CHAPTER 5: Classifier Design
5.4 Classification Results and Statistics
Table III: Maternal socio-demographic characteristics vs preterm delivery
Parameter Babies studied OR (95% CI) p value
Age (years)
<20 20-34
≥35
Preterm (%) 5(16.7) 137(10.9)
43(18.9)
Term (%) 25(83.3) 1125(89.1)
185(81.1)
1.64 (0.48-4.46)$ 1.91 (1.29-2.82)&
0.002*
S
Body Mass Index (Kg/m2) Low(<19.9)
High(≥19.9)
5(15.2) 15(7.8)
28(84.8) 178(92.2)
2.12 (0.56-6.75) 0.184#
NS Social Class
Low(III,IV,V) High(I and II)
103(15.7) 80(9.3)
551(84.3) 780(90.7)
1.82 (1.32-2.52) 0.000*
S Marital Status
Single Married
5(15.2) 180(12.1)
28(84.8) 1309(87.9)
1.30 (0.39-3.47) 0.588#
NS Parity
0 1-4
≥5
5(7.7) 154(11.5)
26(23.0)
60(92.3) 1184(88.5)
87(77.0)
0.64 (0.20-1.62)^
2.30 (1.40-3.76)@
0.001*
S
* Chi square test
# Fisher’s’s Exact test S Significant
NS not significant
$ Odds ratio and 95% confidence interval for low maternal age
& Odds ratio and 95% confidence interval for high maternal age
^ Odds ratio and 95% confidence interval for primiparity
@ Odds ratio and 95% confidence interval for high parity
Relationship between maternal antenatal factors and preterm delivery
A number of antenatal factors were evaluated to determine their relationship to preterm deliveries. Those characteristics significant as determinants of preterm delivery were previous preterm delivery, febrile illness, antepartum haemorrhage, prolonged rupture of membrane and premature rupture of membrane. Others include urinary tract infection chorioamnionitis, pregnancy induced hypertension, type of labour and booking status as shown in Table IV.
Table IV: Maternal antenatal and delivery characteristics vs preterm delivery
Parameter Babies studied OR (95% CI) p value
Packed cell volume (%) Preterm (%) Term (%)
<30
≥30
19(8.2) 62(7.6)
212(91.8) 756(92.4)
1.09 (0.62-1.92) 0.745*
NS Previous abortion
Yes No
53(13.2) 131(11.7)
350(86.8) 987(88.3)
1.14 (0.80-1.63) 0.449*
NS Previous uterine curettage
Yes No
31(12.9) 153(11.9)
209(87.1) 1128(88.1)
1.09 (0.71-1.68) 0.671*
NS Previous preterm delivery
Yes No
13(31.7) 171(11.6)
28(68.3) 1309(88.4)
3.55 (1.71-7.30) 0.001#
S Febrile illness
Yes No
11(22.9) 174(11.8)
37(77.1) 1300(88.2)
2.22 (1.05-4.62) 0.020*
S Antepartum haemorrhage
Yes No
16(53.3) 169(11.3)
14(46.7) 1323(88.7)
8.95 (4.06-19.78) 0.000#
S Prolonged rupture of membrane
Yes No
43(43.0) 142(10.0)
57(57.0) 1280(90.0)
6.80 (4.32-10.71) 0.000*
S Premature rupture of membrane
Yes No
56(40.0) 129(9.3)
84(60.0) 1253(90.7)
6.48 (4.33-9.67) 0.000*
S Urinary tract infection
Yes No
4(44.4) 181(12.0)
5(55.6) 1332(88.0)
5.89 (1.16-27.57) 0.016#
S Chorioamnionitis
Yes No
6(42.9) 179(11.9)
8(57.1) 1329(88.1)
5.57 (1.57-18.50) 0.004#
S Pregnancy induced hypertension
Yes No
37(27.8) 148(10.7)
96(72.2) 1241(89.3)
3.23 (2.09-4.99) 0.000*
S Type of labour
Medically induced Spontaneous
66(38.4) 119(8.8)
106(61.6) 1231(91.2)
6.44 (4.42-9.38) 0.000*
S Booking status
Elsewhere UITH
84(29.4) 101(8.2)
202(70.6) 1135(91.8)
4.67 (3.33-6.56) 0.000*
S
* Chi square test
# Fisher’s Exact test
S Significant NS not significant
Analysis to exclude confounders of the determinants of preterm delivery
To exclude confounders, factors identified above to be significantly associated with preterm delivery were included in a linear regression model. A logistic regression analysis was thereafter carried out to exclude their relative contribution listwise. Table V shows that after regression analysis, socio-economic class, previous preterm delivery, antepartum haemorrhage, premature rupture of fetal membranes, urinary tract infection, pregnancy induced hypertension, type of labour and booking status remained significant. The model accounted for 19.6% of preterm deliveries.
