|Year : 2019 | Volume
| Issue : 1 | Page : 27-32
Determinants of infant mortality in rural India: An ecological study
Abhijit Mukherjee1, Sharmistha Bhattacherjee1, Samir Dasgupta2
1 Assistant Professor, Department of Community Medicine, North Bengal Medical College, Darjeeling, West Bengal, India
2 Professor, Department of Community Medicine, North Bengal Medical College, Darjeeling, West Bengal, India
|Date of Web Publication||12-Mar-2019|
Dr. Sharmistha Bhattacherjee
Department of Community Medicine, North Bengal Medical College, Sushrutanagar, Darjeeling - 734 012, West Bengal
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background: Long-term reductions in infant mortality (IM) are possible only by addressing the distal determinants. Objectives: The objective of the present study was to determine the relationship between IM and its major distal determinants in rural India. Methods: The dependent variable used in the study was state wise IM rate (IMR), the values of which were obtained from the Sample Registration System, 2015. State level literacy rate in females, unemployment rates of females, GINI index, and round-the-clock neonatal services in primary health centers in the rural areas and the per capita gross state domestic product at purchasing power parity (GSDP at PPP) of the states, were used as the predictor variables for IM. Relationship between the variables was obtained by the Pearson's correlation coefficient. Bivariate and multivariable linear regressions were used to identify the magnitude and direction of the predictors on IM. Results: Correlation statistics showed none or weak positive correlation between the Gini coefficient and 24 × 7 primary health-care services and IMR. There was a strong negative correlation between female literacy rate and IMR, while the unemployment rates and per capita gross state domestic product (GSDP) were moderately negatively correlated to IMR. Bivariate analysis revealed that, for unit increase in unemployment rates in females, proportion of literate women, and 1000$ increase in the GSDP at current prices, IMR decreased by 0.07, 0.763, and 1.702, respectively. However, after adjustment, only the female literacy rates showed significant association with IMR. Conclusions: Of the major determinants included in the study, rural female literacy is the most important distal determinant of IM in rural areas of India.
Keywords: Ecology, rural, India, infant mortality, linear regression
|How to cite this article:|
Mukherjee A, Bhattacherjee S, Dasgupta S. Determinants of infant mortality in rural India: An ecological study. Indian J Public Health 2019;63:27-32
|How to cite this URL:|
Mukherjee A, Bhattacherjee S, Dasgupta S. Determinants of infant mortality in rural India: An ecological study. Indian J Public Health [serial online] 2019 [cited 2019 May 25];63:27-32. Available from: http://www.ijph.in/text.asp?2019/63/1/27/253895
| Introduction|| |
Infant mortality (IM) is an indicator of the health of a population and a measure of health inequalities within it. While a few urban areas of India have reached very low IM rates (IMRs), these indicators remain significantly higher in rural areas. This phenomenon of increased IM in rural areas, termed the “urban bias,” has also been observed in many studies done in the developing countries.,,
Sample Registration System (SRS) data published in December 2016 show that significant differences still persist in the IMR in rural and urban areas across all states, as was seen four decades ago. While it is as high as 54 in rural Madhya Pradesh, it is only 7 in urban Puducherry. This urban–rural division is of greater significance in India because approximately 70% of its population live in the villages.
According to the conceptual model proposed by Chen and Mosley, distal ecological factors operate through intermediate factors on the proximate determinants such as maternal nutrition, infections, and others to affect IM. Immediate decreases in IM can be achieved by addressing the proximal causes. However, long-term achievements can only be sustained through the improvement of the distal determinants [Figure 1].
|Figure 1: Conceptual framework of infant mortality (based on the Chen and Mosley model).|
Click here to view
Studies worldwide have shown per capita income, labor force participation of women, literacy rates in women, and wealth inequality to be the important distal socioeconomic predictors of IM.,, In addition, the availability of round-the-clock neonatal health-care services, an important social indicator, influences IM considerably. These determinants vary widely between the urban and rural areas as also between rural areas of the different states of India. Such regional differences along with state-level decision-making in health care make it necessary to identify the ecological factors and to quantify the effect of these factors on IM in these rural areas of India.
| Materials and Methods|| |
An ecological study was designed to evaluate the relationship between rural IMRs in states of India with important distal determinants.
