In-birth CPR in neonates; Predictable or Unpredictable?

Abstract

Background: Anticipating on in-birth Cardiopulmonary Resuscitation(CPR) in neonates is very important and complex. Timely identification and rapid CPR in neonates in the delivery room significantly affect reducing the mortality and other neurological disabilities. The aim of this study is to create a prediction system for identifying the need to in-birth CPR in neonates based on Machine Learning(ML) algorithms.

Methods: In this study, 3882 neonatal medical records were retrospectively reviewed. Records were extracted from the maternal, fetal, and neonatal registry of Valiasr hospital in Tehran. A total of 60 risk factors were extracted, and five ML algorithms including J48, Naïve Bayesian, Multilayer Perceptron (MLP)، Support Vector Machine (SVM) and Random Forest were compared to predict the need to in-birth CPR in neonates. Also, using 10 feature selection algorithms, the features were ranked based on the importance, and using the ML algorithms, the important risk factors were identified.

Results: In order to predict the need to in-birth CPR in neonates, SVM using all risk factors reached the accuracy of 88.43% and F-measure of 88.4%, while MLP using the 15 first important features reached the accuracy of 90.86% and the F-measure of 90.8%. The most important risk factors included gestational age, delivery type, presentation, steroid administration, macrosomia, prenatal care, infant number and rank, mother addiction, maternal chronic disease history, fetal hydrops, amniotic fluid, gestational hypertension, infertility and placental abruption.

Conclusions: The proposed system can be useful in predicting the need to CPR in neonates in the delivery room.

1. Background

Annually, around 1 million neonates die because of birth asphyxia worldwide(1). According to the WHO guideline on basic newborn resuscitation, although around one fourth of neonatal mortality is due to birth asphyxia, effective CPR at the childbirth moment can prevent a large number of these deaths(2). Most neonates enter from intrauterine to extrauterine life with no special assistance. However, less than 1% of all neonates (3) and around 0.1% of term neonates require advanced CPR at the moment of childbirth(4, 5). On the other hand, these statistics are very higher for preterm infants; 6–7% of preterm neonates (GA < 32 weeks) (6) and around 6–10% of Very Low Birth Weight (VLBW) and Extremely Low Birth Weight (ELBW) infants require advanced CPR, i.e. chest compression with or without injecting epinephrine (7). Many studies have been performed on CPR consequences, and it has been found that mortality, neurological morbidity, neurodevelopmental impairment, lower motor scores and ROP are more prevalent among the preterm infants who have received CPR(6, 8, 9). Thus, timely identification and rapid CPR of neonates in the delivery room can reduce the neonatal mortality and morbidity(7).

Currently, in-birth CPR is suggested for neonates with asystole, profound bradycardia (HR < than 60 per minutes), and pulseless electrical activity despite effective ventilation. Absence of heart rate or other vital signs, which is recorded as zero APGAR, can also be used as a guideline for decision-making on beginning CPR (8). Different studies have shown that the severity scoring systems have many limitations, and the systems based on machine learning have a better performance in prediction (10, 11). Accordingly, considering the importance of in-birth CPR, usage of ML based systems can be useful to help anticipating on the need to neonatal CPR. Application of ML algorithms in the medicine and especially in neonatal medicine has shown that these techniques have a very suitable performance in prediction and diagnosis.

Nevertheless, sparse studies have dealt with CPR in neonates, most of which have a small set of samples and risk factors because of the challenges in data collection (12–16). The aim of most of them is only identifying the risk factors affecting the need to CPR(12, 13, 17–19). Further, most studies have dealt with neonatal CPR in NICU, and fewer of them have addressed in-birth CPR, due to examine in-birth CPR, only maternal and fetal factors should be considered. To the best of our knowledge, no study has been performed on predicting the need to in-birth CPR in neonates using machine learning algorithms. Accordingly, our aim is to design a ML-based Clinical Decision Support System (CDSS) which predicts the need to in-birth CPR in neonates based on maternal and fetal factors.

2. Method

This retrospective study was conducted based on the maternal, prenatal and the fetal data, with the aim of predicting the need to in-birth CPR in neonates. In order to develop the prediction model, ML algorithms were used. Also, the models were evaluated to examine the performance and to determine the best model. The details related to the data, setting, method of development, and evaluation of the prediction models have been presented in this section.

2.1. Data source

The data were obtained through the neonatal registry system in Valiasr hospital affiliated to Tehran University of Medical Sciences (TUMS). This registry includes the information related to all neonates hospitalized in the NICU of Valiasr Hospital that has a grade of B3. The data related to the mother and fetus are recorded by pediatric residents in the data collection form, and are then entered into the registry by its person in charge. In this retrospective study, the data available in this registry were retrieved anonymously from March 2016 to March 2020.

