The Role of Place of Delivery in Preventing Neonatal and Infant Mortality Rate in India

: Skilled birth attendance and institutional delivery have been advocated for reducing maternal, neonatal mortality and infant mortality (NMR and IMR). This paper examines the role of place of delivery with respect to neo-natal and infant mortality in India using four rounds of the Indian National Family Health Survey conducted in 2015–2016. The place of birth has been categorized as “at home” or “public and private institution.” The role of place of delivery on neo-natal and infant mortality was examined by using multivariate hazard regression models adjusted for clus-tering and relevant maternal, socio-economic, pregnancy and new-born characteristics. There were 141,028 deliveries recorded in public institutions and 54,338 in private institutions. The esti-mated neonatal mortality rate in public and private institutions during this period was 27 and 26 per 1000 live births respectively. The study shows that when the mother delivers child at home, the chances of neonatal mortality risks are higher than the mortality among children born at the health facility centers. Regression analysis also indicates that a professionally qualiﬁed provider (cid:48) s antenatal treatment and assistance greatly decreases the risks of neonatal mortality. The results of the study illustrate the importance of the provision of institutional facilities and proper pregnancy in the prevention of neonatal and infant deaths. To improve the quality of care during and imme-diately after delivery in health facilities, particularly in public hospitals and in rural areas, accel-erated strengthening is required.


Introduction
Neonatal deaths account for a major proportion of the world's pediatric deaths. Globally four million deaths occur every year in the first month of life [1]. Almost all (99%) neonatal deaths found in low-income and middle-income countries [1,2]. Glob-ally, the neonatal mortality declined from 37 deaths in 1990 to 18 deaths per 1000 live births in 2017, but that decline was slower than the decline in mortality among children aged 1-59 months [3]. In 2018, 2.5 million children died in the first month of life-approximately 7000 neonatal deaths every day, with one third dying on the first day of life [4]. Consequently, the proportion of deaths in the neonatal period rose from 38% (4 million) of total deaths in 2000 to about 41% (3.3 million) have been studied in various locations, but no association between home-based deliveries and neonatal mortality has been found, regardless of the presence of qualified assistants. As a result, the study s goal is to identify the determinants of neonatal and infant mortality in India, as well as their relationship with facility-based delivery.

Data
The study is based on secondary data collected from the most recent round of the National Family Health Survey (NFHS-4) 2015-2016, which is a nationally representa-tive, multi-round, population-based survey conducted in all the states and union terri-tories of India under the direction of the Ministry of Health and Welfare, Government of India. The NFHS 4 is a nationwide cross-sectional household survey conducted with a representative sample of 601,509 households throughout the country during 2015-2016. The NFHS collects information on fertility, mortality, morbidity, and maternal and child health. The data were collected using face to face interview schedules and the household and eligible female informant response rates were consistently above 90% across the three rounds. The study used appropriate weights in the analysis to make the estimates representative and comparable across the four survey rounds, as well as to account for the multistage sampling design used in the four rounds of the NFHS. The unit of the analysis for this study is the children below 1 year of age; infor-mation was collected based on 259,627 observations from the kid file for the last five years with the women age group of 15-49. All the woman surveyed were asked to pro-vide complete birth histories including sex, date of birth, and survival status for each live birth; the analysis uses the birth history data.

Outcome Variables
Neo-natal and infant mortality are two dependent variables of the study. Where the neo-natal mortality defined as the death of life born baby within 28 days of life, while infant mortality defined as dead of life born baby before reaching age 12 months. The outcome variable for the study was recorded as a binary variable in the datasets, assigned a value of 1 if the baby died and 0 if the baby was survived.

Independent Variables
The primary independent variable of interest is the child's reported place of de-livery. The NFHS survey question on place of delivery includes the following options: Home (respondent's home or other home); Public Sector (government hospital, gov-ernment health centre, government health post or other public sector); or Private Medical Sector (private hospital or clinic, other private medical facility).
Covariates were included based on previous literature with the primary objective of adjusting for socio-demographic information about the mother, child and house-hold, such as maternal age below 19 and above 30, birth order as 1, 2 and above 3, birth interval as less or more than 24 months, mother BMI under, normal and obesity, ma-ternal education as no education, primary, secondary and higher secondary, sex of child as male/female, child birth weight as low/high, household wealth as poorest, middle and richest, place of residence as rural/urban, caste SCs, STs, OBCs and others, religion as Hindu, Muslim and Christian, mass media (newspaper, television and ra-dio), antenatal care received/not received, assistant by doctor and nurse and delivered by casearation or not.