IMR, the outcome variable of the study, is defined as the ratio of infant deaths registered in a given year to the total number of live births registered in the same year; usually expressed as a rate per 1000 live births. Data on IMRs for the current study were obtained from the SRS 2015, the results for which were published in December 2016. The SRS is a large-scale field-level continuous enumeration of births and deaths, done every 6 months since 1964–1965 for providing reliable annual estimates of birth rate, death rate, and other fertility and mortality indicators at the national and sub-national levels. The sample unit in rural areas is a village or a segment of it, whereas in urban areas, the sampling unit is a census enumeration block. The SRS sample is replaced every 10 years based on the latest census frame. The current sample is based on the 2011 Census frame and is effective from January 2014. The present SRS has been done in 8854 sample units (4962 rural and 3892 urban) spread across all states and union territories (UTs).
Unemployment rates (per 1000) for women of age 15 years and above, in rural areas, according to the usual status approach (usual principal and subsidiary status) for each state/UT were obtained from the second annual survey of employment and unemployment by the Labour Bureau of the Government of India, 2015–2016.
Literacy status indicates the total percentage of the population of an area at a particular time aged 7 years or above who can read and write with understanding. Data on literacy rates of rural females for the present study were obtained from the District Level Household and Facility Survey (DLHS 4). The DLHS is a countrywide assessment on family planning, maternal and child health, reproductive health of ever-married women and adolescent girls, and utilization of maternal and child health-care services at the district level for India.
The present study used the Gini coefficient in the rural areas of the Indian states for measuring economic inequality. The Gini coefficient measures the extent to which the distribution of income (or, in some cases, consumption expenditure) among individuals or households within an economy deviates from a perfectly equal distribution. A Gini coefficient of 0 means perfect equality, where everyone has the same income. A Gini coefficient of 1 (or 100%) means perfect inequality where only one person has all the income and all the others have none. The Gini coefficients for the present study were obtained from the unofficial estimates of Planning Commission, 66th Round 2009–2010.
There are two ways to determining gross domestic product (GDP), the most comprehensive measure of national economic activity, GDP at exchange rate and GDP at purchasing power parity (PPP). Gross state domestic product (GSDP) is the state counterpart to a country's GDP. The estimates of the per capita GSDP at PPP (in 1000$) at current prices (2012–2013 Series) for the state in the present study have been obtained from reports of the Ministry of Statistics and Programme Implementation, Government of India, 2015.
Accessibility to neonatal care services in the rural areas is closely related to IM as it translates to early and appropriate care of infants in sickness. In the absence of indicators for assessing the accessibility of the neonatal care facilities, the present study uses the proportion of primary health care centres (PHCs) offering 24 × 7 newborn care services as obtained from DLHS 4 to represent accessibility to newborn services in the states.
The correlation between IMR and the predictor variables was assessed using the Pearson's correlation coefficient. Bivariate linear regression was done to assess the direction and magnitude of the effects of the predictor variables on IMR. Outliers were tested by Mahalanobis distance in multivariable regression program and leverage values. After ruling out multicollinearity (variance inflation factor <2), multivariable backward linear regressions were done with all the predictor variables, with 0.05 as the level of removal. All statistical analyses were carried out using IBM SPSS Statistics for Windows, Version 20.0. (Armonk, NY: IBM Corp). Graphs with 95% confidence intervals were built with the use of JMP®, Version 12.2.0. SAS Institute Inc., Cary, NC (trial version).
This study is based on the national-level survey data, which are freely available in the public domain with no identifiable information on the survey participants; therefore, this work was exempted from ethical review.
| Results|| |
The present study included 29 states and 3 UTs of India. The UTs of Chandigarh, Daman and Diu, Dadra and Nagar Haveli, and Lakshadweep were excluded from the analysis as data on all parameters were not available. IMRs ranged from 9/1000 live births in rural Goa to 57/1000 in rural Madhya Pradesh.