All neonates hospitalized in NICU of Valisasr Hospital from March 2016 to March 2020 were included in this study. The post-delivery data such as Apgar score, height, and weight of the neonate were excluded.

2.3. Definition

In this study, delivery room CPR and CPR immediately after birth have been examined. CPR refers to any activity in the delivery room taken to simulate the cardiorespiratory activity of neonates who met the conditions of CPR according to the American Academy of pediatrics (AAP) guidelines(20, 21). These activities can be categorized into two groups: primary CPR (use of oxygen mask, Nasal continuous positive airway pressure (CPAP) and positive pressure ventilation(PPV)) and advanced CPR (primary CPR plus epinephrine injection, chest compression and intubation) (22).

2.4. Data extraction and preprocessing

After removing the any identifiers, the data was extracted from the registry as a .sav file and includes six groups:

Gestational risk factors: prenatal care, chorioamnionitis, steroid administration, magnesium sulfate administration
Maternal risk factors: age, hypertension (chronic, gestational, eclampsia), diabetes (chronic, gestational), addiction, HIV, chronic disease history, history of abortion (less than 20 weeks) and Intrauterine Fetal Death (IUFD)
Female infertility: use of Assisted reproductive techniques (ART), name of ART
Accreta status: decollement / placenta abruption, vasa previa, previa, placenta accreta
Fetal Data: gender, gestational age, rank and number of infant, Intrauterine growth restriction(IUGR), congenital problems, fetal hydrops
Delivery risk factors: mode of delivery, Prelabor rupture of membranes(PRoM), duration of PRoM, presentation, cord status, thick meconium, amniotic fluid status, Fetal Heart Rate (FHR)

The outcome variable is whether or not CPR is performed for a baby in the delivery room. The discrete and continuous missing values in the data set were imputed by the mode and mean of each variable, respectively.

2.5. Prediction models construction

In order to develop the prediction model for the need to in-birth CPR, ML algorithms were used including J48, MLP, SVM, Naïve Bayesian (NB) and Random Forest (RF). All of these algorithms were implemented for the original data set. Next, using Feature Selection (FS) algorithms, the importance of each feature in predicting CPR was examined. For this purpose, filter FS algorithms including Relief, Correlation-based feature selection (CFS), Pearson's Correlation, gain ratio, info gain, OneR and symmetrical uncertainty as well as wrapper methods using three classifiers SVM, J48 and NB were used (Table 1). Then, the risk factors were organized based on the total importance resulting from implementing 10 FS algorithms. Based on the ordered list of variables, various data subsets were created, and ML algorithms were implemented on both the original data set as well as these data subsets.

Table 1

Characteristics of feature selection methods
Type of FS method	Evaluation algorithm	Weka class name	Parameters tuning
Filter	Attribute Evaluation using RELIEF	ReliefFAttributeEval
	Correlation-based feature selection (CFS) Evaluation	CfsSubsetEval
	Attribute Evaluation using Pearson's Correlation	CorrelationAttributeEval
	Attribute Evaluation using Gain ratio	GainRatioAttributeEval
	Attribute Evaluation using Information Gain	InfoGainAttributeEval
	OneR uses the accuracy of a single-attribute classifier	OneRAttributeEval
	Attribute Evaluation using Symmetrical uncertainty	SymmetricalUncertAttributeEval
Wrapper	Subset Evaluation by using a user-specified classifier and separate held-out test set	ClassifierSubsetEval	Classifier = SVM
	Subset Evaluation by using a user-specified classifier and internal cross-validation	WrapperSubsetEval -weka.classifiers.trees.J48	Classifier = J48
	Subset Evaluation by using a user-specified classifier and internal cross-validation	WrapperSubsetEval -weka.classifiers.bayes.NaiveBayes	Classifier = NB

2.6. Statistical analysis and performance measurements

For the continuous data, mean and standard deviation, while for discrete data, frequency and percentage have been reported. To investigate the distribution of variables in the two groups (neonates receiving CPR vs those not receiving CPR), independent samples t-test, chi-square, Fisher exact and Mann-Whitney tests have been used. The significance level in all tests was considered p < 0.05. All statistical analyses were performed by SPSS 20. After analyzing the role of risk factors in predicting the outcome and developing the prediction models for the need to in- birth CPR, the performance of the developed models was evaluated based on the accuracy, precision, sensitivity, specificity and F-measure criteria and using the 10-fold cross validation method. The role of variables was analyzed using the FS algorithms in WEKA software. Development and assessment of models were also performed using R v3.4.1.