The goodness of fit for the model gave an F-value of 28.047; df: 13, p=0.000.
Booking status was the strongest determinant of preterm delivery.
Table V: Linear logistic regression of maternal socio-demographic and antenatal variables on preterm deliveries
Beta Coefficients t p value
Maternal age -.033 -1.360 .174
Socio-economic class -.057 -2.412 .016
Parity -.029 -1.188 .235
previous preterm delivery .081 3.488 .001
Febrile illness .026 1.017 .309
antepartum haemorrhage .102 4.240 .000
prolonged rupture of membranes .051 1.285 .199
premature rupture of membranes .153 3.931 .000
urinary tract infection .066 2.746 .006
Chorioamnionitis .013 .480 .632
hypertension in pregnancy .068 2.702 .007
Type of labour -.167 -6.408 .000
booking status -.172 -7.144 .000
Comparison between determinants of early versus late preterm deliveries
Maternal socio-demographic variables were compared between early and late preterm deliveries. Table VI illustrates that none of the factors studied was significantly associated with timing of preterm delivery.
Table VI: maternal socio-demographic variables vs early and late preterm deliveries
Parameter Babies studied OR (95% CI) P value
Age (years)
<20 20-34
≥35
Early preterm (%) 0
38(27.9) 12(27.9)
Late preterm (%) 5(100.0) 98(72.1) 31(72.1)
0.00 (0.00-2.95)$ 1.00 (0.43-2.28)&
0.383*
NS
Body Mass Index (Kg/m2) Low(<19.9)
High(≥19.9)
3(25.0) 11(12.5)
9(75.0) 77(87.5)
2.33 (0.35-11.34) 0.222#
NS Social Class
Low(III,IV,V) High(I and II)
31(30.1) 18(22.8)
72(69.9) 61(77.2)
1.46 (0.71-3.03) 0.270*
NS Marital Status
Single Married
0 50(27.9)
5(100.0) 129(72.1)
0.00 (0.00-2.92) 0.201#
NS Parity
0 1-4
≥5
1(20.0) 42(26.9)
9(33.3)
4(80.0) 114(73.1)
18(66.7)
0.68 (0.01-7.13)^
1.36 (0.52-3.51)@
0.732*
NS
* Chi square test
# Fisher’s Exact test S Significant
NS not significant
$ Odds ratio and 95% confidence interval for low maternal age
& Odds ratio and 95% confidence interval for high maternal age
^ Odds ratio and 95% confidence interval for primiparity
@ Odds ratio and 95% confidence interval for high parity
Maternal antenatal and delivery characteristics were compared between early and late preterm deliveries. Factors significantly associated with early preterm delivery were prolonged rupture of membrane, premature rupture of membrane, chorioamnionitis and booking status as depicted in Table VII.