Statistical Methods
Descriptive analysis was used to determine the distribution of the variables and bivariate analysis to measure the relationship between each of the independent varia-bles and the main exposure variable at the place of delivery (home vs. facility). Neo-natal and infant mortality rates were estimated based on the simple life table ap-proach using STATA 14.0. This approach builds up a 12-59 month matrix of exposure and deaths for each month that the child was alive between the first and 11 month of life.
The Cox Proportional Hazard Regression Model was used to measure hazard ra-tios (HRs) and 95% confidence intervals (CIs) to assess the odds of neonatal death and infant death by maternal, pregnancy level delivery care and socio-demographic status.
Multivariable analyses were assessed between place of delivery and the other variables in the model. Model 1 describes the adjusted hazard ratios between neonatal deaths, maternal, child, and childbirth characteristics at home, while Model 2 includes place of delivery and pregnancy delivery characteristics, and Model 3 calculates the adjusted hazard ratio by incorporating maternal, sociodemographic, and mater-nal-pregnancy delivery care factors. All analysis was performed using statistical soft-ware package STATA®(Version 14.0). Table 1 present the state-wise neonatal and infant mortality by place of delivery 2015-2016. It was found that Overall, 141,028 (54.5%) and 54,338 (20.9%) deliveries occurred in a public and private hospital, and 63,742 deliveries (24.5%) occurred in a home setting. State wise highest percentage of births delivered by public hospital are found in Sikkim (85%), Jammu and Kashmir (77%), Odisha (76%), Madhya Pradesh (69%) and Tripura (68%). Looking at private health facilities, highest percentage of births delivered in southern states, e.g., in Kerala 61% of birth delivered in public hos-pital followed by Telangana 58%, Andhra Pradesh 53% and Tamil Nadu 32%. Gujarat is only state of northern India where 51% births have been delivered in the Public hos-pitals. Similarly, highest percentage of births without health facility for home are found in north-eastern states, such as Nagaland (68%) followed by Arunachal Pradesh (49%), Meghalaya (47%) and Manipur (36%). In complete contract in neo-natal and infant mortality due to health facility highest mortality in public and private hospital both are found in the states of Uttar Pradesh followed by Chhattisgarh and Madhya Pradesh and Bihar respectively. Neonatal and infant mortality rate was significantly higher for home deliveries (39) than for deliveries at health facilities, but the difference is negligible for public and private hospitals, where 27 per thousand live births died in public hospital and 26 per thousand live births died in private hospitals, respectively.   Table 2 describes the Bio-demographic and Socio-economic factor affecting the neo-natal and infant mortality rate in the five years preceding the survey by place of delivery status. Almost half of the neonatal deaths were of children with mothers between the ages of 15-19 years. Both neonatal and infant mortality rates were higher in both public and private hospitals for mothers with birth intervals of less than 24 months. At both places of delivery, the male child was at greater risk of neonatal death than the female child. There was a higher risk of mortality among mothers living in the rural areas than the mothers living in urban areas due to lower health facilities, low awareness and low maternal education in rural areas as compare to urban areas. With regard to wealth, the quintile risk of mortality relatively higher when the poorest mother given birth at home then in institutional health facilities. Coming into pregnancy and delivery related characteristics higher number of neonatal died for both public and private hospital, those mothers had received less than four antenatal care during the pregnancy. The risk of survival of a new born baby was lower if delivered assistant by doctor as compare to the nurse. Similarly, higher chance of a new born baby died if the birth was delivered by casearation section as compared to normal births for both public and private hospital. Table 2 shows that the risk of neo-natal and infant mortality is the same in both public and private institutions, but the risk of death at home without a health facility is relatively higher than in an institutional health facility.  Male  44  30  28  58  41  35  Female  33  23  22  49  34  31  Mother education  No education  38  32  29  56  46  66  Primary  43  30  31  56  43  49  Secondary  37  24  22  51  33  29  Higher education  20  20  13  32  26  17  Place of residence  Urban  45  25  24  58  39  37  Rural  84  36  37  68  29  25  Wealth Status  Poorest  40  33  51  56  46  64  Middle  42  24  29  54  35  38  Richest  28  18  16  44  25  21  Caste  SC  42  28  37  57  40  45  ST  35  28  27  52  39  35  OBC  39  27  27  56  38  35  others  34  24  19  46  34  24  Religions  Hindu  41  27  27  56  38  34  Muslim  33  26  22  50  39  28  Christian  23  18  6  37  30  10  Mass media  No  40  34  50  57  47  63  At least one  35  23  20  50  33  27  Mothers received antenatal care  No  29  27  35  42  39  46  At least four  25  17  23  37  26  30  Assistant by Nurse  26  28  37  36  Assistant by Doctor  25  23  35  29  Casearation birth  No  26  29  37  37  Yes  31  21  41  27 Source: Authors calculation from NFHS-IV (2015-2016) data. Figure 1 presents the percentage distribution of live births and per thousand neo-natal mortality rates by place of delivery due to mother's education across the India. Around 40% of the deliveries in India were carried out at home with mother without education, while 56% in public institutions with primary education and 54% of the deliveries took place on the way to private hospital to mother of higher education. On the contrary home-based delivery mother without education neo-natal mortality rate was higher than those mothers delivered in births in public and private institution respectively. The declined in neo-natal mortality rate for each unit with increased mother education and in the institutional delivery.  