2.7. Clinical decision support system design

After selecting the best algorithm in the predicting the need to delivery room CPR in neonates, the system user interface was designed based on the best algorithm in Visual Studio platform 2015.

3. Results

A total of 3882 infants were included into the study according to the inclusion/exclusion criteria, out of them 2011(51.8%) had received delivery room CPR. The efforts were taken for the CPR, in the order of frequency, were: nasal CPAP (n = 1120, P = 28.8%), PPV (n = 891, P = 22.9%), Oxygen mask (n = 723, P = 18.6%), intubation (n = 494, P = 12.7%), chest compression (n = 86, P = 2.2%) and epinephrine injection (n = 68, P = 1.7%). Data statistics are shown in Tables 2 and 3.

Table 2

Descriptive statistics of discrete risk factors
	Variable name	Values	Frequency	Percent	P-Value
Gestational risk factors
	Prenatal care	Yes No	3426 456	88.25 11.75	.000
	Chorioamnionitis	Yes No	71 3811	1.83 98.17	0.958
	Steroid administration	Yes No	933 2949	24.03 75.97	.000
	Magnesium sulfate administration	Yes No	333 3549	8.58 91.42	.000
Maternal risk factors
	Hypertension	Yes No	184 3698	4.74 95.26	0.00752
	Gestational hypertension (Ghypertension)	Yes No	654 3228	16.85 83.15	0.0005595
	Diabetes	Yes No	105 3777	2.71 97.29	0.7826
	Gestational diabetes (Gdiabetes)	Yes No	600 3282	15.46 84.54	0.5829
	Mother addiction	Yes No	63 3819	1.62 98.38	0.006855
	Mother HIV	Yes No	28 3854	0.72 99.28	0.848
	Cardiac diseases	Yes No	304 3578	7.83 92.17	0.06417
	Blood diseases	Yes No	187 3695	4.82 95.18	0.8656
	Kidney diseases	Yes No	63 3819	1.62 98.38	0.9263
	Thyroid Disorders	Hyperthyroidism Hypothyroidism Thyroidectomy No	15 694 2 3171	0.39 17.88 0.05 81.68	0.2737
	Respiratory diseases	Yes No	28 3854	0.72 99.28	0.8509
	Mental diseases	Yes No	21 3861	0.54 99.46	0.9576
	Infectious diseases	Yes No	16 3866	0.41 99.59	0.885
	Brain diseases	Yes No	62 3820	1.6 98.4	0.211
	Cancer diseases	Yes No	33 3849	0.85 99.15	0.5051
	Skin diseases	Yes No	7 3875	0.18 99.82	0.6354
	Liver diseases	Yes No	63 3819	1.62 98.38	0.8716
	Autoimmune diseases	Yes No	64 3818	1.65 98.35	0.771
	Uterus diseases	Yes No	41 3841	1.06 98.94	0.5801
	Digestive diseases	Yes No	34 3848	0.88 99.12	0.3677
	Eye diseases	Yes No	4 3878	0.10 99.9	0.9424
	Other chronic disease ( Mother)	Yes No	12 3870	0.31 99.69	0.1073
	(Pre)Eclampsia	Eclampsia Preeclampsia No	8 198 3676	0.21 5.10 94.69	0.192
	Abortion history	Yes No	17 3865	0.44 99.56	0.925
	Intrauterine Fetal Death (IUFD)	Yes No	10 3872	0.26 99.74	0.1671
Infertility
	Female infertility	Yes No	214 3668	5.51 94.49	.000
	Assisted reproductive techniques (ART)	Yes No	144 3738	3.71 96.29	0.0009746
	Name of ART technique	No Drug administration IUI IVF	3738 26 18 100	96.29 0.67 0.46 2.58	0.000333
Accreta status
	Decollement / Placenta abruption	Yes No	41 3841	1.06 98.94	0.03364
	Vasa Previa	Yes No	1 3881	0.03 99.97	0.3347
	Previa	Yes No	113 3769	2.91 97.09	0.5083
	Placenta Accreta	Yes No	163 3719	4.2 95.8	0.1257
Fetal Data
	Number of infants	1 2 3 4	3407 419 55 1	87.76 10.79 1.42 0.03	.000
	Sex	Female Male Ambiguous Genitalia	1730 2146 6	44.57 55.28 0.15	0.3964
	Rank of infant	1 2 3	3628 235 19	93.46 6.05 0.49	.000
	IUGR	Yes No	223 3659	5.75 94.25	0.01075
	Tumors	Yes No	14 3868	0.36 99.64	0.6887
	Genetic problems/ Anomaly	Yes No	18 3864	0.46 99.54	0.08225
	Macrosomia	Yes No	19 3863	0.49 99.51	0.001636
	Cardiac problems	Yes No	31 3851	0.8 99.2	0.983
	Surgery (Defect of the abdominal)	Yes (including Colonic atresia, diaphragmatic hernia, duodenal atresia, Esophageal atresia, Gastroschisis, internal hernia, Intestinal atresia, Jejunal atresia, Omphalocele) No	54 3828	1.39 98.61	0.2758
	Blood problems	Yes No	4 3878	0.10 99.9	0.9424
	Pulmonary problems	Yes No	12 3870	0.31 99.69	0.1073
	Brain problem	Yes No	25 3857	0.64 99.36	0.9842
	Fetal hydrops	Yes No	12 3870	0.31 99.69	0.02858
	Other Problems(Fetus)	Yes No	6 3876	0.15 99.85	0.9295
Delivery risk factors
	Delivery type	Cesarean Vaginal	3617 265	93.17 6.83	.000
	PROM	Yes No	549 3333	14.14 85.86	0.0708
	Presentation	Breech Transverse Hand Normal	106 6 1 3769	2.73 0.15 0.03 97.09	0.07257
	Cord	Absent Doppler Cord prolapse Reverse No	27 4 1 3850	0.69 0.10 0.03 99.18	0.2395
	Thick meconium	Yes No	24 3858	0.62 99.38	0.1438
	Amniotic fluid	Oligohydramnios Polyhydramnios Normal	43 26 3813	1.11 0.67 98.22	0.04057
	Fetal Heart Rate (FHR)	Arrhythmia BPP Bradycardia Tachycardia Decreased FHR Fetal distress Sinusoidal PVC No	1 2 6 10 269 8 1 1 3584	0.03 0.05 0.15 0.26 6.93 0.21 0.03 0.03 92.31	0.3952

