Neonatal mortality in India’s rural poor: Findings of a household survey and verbal autopsy study in Rajasthan, Bihar and Odisha

In 2011, Save the Children India launched a project for the disadvantaged population of Rajasthan, Bihar and Odisha. As a baseline activity, neonatal deaths during January–December 2012 were inves-tigated using modiﬁed verbal autopsy tool in six sub-district-level administrative units (blocks) adopting 30-cluster sample survey approach. Our study reported a total of 189 neonatal deaths of which 50% occurred at home and 39% happened on Day 1. About half of the deaths occurred in blocks from Bihar. High number of neonatal deaths belonged to households that were below poverty line (64%) and other disadvantaged classes (46%); among mothers who were illiterate (65%), < 20 years of age (54%) and during their ﬁrst-order births (36%). Birth asphyxia was a major cause of neonatal deaths across all blocks. These ﬁndings indicate need for easy and early access to transport services, specialized neonatal care and advocacy targeted towards increasing community awareness.


I N T R O D U C T I O N
India has made significant progress towards achieving the 2015 Millennium Development Goal (MDG) 4 target of reducing the under-five child mortality rate to less than 38 per 1000 live births with a 58% reduction from 126 per 1000 live births in 1990 to 53 in 2013. Neonatal deaths (in the first 28 days of life) as a proportion of all child deaths, increased from 40% in 1990 to 55% in 2012 [1]. The national neonatal mortality rate (NMR) in 2012 was 29 per 1000 live births. The overall NMR of 33 per 1000 live births in rural India is twice the rate of V C The Author [2015]. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com 16 in urban India [2]. Achievement of MDG 4 in India will depend on addressing neonatal mortality and focusing on the rural poor [3]. Government of India has identified 264 High Focus Districts (HFDs) based on indicators from district-level household survey including traditional social hierarchy, which remains a structural barrier to health equity in India, and the terms 'Schedule Caste', 'Tribal Caste' and 'Other Backward Class' are used to denote these traditionally marginalized and excluded communities. However, we prefer to use the term 'socially disadvantaged groups'. In 2011, Save the Children India (SCI), an international nongovernmental organization, launched a project to strengthen maternal, newborn and child health and nutrition services in six HFDs. The SCI intervention areas included one block in each district selected in consultation with respective state health missions. These districts (blocks) were Sitamarhi (Riga), Gaya (Mohanpur) in Bihar, Tonk (Urban Tonk), Churu (Rajgarh) in Rajasthan and Kandhmal (Phiringiya), Nuapada (Boden) in Odisha. Table 1 shows key health indicators in these six districts. In this article, we report on a situational analysis conducted at project baseline to determine number, time and causes of neonatal deaths that occurred in 2012.

M E T H O D S
A household survey was conducted in March 2013 using a cluster-sampling technique. Thirty villages were randomly selected in each block as primary sampling units (PSUs). All households in the PSUs were surveyed during the period January-December 2012. When a neonatal death was identified, verbal autopsies were conducted by trained health workers using a World Health Organization verbal autopsy tool adapted and modified to local context [4]. The cause of death was ascertained independently by two physicians using International Statistical Classification of Diseases and Related Health Problems, 10th revision. In case of disagreement, cases were solved through consensus.
Data variables included in the analysis were the neonate's sex, family characteristics (religion, socio-economic status, mothers education, age at first delivery, birth order) and place of death. Birth asphyxia and birth trauma were classified together as it was done in previous studies [5,6]. Data were entered in MS Excel, and analysed in SPSS (IBM, SPSS Statistics, Version 20). Ethics approval was granted by the Ethics Advisory Group, International Union Against Tuberculosis and Lung Disease, Paris, France.

R E S U L T S
A total of 189 neonatal deaths were reported of which 57% occurred at home, 12% occurred on the way to a facility and 39% occurred in first 24 hours of life. About two-thirds of neonatal deaths occurred in households below the poverty line, 90% belonged to disadvantaged groups, 65% of mothers were illiterate, 54% of mothers were less than 20 years of age and 36% were first-order births ( Table 2).

D I S C U S S I O N
This sub-national (block level) data on neonatal deaths in India provides several significant findings. First, our study corroborated known structural determinants of neonatal mortality in India-poverty, rural location, disadvantaged social status, low levels of maternal education, young maternal age at childbirth-and causes of neonatal deaths [7]. Second, while our study concords with other similar studies identifying birth asphyxia/birth injuries, as leading cause of death [8][9][10], the attributable proportion  (utilization) and limited access to health facilities. A high percentage of birth asphyxia/birth injuries may be due to lack of adequate routine or specialized newborn care at home, such as kangaroo care or neonatal resuscitation. The health literacy of parents may influence decision-making in utilization of health services. Access to health-care centres is an important determinant for reducing neonatal mortality [11]. Barriers to accessing care include geographical distance, financial barriers and socio-cultural factors (e.g. family decision-makers). Finally, the districts in Bihar contained 32% of the study population, but disproportionately accounted for approximately half of all neonatal deaths. Of the three study states, Bihar has the worst household economic status, literacy rates and proportion of deaths outside health facilities. This study provides valuable insights into neonatal mortality in India's rural poor, as district-level cause-specific neonatal mortality data are not available in India [12]. In the project areas, only 10% of deaths were registered and among these a death certificate was rarely issued. It was unlikely that neonatal deaths were missed as the study screened all sampled 180 villages through house-to-house survey. The study had a small sample population, which limits the generalizability of the findings. Recall bias may have affected the causes of death in the verbal autopsy as the participant recall period ranged from 3-15 months.

C O N C L U S I O N
This household survey and verbal autopsy study has identified a high proportion of neonatal deaths occurring outside health facilities and birth asphyxia/ injuries as a major cause of death in the SCI project areas in Rajasthan, Bihar and Odisha. These findings indicate the need for specialized newborn care, early referral and ambulance transport systems and fostering community-facility linkages.