Figure 2 depicts the percentage distribution of live births and per thousand neo-natal mortality rates by place of delivery due to place of residence in India. It was identified that only 13% of births in urban areas were delivered at home, while 28% were delivered in rural areas. There was no discernible difference between urban and rural public hospital births, but private institution-based delivery was higher in cities than in rural areas. In the context of a complete neonatal mortality contract based on place of residence, increased mortality rates were found in rural areas compared to urban areas for both place of delivery due to a lack of well-trained personnel, a lack of adequate health care facilities, a lack of medical coverage, and poor transportation facilities. Finally, results of the hazard regression models 1, 2 and 3 for neonatal and infant mortality were presented in Table 3 and Table 4, where Model 1 shows maternal level characteristics at home delivery, and Models 2 is pregnancy and delivery-based characteristics and Model 3 was adjusted with maternal socio-demographic, pregnancy and delivery characteristics. Model 1 revealed that the risk of neonatal mortality among children born to mothers who gave birth at home is 1.26 times  Figure 2 depicts the percentage distribution of live births and per thousand neo-natal mortality rates by place of delivery due to place of residence in India. It was identified that only 13% of births in urban areas were delivered at home, while 28% were delivered in rural areas. There was no discernible difference between urban and rural public hospital births, but private institution-based delivery was higher in cities than in rural areas. In the context of a complete neonatal mortality contract based on place of residence, increased mortality rates were found in rural areas compared to urban areas for both place of delivery due to a lack of well-trained personnel, a lack of adequate health care facilities, a lack of medical coverage, and poor transportation facilities.  Figure 2 depicts the percentage distribution of live births and per thousand neo-natal mortality rates by place of delivery due to place of residence in India. It was identified that only 13% of births in urban areas were delivered at home, while 28% were delivered in rural areas. There was no discernible difference between urban and rural public hospital births, but private institution-based delivery was higher in cities than in rural areas. In the context of a complete neonatal mortality contract based on place of residence, increased mortality rates were found in rural areas compared to urban areas for both place of delivery due to a lack of well-trained personnel, a lack of adequate health care facilities, a lack of medical coverage, and poor transportation facilities. Finally, results of the hazard regression models 1, 2 and 3 for neonatal and infant mortality were presented in Table 3 and Table 4, where Model 1 shows maternal level characteristics at home delivery, and Models 2 is pregnancy and delivery-based characteristics and Model 3 was adjusted with maternal socio-demographic, pregnancy and delivery characteristics. Model 1 revealed that the Finally, results of the hazard regression models 1, 2 and 3 for neonatal and infant mortality were presented in Tables 3 and 4, where Model 1 shows maternal level characteristics at home delivery, and Models 2 is pregnancy and delivery-based characteristics and Model 3 was adjusted with maternal socio-demographic, pregnancy and delivery characteristics. Model 1 revealed that the risk of neonatal mortality among children born to mothers who gave birth at home is 1.26 times higher than the risk of neonatal mortality among children born to women who have access to health care (HR: 1.26; 95 percent CI: 1.19, 1.34). The results also showed that mother age at births, birth interval more than 24 months, mothers' highest level of education, wealth quintile, place of birth (urban-rural), caste and religion emerged as statistically significant predictors of early-neonatal deaths in India in multivariate analysis. Table 3. Hazard ratios (95% confidence intervals) for neo-natal mortality by socio-economic, demographic and residence related characteristics, India, 2015-2016.   Table 3. Cont.    Hazard result Model 2 showed that the risk of neonatal death in a private hospital was 0.12 times higher than in a public hospital, with no statistically significant 95 percent confidence interval. Those mothers had received at least 4 ANC components during antenatal visits had 22 percent lower risks as compared to babies born to mothers who had not received any components of ANC during their visits (HR 0.78; 95 percent CI: 0.73, 0.83). The risk of new born baby dying 1.01 times higher when delivered by nurse as compare to assistant by doctor without statistically significant. At the same time risk of new born babies delivered by casearation section as compared to normal birth delivery (HR 1.01; 95 percent CI: 0.919-1.102) respectively.