Table 3

Descriptive statistics of continuous risk factors
Variable name	Mean	SD	P-Value [95% confidence interval]
Maternal age (Year)	30.89	5.9	0.4742 [-0.1497735, 0.3219842]
Gestational age(day)	247.15	25.17	1.179099e-158 [19.21917, 22.09842]
PROM hrs.	7.84	62.41	0.0004639 [-10.95175, -3.09293]

In order to develop the prediction model for the need to in-birth CPR, ML algorithms were used. Figure 1 displays the results obtained from implementing these algorithms on the original data set.

Figure 1 shows that based on all performance criteria the SVM method has had the best performance in predicting the need to in-birth CPR. Also, J48 method reached comparable results. In the next step of simulation, FS algorithms were employed. For this purpose, seven filter FS algorithms including Relief, Correlation-based feature selection (CFS), Pearson's Correlation, gain ratio, info gain, OneR and symmetrical uncertainty as well as three wrapper methods using three classifiers SVM, J48 and NB were implemented (appendix A). Then, for each risk factor, the total importance resulting from implementing 10 FS algorithms was calculated. Table 4 presents the rank resulting from implementing the FS algorithms as well as the averaged rank for each variable. The averaged rank was calculated using the following relation, where r_i represents the rank of variable in the i^th feature selection algorithm.