Table VII: Maternal antenatal and delivery characteristics vs early and late preterm delivery
Parameter Babies studied OR (95% CI) P value
Packed cell volume (%) Early preterm (%) Late preterm (%)
<30
≥30
4(22.2) 11(17.7)
14(77.8) 51(82.3)
1.32 (0.27-5.41) 0.450#
NS Previous abortion
Yes No
17(32.1) 32(24.6)
36(67.9) 98(75.4)
1.45 (0.68-3.08) 0.301*
NS Previous uterine curettage
Yes No
8(25.8) 41(27.0)
23(74.2) 111(73.0)
0.94 (0.35-2.44) 0.894*
NS Previous preterm delivery
Yes No
4(30.8) 45(26.5)
9(69.2) 125(73.5)
1.23 (0.26-4.69) 0.477#
NS Febrile illness
Yes No
5(45.5) 45(26.0)
6(54.5) 128(74.0)
2.37 (0.54-9.78) 0.146#
NS Antepartum haemorrhage
Yes No
5(31.3) 45(26.8)
11(68.8) 123(73.2)
1.24 (0.32-4.14) 0.450#
NS Prolonged rupture of membrane
Yes No
19(44.2) 31(22.0)
24(55.8) 110(78.0)
2.81 (1.28-6.15) 0.004*
S Premature rupture of membrane
Yes No
24(42.9) 26(20.3)
32(57.1) 102(79.7)
2.94 (1.41-6.16) 0.002*
S Urinary tract infection
Yes No
1(25.0) 49(27.2)
3(75.0) 131(72.8)
0.89 (0.02-11.41) 0.702#
NS Chorioamnionitis
Yes No
5(83.3) 45(25.3)
1(16.7) 133(74.7)
14.78 (1.57-703.99) 0.006#
S Pregnancy induced hypertension
Yes No
8(21.6) 42(28.6)
29(78.4) 105(71.4)
0.69 (0.27-1.74) 0.396*
NS Type of labour
Medically induced Spontaneous
18(27.3) 32(27.1)
48(72.7) 86(72.9)
1.01 (0.48-2.09) 0.982*
NS Booking status
Elsewhere UITH
33(39.3) 17(17.0)
51(60.7) 83(83.0)
3.16 (1.52-6.62) 0.001*
S
* Chi square test
# Fisher’s Exact test S Significant NS not significant
After linear logistic regression analysis, chorioamnionitis and booking status remained significantly associated with early preterm delivery. The model could only account for 14% of early preterm delivery. The goodness of fit for the model gave an F-value of 7.383; df: 4, p=0.000. Booking status was the strongest determinant of early preterm delivery as shown in Table VIII.
Table VIII: Linear logistic regression of maternal socio-demographic and antenatal variables on early preterm deliveries
Beta Coefficients T p value
Prolonged rupture of membrane -0.018 -0.138 0.890
Premature rupture of membrane 0.162 1.283 0.201
Chorioamnionitis 0.213 2.955 0.004
Booking status -0.234 -3.304 0.001
Mortality outcome of the deliveries
Twenty two (11.9%) out of the 185 recruited babies delivered preterm died compared to 17 (1.3%) out of 1337 recruited babies delivered at term. This difference was statistically significant with Fisher’s’s Exact test= 0.000.
Outcome by gestational age
Three (75%) of the 4 babies delivered before 28 completed weeks of gestation died while 15(32.6%) of 46 babies delivered between 28 weeks and 33 weeks died.
Mortality rates for those between 34-36 weeks and 37-42 weeks were 3% and 1.3%
respectively as shown in Figure 2. These differences were significant with p value = 0.000.
0 200 400 600 800 1000 1200 1400 1600
<28 28-33 34-36 37-42 Gestational age (weeks)
Number of babies
0 10 20 30 40 50 60 70 80
percentage mortality
Total Dead (%)
Figure 2: Outcome by gestational age
One hundred percent mortality was observed among babies delivered at 24, 26, 28 and 29 weeks gestational age as shown in Table IX. All the deliveries below 28 weeks recorded were livebirths.