Wealth Status
After adjustment maternal socio-demographic, pregnancy and delivery characteristics in Model 3, depicts those mother who had given birth in a private or public facility had significantly lower value of neonatal death than those who had given birth at home (HR 0.86 95 percent CI 0.76-0.97 and HR 1.24, 95 percent CI 1.08-1.43 respectively).The risks of neo-natal death has increased, while at the same time increasing the maternal age of birth, birth order and mother BMI in 0.59, 0.88 and 1.44 time with statistical significance respectively. On the contrary higher maternal age with education decreased by 22 percent in the early neo-natal death with 95 percent statistically significant. The hazard risks of dying in the early-neonatal period is 43 percent lower for a child born in the richest wealth quintile compared to a child born in the poorest wealth quintile group. It depicts from the results that babies born to women who had received more than 4 ANC components during antenatal visits had 32 percent lower risks of neo-natal mortality compared to babies born to mothers who had not received any components of ANC during their visits (HR 0.68; 95 percent CI: 0.61, 0.76).The odds of neonatal mortality among women who gave birth with assistant of doctor at health institutions was 1.03 lower than among those who gave birth attendants by nurses at institution, and this difference was not statistically significant.
The results of hazard models for infant mortality are presented in Table 4. Like neo-natal mortality, for model 1 we found mother age at birth, birth interval, mothers' highest level of education, wealth quintile, place of birth (urban-rural), religion, are statistically significant predictors of infant deaths in India in multivariate models.
For model 2, adjusted with pregnancy and delivery-based characteristics showed that the differential risk of infant death between public and private hospital was narrowed. The mothers who had received more than 4 antenatal care 46 percent have lower risk then who had not received any antenatal care.
For model 3 presented adjusted maternal, socio-demographic and delivery level characteristics. A significant association was found between mother's age and infant mortality; birth order and interval were also significantly associated with infant mortality. Irrespective of the birth order, new-borns with a birth interval less than 24 months had a higher risk of mortality during infancy than new-borns with a birth interval of 24 months or more. Another variable that was significantly associated with infant mortality was the region of residence. Table 5 presents the AHRs (with 95% confidence interval) of neo-natal and infant mortality by place of delivery controlling for all other socio demographic correlations between the states of India. These results are robust with the pattern obtained at the national level. The results clearly indicate that there is a higher risk of neo-natal and infant death at home as compared to those who gave birth in public and private health facilities. The higher risk neonatal died at home found in the state of Delhi, Meghalaya, Tamil Nadu and Jammu and Kashmir. When comparing public and private hospital health facilities, it was found that a private health institution had a higher risk of new-born babies than a public health institution due to untrained health workers and a lack of hospital facilities. These hospitals belonged to backward states such as, Mizoram, Bihar, Tripura, Arunachal Pradesh and Odisha. On contrary Delhi is one of the state, where risk of neo-natal mortality is lower than that of other states in private institution. While considering public health institution, higher risk of neo-natal death found in north-eastern states like Tripura, Mizoram, and Nagaland etc. similarly, results were found in infant mortality, the hazard risks of infant death higher at home as compared to health institution. The difference in infant mortality risk between public and private health institutions varied by state. * Adjusted hazard ratios estimates are controlled for mother age, birth order, birth interval, size at birth of the child, mother's education, BMI of the mother, place of residence, caste, religion, wealth quintile, assistant by doctor and nurse. *** p < 0.01, ** p < 0.05, * p< 0.10. Source: Authors' calculation from NFHS-IV (2015-2016) data.