Table 4

Rank of attributes based on 10 feature selection methods
#	Variable names	Relief	CFS	Correlation	Gain ratio	Info gain	OneR	symmetrical uncertainty	Wrapper (SVM)	Wrapper (NB)	Wrapper (J48)	Averaged rank
1	GA	1	1	1	1	1	1	1	1	1	1	1
2	Delivery type	5	4	5	8	6	4	5	4	2	2	4.5
3	Presentation	10	7	17	21	15	6	14	6	17	8	12.1
4	Steroids administration	2	2	2	3	2	2	2	2	46	60	12.3
5	Macrosomia	22	18	13	2	13	9	11	9	3	24	12.4
6	Prenatal care	6	6	6	14	5	3	6	3	47	39	13.5
7	Infant number	8	3	3	5	3	5	3	5	58	45	13.8
8	Addiction	12	12	15	19	17	7	17	7	14	19	13.9
9	Fetal hydrops	32	16	19	4	20	18	20	21	6	15	17.1
10	Infant rank	11	15	7	11	7	23	7	15	59	31	18.6
11	Amniotic fluid	17	10	18	22	18	8	18	8	42	27	18.8
12	Ghypertension	9	14	10	28	10	21	16	25	10	59	20.2
13	Infertility	13	13	8	12	8	17	8	19	57	48	20.3
14	Decollement / Placenta abruption	27	11	20	20	23	30	23	31	4	14	20.3
15	Other Chronic disease (Mother)	29	23	26	10	28	28	26	32	5	6	21.3
16	Magnesium sulfate	3	5	4	6	4	52	4	60	55	46	23.9
17	IUFD	48	21	31	16	33	12	31	10	30	9	24.1
18	Hypertension	16	17	14	26	16	20	19	17	40	57	24.2
19	Genetic problems/Anomaly	26	36	25	15	26	29	23	33	11	25	24.9
20	IUGR	20	30	16	27	19	24	21	14	41	38	25.0
21	Vasa Previa	41	27	37	7	36	39	35	27	7	4	26.0
22	Surgery	28	34	34	31	35	14	34	13	18	23	26.4
23	Cardiac diseases	14	24	22	35	25	11	28	11	52	51	27.3
24	ART use	37	59	11	23	12	19	12	16	56	30	27.5
25	Cord	39	19	21	18	21	44	22	28	51	16	27.9
26	PROM hrs.	50	9	9	17	14	33	15	30	60	44	28.1
27	Pulmonary problems	34	26	27	9	29	31	25	35	31	35	28.2
28	ART name	23	22	12	13	9	51	9	49	50	47	28.5
29	Thick meconium	31	25	30	24	31	27	30	29	37	21	28.5
30	Brain diseases	56	33	33	29	34	15	33	12	9	43	29.7
31	Digestive diseases	38	41	36	30	37	13	37	24	29	33	31.8
32	Placenta Accreta	51	31	28	34	30	32	32	26	28	36	32.8
33	PROM	24	57	23	37	27	16	29	20	49	52	33.4
34	FHR	44	8	29	25	11	53	13	54	54	54	34.5
35	Sex	4	28	35	44	32	23	36	46	43	55	34.6
36	Skin diseases	53	49	43	33	42	22	41	36	25	3	34.7
37	(Pre) Eclampsia	54	29	24	32	24	46	27	42	16	58	35.2
38	Thyroid Disorders	19	20	32	36	22	43	24	55	48	53	35.2
39	Eye diseases	33	32	53	40	52	40	52	37	12	7	35.8
40	Abortion	30	43	45	43	44	41	44	48	13	10	36.1
41	Infectious diseases	47	35	48	42	46	36	46	40	19	5	36.4
42	Cancer	35	38	40	39	39	26	38	23	38	49	36.5
43	Tumors	40	40	44	38	43	49	42	44	26	13	37.9
44	Gdiabetes	7	51	42	57	41	25	43	18	44	56	38.4
45	Previa	25	53	39	45	38	35	39	39	45	32	39.0
46	Uterus diseases	59	39	41	41	40	34	40	34	33	42	40.3
47	Blood problems	43	37	60	49	59	42	59	45	22	12	42.8
48	HIV	58	45	47	47	47	48	45	50	20	28	43.5
49	Other problems (Fetus)	49	48	55	46	54	37	54	47	34	11	43.5
50	Chorioamnionitis	15	58	58	58	56	38	55	53	27	26	44.4
51	Diabetes	55	56	46	56	45	59	47	57	15	18	45.4
52	Cardiac problems	36	54	50	48	48	57	48	52	24	37	45.4
53	Blood diseases	18	47	59	59	58	47	58	38	36	34	45.4
54	Maternal age	21	60	38	60	60	60	60	22	53	22	45.6
55	Mental diseases	57	52	56	53	55	50	57	43	8	29	46.0
56	Respiratory diseases	42	44	57	50	57	54	56	41	21	41	46.3
57	Liver diseases	45	46	52	55	51	56	51	59	32	17	46.4
58	Brain problem	46	42	54	51	53	55	53	56	23	40	47.3
59	Autoimmune	60	55	49	52	49	45	49	51	35	29	47.4
60	Kidney diseases	52	50	51	54	50	58	50	58	39	50	51.2

Averaged rank: (r₁ + r₂ + … + r₁₀) / 10

According to Table 4, GA is the most important risk factor based on implementation of all FS algorithms. Also, the averaged rank of the "maternal kidney disease" variable was the lowest, suggesting that it is the least important risk factor. The variables have been sorted based on the averaged rank, then, 20 feature subsets were created, including 1, 2, …,10,15,20,25,30,35,40,45,50,55 and 60 important variables respectively. According to these subsets, 20 data subsets were created, and ML algorithms were implemented on these data subsets. Figures 2 and 3 reveal the results related to implementing the ML algorithms for 20 data subsets obtained from FS.