Table IX: Distribution of mortality by specific gestational age
GA (weeks) Died (%) Alive (%) Total
24 1(100) - 1
26 1(100) - 1
27 1(50) 1(50) 2
28 3(100) - 3
29 1(100) - 1
30 4(57.1) 3(42.9) 7
31 3(42.9) 4(57.1) 7
32 - 13(100) 13
33 4(26.7) 11(73.3) 15
34 2(6.1) 31(93.9) 33
35 2(4.8) 40(95.2) 42
36 - 59(100) 59
37 4(3.2) 120(96.8) 124
38 1(0.4) 249(99.6) 250
39 2(0.5) 370(99.5) 372
40 5(1.3) 380(98.7) 385
41 5(2.5) 193(97.5) 198
42 - 9(100) 9
Total 39 1483 1522
Outcome by birthweight
Seventy five percent of 4 babies weighing less than 1000g at delivery and 52.6% of those between 1000 and 1499g died. Mortality rates for those between 1500-2499g, 2500-3500g and greater than 3500g were 6%, 1.4% and 0.4% respectively as shown in Figure 3. These differences were significant with p value = 0.000.
0 200 400 600 800 1000 1200
<1000 1000-1499
1500-2499
2500-3500
>3500
birth weight (grammes)
Number of babies
0 10 20 30 40 50 60 70 80
percentage mortality
Total Dead (%)
Figure 3: Outcome by birthweight Outcome of all preterm babies
Three percent of late preterm babies died compared to 36% of early preterm babies. The difference was statistically significant as shown in Table X.
Table X: Outcome of early preterm vs late preterm babies
Gestation age Dead (%) Alive (%) Total
Early preterm 18(36.0) 32(64.0) 50
Late preterm 4(3.0) 130(97.0) 134
Total 22 162 184
χ2 at 1 df= 37.703; p value= 0.000; OR = 18.28; 95% CI = 5.41-77.74 Duration on admission before death
Fifteen percent of mortality among all recruited subjects took place within 6 hours of birth. More than two thirds of the mortalities (74.4%) took place within 72 hrs of birth and less than one third of all death (25.6%) took place after 72 hours as illustrated in Figure 4.
6
23 10
<6hrs 6hrs to 72hrs
>72hrs
Figure 4: Duration on admission before death among all babies
Fourteen percent of mortality among all recruited preterm subjects took place within 6 hours of birth. More than two thirds of the mortalities (72.7%) took place within 72 hrs of birth. Less than one third of all death (27.3%) took place after 72 hours as illustrated in Figure 5.
3
13 6
<6hrs 6hrs to 72hrs
>72hrs
Figure 5: Duration on admission before death among preterm babies
Duration of admission among babies admitted and discharged home alive
Table XI illustrates the mean duration of admission among the various categories of babies. The mean duration of admission for preterm babies discharged home alive was 8.5 ± 7.5 days while that for term babies was 4.3 ± 2.9 days. The mean duration of admission among early preterm babies discharged home alive was 10.8 ± 10.0 days compared to that for late preterm babies which was 7.0 ± 5.0 days.
The differences between these various categories were statistically significant with a p value < 0.05.
Table XI: Mean duration of admission among babies admitted and discharged home alive
Category of Baby n Mean ± S D (days) Significance * Preterm
Term
64 121
8.5 ± 7.5 4.3 ± 2.9
0.000 S Early Preterm
Term
25 121
10.8 ± 10.0 4.3 ± 2.9
0.000 S Early Preterm
Late Preterm
25 39
10.8 ± 10.0 7.0 ± 5.0
0.016 S
* = Mann-Whitney U S Significant
Duration of admission among babies that died
Mean duration of admission among preterm babies that died was 2.0 ± 2.3 days while that for term babies was 2.1 ± 2.7 days. The mean duration of admission among early preterm babies that died was 1.7 ± 2.2 days compared to late preterm babies; 3.5
± 2.4 days. None of the differences between these various means was statistically significant with p values > 0.05 as shown in Table XII.