Discussion
The present study provides evidence of disparity in neo-natal and infant mortality by place institutional delivery across India. The findings revealed that health facilities during delivery have statistically significant effect on neonatal and infant mortality. The new born baby at home had an increased risk of death during the neonatal period compared to those who gave birth in health facilities. Women who gave birth at home had a 1.3-fold increase in the risk of neonatal death compared to their counterparts. Previous research Ethiopia reported that most women had a strong aspiration to give birth at home considering it a natural space for delivery and allowing for traditional births related events to take place and thus making it more enjoyable. Even those who gave birth in health institutions appreciated events in connection to home delivery [25]. There are several factors affected the neo-natal and infant mortality. Maternal complications during delivery have been reported as a key factor in neonatal mortality in India [26]. Low mother age and lower female education ware one of leading cause of neo-natal and infant mortality in India. Similarly, baby with low birth weight were higher risk of death than those with a normal birth weight in the first month of life. When stratified by residence, institutional delivery appears to be consistently correlated with a decreased odd of neonatal mortality in both urban and rural areas, although the extent and substantial correlation have only been found in urban environments. In this study, rural dwellers experienced higher neonatal mortality than their urban counterparts. The possible reason may be related to a low level of maternal health service utilization amongst rural women compared with their urban counterpart [27]. Previous research has found that rural women are more readily influenced by traditional practices that are contrary to modern health care [25]. Among the various constituents of health care variables, ANC checks during delivery were found to have a significant impact on neonatal mortality and infant mortality in the states. Neonatal mortality was lowered by 32% among children whose mothers received ANC checks during pregnancy. Similarly, mortality risk of neo-born babies was higher when the birth delivered by casearation as compare to normal birth delivery.
The state wise difference of neonatal and infant mortality was remarkable. The neonatal mortality rate for central states was quite high compared to neonatal mortality rate in other states. On the other hand, the comparison of neonatal mortality rate for public and private hospital did not reveal much difference. This disparity might reflect differences in the quality of the public and private obstetric care available between states. Sikkim experienced the maximum in public institutional delivery coverage (85%), and minimum by Nagaland having 25% coverage. Similarly, minimum coverage in private institution were found Nagaland, Jammu and Kashmir and Odisha respectively. The growing number of studies found low coverage in birth in private institution mainly for high-cost delivery and unavailability services [15,[28][29][30]. It was also main barrier among poor people in economic backward states. Kerala is the only state with almost 61% deliveries being conducted in private institutions. Due to this institutional coverage delivery, mainly for high immunizations and universal health care services, the state of Kerala is becoming an ideal model for the developing countries [8]. Similarly considering in mortality Kerala tops the rank due to its lowest neonatal mortality rate in the country because of its rigorous training programmes of healthcare workers and inherent superior healthcare infrastructure. Compare to other state like Madhya Pradesh and Uttar Pradesh 69% and 48% of the institutional deliveries were highest neonatal mortality and infant mortality. This indicates poor quality of care received by a new born after childbirth even at health centres. Home-delivered children in Chhattisgarh and Jharkhand were found to have high neonatal mortality rate due to lack of medical attention at home.
Another significant finding of the study was the identification of neo-natal and infant mortality in both public and private hospitals. Public hospitals have a higher mortality rate than private hospitals, but the risk of death in private hospitals is relatively higher. It is mainly due to the pregnancies resulting in complications from prolonged labour and prematurity are often referred to medical institutions. However, lack of well-equipped infrastructure to respond to crisis, non-availability of obstetric emergency care, lack of transportation facilities to deal with emergency cases add to the loss of lives in health private institutions [22]. On the contrary higher mortality in public hospitals due to high coverage birth in public institutions compared to private institution and another hand delay in decision to seek care, delay in reaching care and delay in receiving adequate health care for new babies by doctor. Sometime they do not properly visit in hospital; they are engaged with various private institutions for pecuniary benefits and hand over the duty to new untrained doctors and nurses. For these delays might result in reporting of institutional neonatal death. Previous studies found that delivering a public facility was not a protective factor against early neonatal mortality [31,32] and also found that delivering in a public hospital increased the odds of having a neonatal death [32]. However, the findings of present research suggest that delivering in public facility was protective against neonatal mortality even though it has been argued that limited access to emergency obstetric/neonatal care may be reasons for the lack of association between public facility delivery and higher neonatal survival in other settings [32].
The analysis presented in this paper has several limitations. First, the present study s analysis did not assess the quality of emergency obstetric care, new-born care in facilities, or community-based new-born care training, all of which are important distal factors influencing neonatal outcome [17,33,34]. Furthermore, the study did not evaluate staffing and quality of facilities, which are of central importance for realizing the potential health benefits associated with institutional deliveries and understanding the often remarkably low utilization rates. Finally, neonatal deaths were slightly underrepresented in the study, because multiple births and new-born whose mothers have died were not included in the study.

Conclusions
The study emphasized that safe delivery is associated with lower risks of neonatal deaths, but increases in institution-based deliveries were not consistently associated with decreased neonatal mortality rates. Despite increase in delivery facilities, neonatal and infant mortality rates remain high in these settings. The findings of the study have significant implications for new policies aimed at health facilities in order to reduce mortality. Improving the quality and safety of institutional deliveries has the potential to increase the utilisation and health impact of perinatal health services, as well as improve the health status of newly born Indian babies.