According to Figs. 2 and 3, MLP using 15 first important variables with accuracy of 90.86% and F-measure of 90.8% has achieved the best results. The first 15 most important variables are gestational age, delivery type, presentation, steroid administration, macrosomia, prenatal care, infant number and rank, mother addiction, maternal chronic disease history, fetal hydrops, amniotic fluid, gestational hypertension, infertility and placental abruption.

Table 5 provides the best results of every algorithm. Comparing Fig. 1 and Table 5, it is found that use of FS algorithms has caused increased accuracy and F-measure by 4.7% and 4.9% respectively on average.

Table 5

The best performance of each ML methods on various feature subsets
ML Method	Accuracy	F-measure	No of selected features
MLP	90.86	90.8	15
J48	90.29	90.2	6
RF	90.24	90.3	3
SVM	90.76	90.8	9
NB	89.37	89	15

Graphical user interface of the proposed CDSS was designed based on the best algorithm in Visual Studio 2015 and shown in Fig. 4.

4. Discussion And Conclusions

This paper dealt with a prediction system for the need to neonatal CPR immediately after birth in delivery room. In order to achieve a system with proper performance, various ML algorithms were compared with different sets of risk factors to identify the best system as well as the most effective factors in CPR predicting. According to the obtained results, in order to predict the need to CPR in neonates, SVM using all risk factors reached accuracy of 88.43% and F-measure of 88.4%, while MLP using the first 15 most important variables reached accuracy of 90.86% and F- measure of 90.8%.

According to the sixth's edition of the textbook of neonatal resuscitation(23) and the International Liaison Committee On Resuscitation (ILCOR) guideline(24), the risk factors of GA, delivery type, presentation, macrosomia, prenatal care, multiple gestation, fetal hydrops, amniotic fluid status, hypertension, placental abruption and maternal chronic disease history all may contribute to increasing the need to in-birth CPR in neonates. In the study by Afjeh et al. the risk factors affecting the CPR in neonates were examined, whereby placental abruption, multiple gestation, delivery type and infertility were identified as the risk factors which may contribute to increasing the need to delivery room CPR(17). Also, in the study by Jiang et al., it was found that diabetes, hypertension and delivery type affect the need to CPR in neonates (18).

The prevalence of mortality, neurodevelopmental impairment, respiratory support at 28 days, days to full oral feeds and length of stay are very high among the neonates who have received in-birth CPR (8, 25). Even the national institute of child health and developmental neonatal research reported that CPR in delivery room is a prognostic factor for morbidity and later complications up to 18 months of age (26). Thus, the healthcare system should be able to predict which neonates require CPR before delivery, so that the neonatal resuscitation team would present in time (27). Previous studies have shown that antenatal transfer of high-risk mothers reduces pre-discharge neonatal mortality (28, 29). Thus, predicting the need to in-birth CPR can be very effective, as it increases the preparation of the neonatal resuscitation team and provides the possibility of consultation with the family before delivery (27). Therefore, according to the results obtained from this study, use of the proposed system for predicting the need to in-birth CPR in neonates will have a great significance in reducing the adverse outcomes in childbirth and better preparation of the CPR team.

In addition, the coordination between the CPR team and obstetricians can lead to reduced adverse events in the delivery room and improve the care (27). In the Draper's study, intrapartum deaths in the UK were examined and it was found around 25% of mortalities have been due to lack of suitable communication between the multidisciplinary team during delivery (30). Thus, the proposed system can be used as an infant pre-resuscitation guide in order to ensure coordination between the CPR team and obstetricians.

In spite of the great significance of CPR prediction, very few studies have dealt with neonatal CPR, out of which mostly have addressed CPR in NICU (12–15), which have samples with a small sample size and few risk factors because of the challenges in data collection (12–16). However, in this study, in addition to considering a sample with suitable size, attempts were made to capture all fetal and maternal risk factors mentioned in credible guidelines which also had demonstrated their importance in previous studies.

The main limitation of this study, like most previous studies, was that the data related to only one center were examined. Thus, it is suggested to conduct studies with a more diverse sample extracted from multiple centers with different grades of NICU. Comparison of the results can be useful in identifying significant risk factors affecting the need to CPR and thus in its prediction.

