

1125 (69.15)

1002 (69.44)

5834 (65.94)

4162 (62.97)

31,285 (58.03)

25,932 (57.51)

 > 4000 g (macrosomia)

229 (4.22)

204 (4.21)

85 (5.22)

74 (5.13)

368 (4.16)

243 (3.68)

3352 (6.22)

2478 (5.5)

Missing

9 (0.17)

9 (0.19)

1 (0.06)

1 (0.07)

40 (0.45)

26 (0.39)

978 (1.81)

783 (1.74)

Mode of delivery

Breech

43 (0.79)

40 (0.83)

23 (1.41)

19 (1.32)

197 (2.23)

165 (2.5)

1226 (2.27)

1028 (2.28)

Caesarean section (C/S)

2212 (40.76)

1957 (40.43)

574 (35.28)

509 (35.27)

3211 (36.29)

2251 (34.05)

18,634 (34.57)

15,540 (34.46)

Spontaneous vaginal (SVD)e

3014 (55.54)

2698 (55.74)

1012 (62.2)

897 (62.16)

5157 (58.28)

4006 (60.61)

33,203 (61.59)

27,824 (61.71)

Missing

158 (2.91)

145 (3)

18 (1.11)

18 (1.25)

283 (3.2)

188 (2.84)

846 (1.57)

698 (1.55)

Outbornf

Yes

123 (2.27)

107 (2.21)

60 (3.69)

57 (3.95)

495 (5.59)

439 (6.64)

6017 (11.16)

5155 (11.43)

Missing

0 (0)

0 (0)

0 (0)

0 (0)

0 (0)

0 (0)

0 (0)

0 (0)

Apgar score (5 min)

0–3

116 (2.14)

112 (2.31)

33 (2.03)

33 (2.29)

293 (3.31)

274 (4.15)

1146 (2.13)

1005 (2.23)

4–6

602 (11.09)

593 (12.25)

200 (12.29)

199 (13.79)

1334 (15.08)

1184 (17.91)

9121 (16.92)

7986 (17.71)

7–10

3992 (73.56)

3918 (80.95)

1149 (70.62)

1142 (79.14)

6484 (73.28)

4868 (73.65)

38,391 (71.21)

31,531 (69.93)

Missing

717 (13.21)

217 (4.48)

245 (15.06)

69 (4.78)

737 (8.33)

284 (4.3)

5251 (9.74)

4568 (10.13)

HIV exposure

Exposed

319 (5.88)

287 (5.93)

84 (5.16)

74 (5.13)

473 (5.35)

439 (6.64)

2145 (3.98)

1957 (4.34)

Missing

305 (5.62)

277 (5.72)

94 (5.78)

80 (5.54)

276 (3.12)

219 (3.31)

4742 (8.8)

3826 (8.49)

Outcome

Aliveg

4900 (90.29)

4374 (90.37)

1469 (90.29)

1299 (90.02)

8134 (91.93)

5950 (90.02)

46,358 (85.99)

38,576 (85.55)

Dead

508 (9.36)

447 (9.24)

152 (9.34)

138 (9.56)

696 (7.87)

649 (9.82)

7486 (13.89)

6482 (14.38)

Missing

19 (0.35)

19 (0.39)

6 (0.37)

6 (0.42)

18 (0.2)

11 (0.17)

65 (0.12)

32 (0.07)

  1. aThe same hospital was used at the model derivation and temporal validation stage, with the temporal validation stage using data from a specific future period
  2. bData is only from the same hospital used at derivation and temporal validation stage. Data collected between January 2016 and December 2020
  3. cData presented are before multiple imputation. The multiple imputation filled in the missing values while preserving the pattern of distribution observed in the original datasets
  4. dAll cases included in NETS model are subset of those included in SENSS model but with a treatment sheet present; given the models are developed independently of each other, there is no substantive implication on interpretation of findings
  5. eIncludes assisted vaginal deliveries (e.g. forceps, vacuum)
  6. fOutborn refers to neonates admitted to the unit having been born either in another facility, at home or on the way to hospital
  7. gPatients referred out of hospital recoded were also treated as being ‘alive’ at discharge
Back to article page

BMC Medicine

ISSN: 1741-7015

Contact us