Table XII: Mean duration of admission among babies that died
Category of Baby n Mean ± S D (days) Significance * Preterm
Term
22 17
2.0 ± 2.3 2.1 ± 2.7
0.954 NS Early Preterm
Term
18 17
1.7 ± 2.2 2.1 ± 2.7
0.641 NS Early Preterm
Late Preterm
18 4
1.7 ± 2.2 3.5 ± 2.4
0.094 NS
* = Mann-Whitney U NS not significant
Pattern of morbidities among the babies
Table XIII shows the morbidities encountered by preterm and term babies. All listed morbidities (except polycythaemia, anaemia and hypoglycaemia) occurred significantly more frequently in preterm than term babies.
Table XIII: Morbidities encountered by subjects
Morbidity Yes (%) No (%) OR (95% CI) P value
Severe perinatal asphyxia Preterm
Term
23(12.5) 39(2.9)
161(87.5) 1294(97.1)
4.74 (2.67-8.40) 0.000*
S Sepsis
Preterm Term
19(10.3) 21(1.6)
165(89.7) 1311(98.4)
7.19 (3.62-14.27) 0.000#
S Apnoea
Preterm Term
12(6.5) 7(0.5)
172(93.5) 1325(99.5)
13.21 (4.76-37.63) 0.000#
S Polycythaemia
Preterm Term
2(1.1) 3(0.2)
182(98.9) 1329(99.8)
4.87 (0.40-42.72) 0.114#
NS Anaemia
Preterm Term
- 4(0.3)
184(100.0) 1328(99.7)
0.00 (0.00-11.00) 0.596#
NS Respiratory distress
Preterm Term
66(35.9) 73(5.5)
118(64.1) 1259(94.5)
9.65 (6.46-14.40) 0.000*
S Jaundice
Preterm Term
32(17.4) 15(1.1)
152(82.6) 1317(98.9)
18.48 (9.41-36.68) 0.000*
S Hypoglycaemia
Preterm Term
2(1.1) 9(0.7)
182((98.9) 1323(99.3)
1.62 (0.17-7.82) 0.633#
NS Necrotizing enterocolitis
Preterm Term
3(1.6) 1(0.1)
181(98.4) 1331(99.9)
22.06 (1.75-1158.93)
0.006#
S
* Chi square test
# Fisher’s Exact test S Significant
NS not significant
Similarly, all listed morbidities (except polycythaemia, necrotizing entero-colitis and hypoglycaemia) occurred significantly more frequently in early preterm than late term babies. This is illustrated in table XIV.
Table XIV: Frequency of occurrence of morbidities in early and late preterm babies
Morbidity Yes(%) No(%) OR (95% CI) P value
Severe perinatal asphyxia Early Preterm
Late Preterm
17(34) 6(4.5)
33(66.0) 127(95.5)
10.90 (3.68-35.92) 0.000*
S Sepsis
Early Preterm Late Preterm
9(18.0) 10(7.5)
41(82.0) 123(92.5)
2.70 (0.93-7.83) 0.038*
S Apnoea
Early Preterm Late Preterm
11(22.0) 1(0.8)
39(78.0) 132(99.2)
37.23 (5.02-1614.07)
0.000#
S Polycythaemia
Early Preterm Late Preterm
1(2.0) 1(0.8)
49(98.0) 132(99.2)
2.69 (0.03-213.03) 0.473#
NS Respiratory distress
Early Preterm Late Preterm
38(76.0) 28(21.1)
12(24.0) 105(78.9)
11.88 (5.17-27.79) 0.000*
S Jaundice
Early Preterm Late Preterm
19(38.0) 13(9.8)
31(62.0) 120(90.2)
5.66 (2.35-13.76) 0.000*
S Hypoglycaemia
Early Preterm Late Preterm
0 2(1.5)
50(100.0) 131(98.5)
0.00 (0.00-14.22) 0.527#
NS Necrotizing enterocolitis
Early Preterm Late Preterm
2(4.0) 1(0.8)
48(96.0) 132(99.2)
5.50 (0.28-326.87) 0.182#
NS
* Chi square test
# Fisher’s’s Exact test S Significant
NS not significant
Determinants of Mortality
Maternal socio-demographic determinants of mortality among preterm babies When outcome among preterm babies was analysed with respect to maternal socio-demographic variables as shown in Table XV, no variable was statistically significant.