List Of Abbreviations

CPR: Cardiopulmonary Resuscitation

ML: Machine Learning

MLP: Multilayer Perceptron

SVM: Support Vector Machine

GA: Gestational Age

VLBW: Very Low Birth Weight

ELBW: Extremely Low Birth Weight

FHR: Fetal Heart Rate

CDSS: Clinical Decision Support System

APGAR: Appearance, Pulse, Grimace, Activity, Respiration

AAP: American Academy of pediatrics

CPAP: Continuous Positive Airway Pressure

PPV: Positive Pressure Ventilation

IUFD: Intrauterine Fetal Death

ART: Assisted Reproductive Techniques

IUGR: Intrauterine Growth Restriction

PRoM: Prelabor Rupture of Membranes

NB: Naïve Bayesian

RF: Random Forest

CFS: Correlation-based Feature Selection

FS: Feature Selection

ILCOR: International Liaison Committee On Resuscitation

Declarations

Ethics approval and consent to participate: Our study was approved by the institutional review board of Tehran University of Medical Sciences and according to Helsinki declaration (Approval ID: IR.TUMS.VCR.REC.1398.591). All participants’ parents gave their written informed consent before loading the data into the registry and for the participation in the study. Participants’ data were considered confidential and no extra cost was imposed on our participants.

Consent for publication: 'Not applicable'

Availability of data and materials: The data that support the findings of this study are available from manager of maternal, fetal, and neonatal registry of Valiasr hospital in Tehran but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of manager of maternal, fetal, and neonatal registry of Valiasr hospital in Tehran.

Competing interests: The authors declare that they have no competing interests

Funding: This research has been supported by Tehran university of medical sciences and health service as well as maternal, fetal and neonatal research center.

Authors' contributions: AO has participated in acquisition, analysis, interpretation of data; the creation of new software used in the work; MZ has participated in interpretation of data and substantively revised the work; AO and MZ have drafted the work. RM has participated in acquisition and interpretation of data. All authors read and approved the final manuscript.

Acknowledgments: This study was supported by Tehran University of Medical Sciences (TUMS) and maternal, fetal and neonatal research center. We acknowledge their kindly supports in this study.