Scatter plots depicting the relationship between IMR and other predictor variables are presented in [Figure 2]. They show a none or weak positive correlation between the Gini coefficient (r = 0.169, P = 0.355) and 24 × 7 PHC services (r = 0.135, P = 0.462) and the IMR. A strong negative correlation was found between the proportion of female literacy (r = −0.691, P = 0.000) with IMR, whereas the unemployment rates (r = 0−.473, P = 0.006) and the per capita GSDP (r = −0.495, P = 0.004) were moderately negatively correlated to IMR.
|Figure 2: Scatter plots showing the relationships between income inequality, per capita gross domestic state product in 1000$ purchasing power parity, female unemployment, female literacy, and proportion of primary health-care centers offering 24 × 7 newborn care services and infant mortality rate (per 1000) in Indian states (rural). Gray area shows 95% confidence interval.|
Click here to view
On bivariate analysis, the study results indicate that, for unit increase in unemployment rates (per 1000), the IMR decreased by 0.07. For 1% increase in the proportion of literate women, IMR decreased by 0.763 and for a 1000$ increase in per capita GSDP (PPP), the IMR decreased by 1.702. For 1% increase in the proportion of PHCs providing 24 × 7 services, the IMR increased by 0.122. For 1 unit change in the Gini coefficient, IMR increased by 0.412 [Table 1].
|Table 1: Bivariate linear regression models on infant mortality rates across states of India|
Click here to view
Multivariable regression analyses with backward elimination method were conducted where all the five predictors were entered at the first step and, then at each step, variables with large P values were deleted sequentially. The first model with all the five predictors produced R2 = 0.476, F = 6.639, and P = 0.000, while the final model had only female literacy as the predictor with R2 = 0.401, F = 21.739, P = 0.000 [Table 2]. This means that the model with all the variables explained 47.6% of the variation in the data, while the female literacy-only model accounted for 40.1% of the variation in the data. Both the models were found to be significantly better than an intercept-only model (F, p).
|Table 2: Multivariable linear regression model on infant mortality rates in different states of India|
Click here to view
| Discussion|| |
Data on IM, unemployment, female education, income inequality, and the availability of 24 × 7 neonatal care services were obtained for the rural areas of the states. However, the total GDP of the state was used as the marker for assessing the prosperity of a state.
Demographic and health surveys and other important data sources suggest that IMR are significantly higher in rural than in urban areas. The fact remains true in both the developed and the developing world., Although successful in the short term, the overreliance on narrow disease specific programs targeting the proximate determinants may be one of the factors that has stalled progress in further reducing IM. It has been argued that long-term gains in reducing IM are only possible by implementing national programs targeting the distal determinants such as health and community development.
The GDP per capita and the income gap are considered the major determinants of the economic prosperity of a country or state. Developed countries or states have shown the per capita national income to be significantly associated with the IMR., However, in their study on 28 Indian states in 2011, Barker did not find any such significant association between the GDP or the per capita GDP and IMR. In the present study, we used the per capita GSDP in 1000 dollars (PPP) to identify states with a healthier economy. With increasing wealth of the state the IMR tends to decrease probably due to an increased state level funding for health activities in general and towards infant health in particular.
Unemployment is an important indicator of a country's economic success or failure., Female employment can affect IMR directly from a reduction in the effective time for child care activities due to the dual burden of employment and household work. Indirectly, lack of employment in women would lead to greater income inequalities in populations.
The greater the income gap between the rich and the poor, the poorer are the health outcomes. Tacke and Waldmann examined the effect of relative income on IM and concluded that IM can be reduced by reducing the income inequality. In their study on twenty Italian counties, Dallolio et al. found that both income and income inequality were associated with IM. The present study, however, failed to detect any significant relationship between IMR and GINI in the Indian states.