Table XV: Maternal socio-demographic variables vs outcome among preterm babies
Parameter outcome OR (95% CI) P value
Marital status Dead (%) Alive (%) Single
Married
0 (0) 22(12.2)
5(100) 158(87.8)
0.00 (0.00-8.33) 0.527#
NS Maternal age
<20 20-34
≥35
0 (0) 18(13.1)
4(9.3)
5(100) 119(86.9)
39(90.7)
0.00 (0.00-7.79)$ 0.68 (0.16-2.24)&
0.562*
NS
Body mass index (Kg/m2)
<19.9
≥19.9
0(0) 2(1.0)
33(100.0) 191(99.0)
0.00 (0.00-31.45) 0.729#
NS Socioeconomic class
Low(3, 4, 5) High(1 & 2)
11(10.7) 10(12.5)
92(89.3) 70(87.5)
0.84 (0.31-2.28) 0.702*
NS Parity
0 1-4
≥5
1(20.0) 21(13.6)
0 (0)
4(80.0) 133(86.4)
26 (100)
1.58 (0.03-17.00)^
0.00 (0.00-1.07)@
0.118*
NS
* Chi square test
# Fisher’s Exact test S Significant
NS not significant
$ Odd ratio and 95% confidence interval for low maternal age
& Odd ratio and 95% confidence interval for high maternal age
^ Odd ratio and 95% confidence interval for primiparity
@ Odd ratio and 95% confidence interval for high parity
Maternal antenatal determinants of mortality among preterm babies
The booking status was significant as a determinant of mortality when maternal antenatal factors were analysed against outcome among preterm babies as illustrated in Table XVI.
Table XVI: Maternal antenatal factors vs outcome among preterm babies
Parameter Outcome OR (95% CI) p value
Booking status Elsewhere UITH
Dead(%) 15(17.9) 7(6.9)
Alive(%) 69(82.1) 94(93.1)
2.92 (1.04-8.41) 0.022*
Previous abortion Yes
No
9(17.0) 13(9.9)
44(83.0) 188(90.1)
2.96 (1.08-7.99) 0.181*
NS Previous preterm delivery
Yes No
2(15.4) 20(11.7)
11(84.6) 151(88.3)
1.37 (0.14-7.00) 0.477#
NS Febrile Illness
Yes No
3(27.3) 19(10.9)
8(72.7) 155(89.1)
3.06 (0.48-14.10) 0.128#
NS Antepartum haemorrhage
Yes No
4(25.0) 18(10.7)
12(75.0) 151(89.3)
2.80 (0.59-10.51) 0.104#
NS Prolonged rupture of membrane
Yes No
3(7.0) 19(13.4)
40(93.0) 123(86.6)
0.49 (0.09-1.78) 0.256#
NS Premature rupture of membrane
Yes No
6(10.7) 16(12.4)
50(89.3) 113(87.6)
0.85 (0.26-2.45) 0.744#
NS Urinary tract infection
Yes No
1(25.0) 21(11.6)
3(75.0) 160(88.4)
2.54 (0.05-33.10) 0.400#
NS Chorioamnionitis
Yes No
2(33.3) 20(11.2)
4(66.7) 159(88.8)
3.97 (0.34-29.49) 0.151#
NS Pregnancy induced hypertension
Yes No
5(13.5) 17(11.5)
32(86.5) 131(88.5)
1.20 (0.32-3.74) 0.460#
NS Type of Labour
Medically induced Spontaneous
9(13.6) 13(10.9)
57(86.4) 106(89.1)
1.29 (0.47-3.45) 0.585*
NS Packed cell volume (%)
<30
≥30
4(21.1) 5(8.1)
15(78.9) 57(91.9)
3.04 (0.53-15.90) 0.125#
NS
* Chi square test
# Fisher’s Exact test S Significant NS not significant
Delivery factors as determinants of mortality among preterm babies
None of the delivery factors studied was significant as a determinant of mortality among recruited preterm babies as shown in Table XVII.