References

Moshiro R, Mdoe P, Perlman JM. A Global View of Neonatal Asphyxia and Resuscitation. Frontiers in Pediatrics. 2019;7(489).
Organization WH. Guidelines on basic newborn resuscitation. 2012.
Reynolds C, Cox J, Livingstone V, Dempsey EM. Rescuer Exertion and Fatigue Using Two-Thumb vs. Two-Finger Method During Simulated Neonatal Cardiopulmonary Resuscitation. Frontiers in pediatrics. 2020;8:133-.
Cheung P-Y, Huang H, Xu C, Liu J-Q, Ting JY, Wong R, et al. Comparing the Quality of Cardiopulmonary Resuscitation Performed at the Over-the-Head Position and Lateral Position of Neonatal Manikin. Frontiers in Pediatrics. 2020;7(559).
Hosono S, Tamura M, Isayama T, Sugiura T, Kusakawa I, Ibara S, et al. Neonatal cardiopulmonary resuscitation project in Japan. Pediatrics International. 2019;61(7):634-40.
Fischer N, Soraisham A, Shah PS, Synnes A, Rabi Y, Singhal N, et al. Extensive cardiopulmonary resuscitation of preterm neonates at birth and mortality and developmental outcomes. Resuscitation. 2019;135:57-65.
Savani M, Upadhyay K, Talati AJ. Characteristics and outcomes of very low birth weight infants receiving epinephrine during delivery room resuscitation. Resuscitation. 2017;115:1-4.
Foglia EE, Weiner G, de Almeida MFB, Wyllie J, Wyckoff MH, Rabi Y, et al. Duration of Resuscitation at Birth, Mortality, and Neurodevelopment: A Systematic Review. Pediatrics. 2020;146(3):e20201449.
Sarrato Z, Sarrato Z, Chamorro P, Castro V, Pacheco G, Luna S. Perinatal risk factors and outcome of preterm infants ≤27 weeks of gestational age who received delivery room cardiopulmonary resuscitation (CPR): P-0381| POSTER| Miscellaneous. Journal of Perinatal Medicine. 2015;43.
Deliberato RO, Escudero GG, Bulgarelli L, Neto AS, Ko SQ, Campos NS, et al. SEVERITAS: An externally validated mortality prediction for critically ill patients in low and middle-income countries. International journal of medical informatics. 2019;131:103959.
Kefi Z, Aloui K, Naceur MS, editors. The early prediction of neonates mortality in Intensive Care Unit. 2019 International Conference on Advanced Systems and Emergent Technologies (IC_ASET); 2019: IEEE.
Ahmad KA, Velasquez SG, Kohlleppel KL, Henderson CL, Stine CN, LeVan JM, et al. The Characteristics and Outcomes of Cardiopulmonary Resuscitation within the Neonatal Intensive Care Unit Based on Gestational Age and Unit Level of Care. Am J Perinatol. 2019(EFirst).
Campbell ME, Byrne PJ. Cardiopulmonary resuscitation and epinephrine infusion in extremely low birth weight infants in the neonatal intensive care unit. Journal of Perinatology. 2004;24(11):691-5.
Kostelanetz AS, Dhanireddy R. Survival of the very-low-birth-weight infants after cardiopulmonary resuscitation in neonatal intensive care unit. Journal of Perinatology. 2004;24(5):279-83.
Lantos JD, Miles SH, Silverstein MD, Stocking CB. Survival after Cardiopulmonary Resuscitation in Babies of Very Low Birth Weight. New England Journal of Medicine. 1988;318(2):91-5.
Salhab WA, Wyckoff MH, Laptook AR, Perlman JM. Initial hypoglycemia and neonatal brain injury in term infants with severe fetal acidemia. Pediatrics. 2004;114(2 I):361-6.
Afjeh SA, Sabzehei MK, Esmaili F. Neonatal resuscitation in the delivery room from a tertiary level hospital: Risk factors and outcome. Iranian Journal of Pediatrics. 2013;23(6):675-80.
Jiang S, Lyu Y, Ye XY, Monterrosa L, Shah PS, Lee SK. Intensity of delivery room resuscitation and neonatal outcomes in infants born at 33 to 36 weeks' gestation. Journal of Perinatology. 2016;36(2):100-5.
Sood S, Giacoia GP. Cardiopulmonary Resuscitation in Very Low Birthweight Infants. American Journal of Perinatology. 1992;9(2):130-3.
Association AH. American Heart Association (AHA) guidelines for cardiopulmonary resuscitation (CPR) and emergency cardiovascular care (ECC) of pediatric and neonatal patients: pediatric advanced life support. Pediatrics. 2006;117(5):e1005-e28.
Association AH, Weiner G, NNP-BC JZMN, Pediatrics AA. Textbook of Neonatal Resuscitation (NRP) 7th Edition ed: Amazon Digital Services LLC - KDP Print US; 2016. 320 p.
Lee AC, Cousens S, Wall SN, Niermeyer S, Darmstadt GL, Carlo WA, et al. Neonatal resuscitation and immediate newborn assessment and stimulation for the prevention of neonatal deaths: a systematic review, meta-analysis and Delphi estimation of mortality effect. BMC public health. 2011;11(S3):S12.
Association AH. NRP: Textbook of Neonatal Resuscitation: American Academy of Pediatrics; 2011.
Kattwinkel J, Niermeyer S, Nadkarni V, Tibballs J, Phillips B, Zideman D, et al. ILCOR advisory statement: resuscitation of the newly born infant: an advisory statement from the pediatric working group of the International Liaison Committee on Resuscitation. Circulation. 1999;99(14):1927-38.
Bajaj M, Natarajan G, Shankaran S, Wyckoff M, Laptook AR, Bell EF, et al. Delivery Room Resuscitation and Short-Term Outcomes in Moderately Preterm Infants. The Journal of Pediatrics. 2018;195:33-8.e2.
Wyckoff MH, Salhab WA, Heyne RJ, Kendrick DE, Stoll BJ, Laptook AR, et al. Outcome of extremely low birth weight infants who received delivery room cardiopulmonary resuscitation. The Journal of pediatrics. 2012;160(2):239-44. e2.
Sawyer T, Lee HC, Aziz K, editors. Anticipation and preparation for every delivery room resuscitation. Seminars in Fetal and Neonatal Medicine; 2018: Elsevier.
Lasswell SM, Barfield WD, Rochat RW, Blackmon L. Perinatal regionalization for very low-birth-weight and very preterm infants: a meta-analysis. Jama. 2010;304(9):992-1000.
Mellander M, editor Perinatal management, counselling and outcome of fetuses with congenital heart disease. Seminars in Fetal and Neonatal Medicine; 2005: Elsevier.
Draper E, Kurinczuk J, Kenyon S. MBRRACE-UK 2017 perinatal confidential enquiry: term, singleton, intrapartum stillbirth and intrapartum-related neonatal death. Leicester: The Infant Mortality and Morbidity Studies, Department of Health Sciences, University of Leicester. 2017.