The availability and accessibility of a round-the-clock health delivery service, especially with neonatal care facilities, is an important social indicator determining IM. In the absence of availability of these services, care seeking may be delayed because of long distances or unavailability of transport from remote rural areas to appropriate care centers. A close association between health-care accessibility and IM has been seen in studies worldwide.,
Increase in the number of specialized neonatal care centers will result in increased handling of the complicated cases, thereby increasing the chances of infant survival. Resource utilization at these centers will be optimal in the presence of a robust referral system at the community level. The accredited social health activists at the village levels, functioning as frontline health-care delivery personnel, are responsible for the early recognition of childhood illnesses and referral to the PHCs and other higher centers for treatment. Since the whole referral system is still evolving, the impact of the neonatal care centers will only be apparent after a substantial period of time. Furthermore, since a large section of the infant deaths are early neonatal deaths, strengthening the institutional delivery system will add to a decrease in the overall IM in a state.
The female literacy rate has a trickle-down effect on IM because it has a catalytic influence on health-care utilization and acts as an additive and conjunctive ingredient to various other factors. States and UTs of India with higher literacy rates were found to have lower IMR. This difference in mortality persists even after adjusting for socioeconomic status and residence in rural and urban areas.
Female literacy has been shown to have the strongest impact on IM in data obtained from 177 developing and industrialized countries from around the globe including India. It has been shown that, in India, a 10% increase in overall literacy rate at the state or UT level will reduce IMR by 12/1000 live births. Women's education has often been cited as one of the most valuable tools to reduce poverty. Increase in the percentage of women with at least a complete primary education has been shown to decrease IM in studies from sub-Saharan Africa., Women who are educated tend to marry later and have smaller families. Women with adequate education could avoid teenage pregnancy and seek appropriate health care.
Our findings suggest that, even after adjusting for the income inequality and unemployment status, female literacy remains the most significant predictor of IM in rural areas. Female literacy alone can explain a signification part of the variability in the deaths of infants in rural areas of Indian states. An urban–rural difference has also been observed in education, especially women's education. Typically, urban populations have better access to schools and also enjoy better quality education. Mobilization of the state and central government machinery to ensure increased enrollment in school for the girl child in rural areas of India can increase the literacy rates and education level attained by women.
The study does not take into account the spatial dependence of the variables between the states. Although the researchers used the latest data available on the internet on the demographic and health surveys for the present study, time period delay (delay between collection of data and online publication of results) could not be avoided.
| Conclusions|| |
Long-term reductions in IM can be brought about by addressing the distal determinants of infant deaths. Of the major determinants assessed in the study, female literacy is the most important distal determinant of IM in rural India. The states will need to improve the female literacy rates as a priority for sustained and significant reductions in infant deaths.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Lau RS, Johnson S, Kamalanabhan TJ. Healthy life expectancy in the context of population health and ageing in India. Asia Pac J Public Health 2012;24:195-207.
Crenshaw E, Ameen A. Dimensions of social inequality in the third world: A cross-national analysis of income inequality and mortality decline. Popul Res Policy Rev 1993;12:297-313.
Balarajan Y, Selvaraj S, Subramanian SV. Health care and equity in India. Lancet 2011;377:505-15.
Yanping W, Lei M, Li D, Chunhua H, Xiaohong L, Mingrong L, et al.
A study on rural-urban differences in neonatal mortality rate in China, 1996-2006. J Epidemiol Community Health 2010;64:935-6.
Pham TL, Kooreman P, Koning RH, Wiersma D. Gender patterns in Vietnam's child mortality. J Popul Econ 2013;26:303-22.
Office of the Registrar General and Census Commissioner of India. Office of the Registrar General and Census Commissioner of India 2011. Houselisting and Housing Census Data Highlights – 2011. New Delhi: Government of India; 2011. Available from: http://www.censusindia.gov.in
. [Last accessed on 2018 Dec 07].
Mosley WH, Chen LC. An analytical framework for the study of child survival in developing countries. Popul Dev Rev 1984;10:25-45.
Schell CO, Reilly M, Rosling H, Peterson S, Ekström AM. Socioeconomic determinants of infant mortality: A worldwide study of 152 low-, middle-, and high-income countries. Scand J Public Health 2007;35:288-97.