Table XVII: Delivery factors vs outcome among preterm babies
Parameter Outcome OR (95% CI) p value
Labour duration
>12hrs
≤12hrs
Dead (%) 2(22.2)
6(7.0)
Alive(%) 7(77.8) 80(93.0)
3.81 (0.31-26.92) 0.166#
NS Mode of delivery
Operative Vaginal
8(15.4) 14(10.5)
44(84.6) 19(89.5)
0.25 (0.08-0.76) 0.359*
NS
* Chi square test
# Fisher’s Exact test NS not significant
Neonatal factors as determinants of mortality among preterm babies
Table XVIII show those neonatal factors that were significant as determinants of mortality among preterm babies. These were severe perinatal asphyxia, sepsis, apnoea, respiratory distress, necrotizing enterocolitis and duration of resuscitation.
Table XVIII: Neonatal factors vs outcome among preterm babies
Parameter Outcome OR (95% CI) p value
Gender Male Female
Dead (%) 10(12.3) 11(10.7)
Alive (%) 71(87.7) 92(89.3)
1.18 (0.43-3.19) 0.724*
NS
Class SGA AGA LGA
2(33.3) 18(10.6)
0
4(66.7) 152(89.4)
7(100.0)
4.22 (0.35-31.52)$ 0.00 (0.00-6.38)&
0.137*
NS
Severe perinatal asphyxia Yes
No
16(69.6) 5(3.1)
7(30.4) 156(96.9)
71.31 (17.63-308.24) 0.000#
S
Sepsis Yes No
5(26.3) 16(9.7)
14(73.7) 149(90.3)
3.33 (0.82-11.43) 0.047#
S
Apnoea Yes No
11(91.7) 10(5.8)
1(8.3) 162(94.2)
178.20 (20.64-7709.02) 0.000#
S
Polycythaemia Yes
No
0 21(11.5)
2(100.0) 161(88.5)
0.00 (0.00-42.03) 0.784#
NS
Respiratory distress Yes
No
18(27.3) 3(2.5)
48(72.7) 115(97.5)
14.38 (3.88-78.46) 0.000*
S
Jaundice Yes No
3(9.4) 18(11.8)
29(90.6) 134(88.2)
0.77 (0.14-2.91) 0.484#
NS
Hypoglycaemia Yes
No
0 21(11.5)
2(100.0) 161(88.5)
0.00 (0.00-42.03) 0.784#
NS
Necrotizing enterocolitis Yes
No
3(100.0) 18(9.9)
0 163(90.1)
Undefined 0.001#
S
Resuscitation duration
>5min
≤5min 14(40.0)
6(11.1)
21(60.0) 48(88.9)
5.33 (1.62-19.02) 0.003##
S
* Chi square test
# Fisher’s Exact test
## Chi square with Yates correction S Significant
NS not significant
$ Odd ratio and 95% confidence interval for SGA babies
Analysis to exclude confounders of the determinants of mortality among preterm babies
After linear regression analysis, severe perinatal asphyxia, apnoea, necrotizing enterocolitis and duration of resuscitation remained significant. The model accounted for 66.8% of mortality among preterm babies. The goodness of fit for the model gave an F-value of 23.008; df: 7, p=0.000. The strongest determinant of mortality was apnoea followed by severe perinatal asphyxia as shown in Table XIX.
Table XIX: Linear logistic regression of determinants of mortality
Beta Coefficients T p value
Severe perinatal asphyxia -0.416 -4.312 0.000
Sepsis -0.093 -1.399 0.166
Apnoea -0.536 -6.182 0.000
Respiratory distress -0.035 -0.468 0.641
Necrotizing enterocolitis -0.236 -3.345 0.001
Resuscitation duration -0.184 -2.207 0.030
Booking status -0.027 -0.394 0.695