Zakir M, Wunnava PV. Factors affecting infant mortality rates: Evidence from cross-sectional data. Appl Econ Lett 1999;6:271-3.
Gokhale MK, Kanade AN, Rao S, Kelkar RS, Joshi SB, Girigosavi ST. Female literacy: The multifactorial influence on child health in India. Ecol Food Nutr 2004;43:257-78.
Frankenberg E. The effects of access to health care on infant mortality in Indonesia. Health Transit Rev 1995;5:143-63.
Park K. Preventive medicine in obstetrics, paediatrics and geriatrics. Park's Textbook of Preventive and Social Medicine. Vol. 10. Jabalpur: Banarsidas Bhanot; 2015. p. 481-560.
International Institute for Population Sciences (IIPS). District Level Household and Facility Survey (DLHS-4), 2012–13. India, Mumbai: IIPS; 2014.
De Maio FG. Income inequality measures. J Epidemiol Community Health 2007;61:849-52.
Bildirici M, Ersin O, Koktener M. Genetic structure, consanguineous marriages and economic development: Panel cointegration and panel cointegration neural network analyses. Expert Syst Appl 2011;38:6153-63.
Sastry N. What explains rural-urban differentials in child mortality in Brazil? Soc Sci Med 1997;44:989-1002.
Gould WT. African mortality and the new 'urban penalty'. Health Place 1998;4:171-81.
Walsh JA, Warren KS. Selective primary health care: An interim strategy for disease control in developing countries. N Engl J Med 1979;301:967-74.
Claeson M, Gillespie D, Mshinda H, Troedsson H, Victora CG; Bellagio Study Group on Child Survival. Knowledge into action for child survival. Lancet 2003;362:323-7.
Barker L. Determinants of Child Mortality Across States in India. Unpublished Thesis for the Bachelor of Arts degree. Colorado: The Colorado College; 2011.
Shaw M, Galobardes B, Lawlor DA, Lynch J, Wheeler B, Davey Smith G. The Handbook of Inequality and Socioeconomic Position: Concepts and Measures. Great Britain: The Policy Press; 2007.
Kapoor S. Infant mortality rates in India: District level variations and correlations. In 6th
Annual Conference on Growth and Development 2010. p. 16-8.
Wilkinson RG. Socioeconomic determinants of health. Health inequalities: Relative or absolute material standards? BMJ 1997;314:591-5.
Tacke T, Waldmann RJ. Infant mortality, relative income and public policy. Appl Econ 2013;45:3240-54.
Dallolio L, Franchino G, Pieri G, Raineri C, Fantini MP. Geographical and temporal trends in infant mortality in Italy and current limits of the routine data. Epidemiol Prev 2011;35:125-30.
Kumar C, Singh PK, Rai RK. Under-five mortality in high focus states in India: A district level geospatial analysis. PLoS One 2012;7:e37515.
Politzer RM, Harris DL, Gaston MH, Mullan F. Primary care physician supply and the medically underserved. A status report and recommendations. JAMA 1991;266:104-9.
Ladusingh L, Gupta AK, Yadav A. Ecological context of infant mortality in high-focus states of India. Epidemiol Health 2016;38:e2016006.
Shetty A, Shetty S. The impact of female literacy on infant mortality rate in Indian States. Curr Pediatr Res 2014;18:49-56.
Saurabh S, Sarkar S, Pandey DK. Female literacy rate is a better predictor of birth rate and infant mortality rate in India. J Family Med Prim Care 2013;2:349-53. [Full text]
Gakidou E, Cowling K, Lozano R, Murray CJ. Increased educational attainment and its effect on child mortality in 175 countries between 1970 and 2009: A systematic analysis. Lancet 2010;376:959-74.
Rao RS, Chakladar BK, Nair NS, Kutty PR, Acharya D, Bhat V, et al.
Influence of parental literacy and socio-economic status on infant mortality. Indian J Pediatr 1996;63:795-800.
[Figure 1], [Figure 2]
[Table 1], [Table 2]