We assessed national and state trends in mean child mortality rates, CI, and HAI in India during 1992, 1998, and 2005, and tested for changes in mean child mortality rates and CI. Results of this secondary analysis are presented here.

We used household survey data from three waves of India’s National Family Health Survey (NFHS) (1992-1993, 1998-1999, and 2005-2006) [8-10] for the empirical analysis. For convenience, we henceforth refer to these periods as 1992, 1998, and 2005.

We estimated mortality rates in children age 5 and younger, the CI, and the HAI and computed separate estimates for rural and urban areas using the Synthetic Cohort Life Table approach [11]. Following Rutstein and Rojas, we adopted the following age segments to calculate the individual probabilities of dying (completed ages, in months): 0, 1-2, 3-5, 6-11, 12-23, 24-35, 36-47, and 48-59. This approach allows full use of recent data and is specific to the periods.

At national and state levels, we used the wealth index to compute the CI, as described in the literature [12]. The CI can range in value from -1 to +1, with 0 indicating no wealth-related health inequality. Negative CI values represent ill health in that they indicate a disproportionate burden of mortality among the poorest sections of society; the larger the value, the greater the inequality. The product of the mean mortality rates and the CI yields the HAI, a weighted average of the health levels of people in the sample, in which higher weights are attached to poorer people [13]. While the concentration curve is not a convenient measure of inequality for comparison, the CI, which is based on the concentration curve, quantifies the degree of socioeconomic inequality in a health variable [6,7].

We computed t-statistics to test for statistically significant differences in the mean mortality rates and inequalities (CI) between 1992 and 1998 and 1998 and 2005 at national and state levels. For the three states formed in 2000 (Uttarkhand, Jharkhand, and Chhattisgarh), we used estimates from the respective "parent" states (Uttar Pradesh, Bihar, and Madhya Pradesh) for the earlier years.

Overall, with the exception of a few small states, mean child mortality rates declined at national and state levels between 1992 and 1998 and 1998 and 2005 (Table I, Fig. 1). Nationally, mean child mortality rates declined significantly between 1992 and 1998 and 1998 and 2005. Mean child mortality for the 10 states that together account for roughly 60% of the annual birth cohort, also declined. However, changes in inequality were not uniform. Nationally, inequalities increased slightly between 1992 and 1998 and declined slightly between 1998 and 2005, with similar trends in rural India; there was a statistically significant decline in mean mortality for both urban and rural India and a corresponding decline in HAI between 1992 and 1998 and 1998 and 2005. Urban areas had lower child mortality rates than rural areas, but inequalities (CIs) were higher in urban areas in all three years.

U5MR, mean mortality rates in children aged 5 years or younger

Fig 1. Mean mortality rates in children aged £5 years, inequality, and health achievement indices in India, nationwide and in rural states and urban populations, 1992, 1998, and 2005.

Results were mixed for the states with the largest birth cohorts (Web Fig.1). Although HAI decreased uniformly, few had statistically significant changes in inequality. Among all states, mean child mortality, inequalities, and HAI varied significantly (Table I). In 1992, eight states had mean child mortality rates and HAI higher than the national average of 108.75/1,000 live births and 127.28/1,000 live births, respectively, while 19 states had inequalities worse than the national average of -0.170.

TABLE I	State-level Trends in Mean Mortality Rates of Children Aged 5 Years or Younger, Health Achievement, 

and Inequality in India, 1992, 1998, and 2005.

Child mortality varied significantly among the states, with the lowest rates in states like Kerala (32/1,000 live births) and the highest in states like Uttar Pradesh (141/1,000 live births). In 1998, 10 states had mean mortality rates and HAIs higher than the national average of 94.25/1,000 live births and 112.31/1,000 live births, respectively, while 23 states had inequalities worse than the national average of 0.192. In 1998, child mortality rates declined in both Kerala (19/1,000 live births) and Uttar Pradesh (122/1,000 live births). In 2005, nine states had mean child mortality rates and HAIs higher than the national average of 74.1/1,000 live births and 87.22/1,000 live births, respectively, while 19 states had inequalities worse than the national average of -0.178. Child mortality rates in Kerala declined further in 2005 to 16/1,000 live births, while those of Uttar Pradesh declined to 96/1,000 live births.

Bihar, Madhya Pradesh, Rajasthan, Uttar Pradesh, and the recent additions Chhattisgarh, Jharkhand, and ttarkhand. Chhattisgarh, Madhya Pradesh, and Rajasthan fared marginally worse on all dimensions of health achievement between 1992 and 1998 (Table I), but the differences were not statistically significant except for the change in inequality for Rajasthan. Between 1998 and 2005, all the BIMARU states except Bihar had statistically significant declines in child mortality; both Bihar and Uttar Pradesh had statistically significant declines in inequality. The exception was Uttarkhand, where inequality increased between 1998 and 2005. Between 1992 and 1998, five states had statistically significant increases in child mortality; 12 states showed a statistically significant decline in child mortality. Four states had worse inequalities in 1998 than in 1992, while three improved in this measure. All states had lower child mortality in 2005 than in 1998, with statistically significant declines in 17 states. However, six states had statistically significant higher inequalities in 2005, while four had lower inequalities.

India’s national child mortality rates mask the vast underlying differences in child health achievement in urban and rural areas among the states and along the socioeconomic gradient. In 2006, India had a birth cohort of more than 27 million. By aggregating data over such large cohorts, with significant socioeconomic, geographic, and political differences between states, we lose important information [14]. Detailed information regarding the performance of the individual states is important to enable policy prescriptions subnationally.

Although child health achievement improved nationally between the periods we compared, the mean child mortality rate and CI presented stark contrasts. While mean child mortality rates have declined significantly, inequalities have not declined commensurately. However, nationally, declines in child mortality were not necessarily accompanied by worsening inequality. Although mean child mortality rates declined in both rural and urban India, significant rural-urban differences persisted, with 50% to 60% higher rates in rural areas. Lack of potable water and sanitation infrastructure is a major cause of infant and child mortality in rural areas [3,14].i>

IIn India, health care is the responsibility of state governments [15]. Although mortality rates declined across all states during 1992 to 2005, a slight increase in some states in 1998 was associated with the drop in public health care expenditure (PHCE). In 1990, India’s PHCE was 1.3% of gross domestic product (GDP) but had declined to 0.9% by 1999 [16]. Among the states, Bihar and Uttar Pradesh had the lowest PHCE. However, the more troubling trend is the widening of the gap in child mortality across states. Increasing the gross state domestic product (GSDP) (and the corresponding PHCE) of the states lagging in child health achievement is not easy; according to Ahluwalia, [17] Bihar, Uttar Pradesh, and Orissa had poor GSDP growth in the 1990s. Bhat and Jain, [15] have shown that all state governments have set targets for PHCE expenditures as a fraction of their GSDP, but they are inadequate to achieve the United Nations Millennium Development Goals.

Nationally, we found no evidence of significant improvements in wealth-related inequalities in child mortality rates. At the state level, inequalities increased in more places than they decreased during 1992 to 1998 and 1998 to 2005. In addition, the poorest performing states failed to reduce wealth-related inequalities in child mortality, an important finding that suggests that improvements in economic opportunities do not necessarily lead to improvements in health care infrastructure [18]. The increasing inequality in 1992 to 1998 corresponds to the general contraction of public expenditures between 1992 and 1996 [15]. /p>

AAlthough previous studies indicate that economic growth is associated with declines in average child mortality rates nationwide, [19-21] this association may not hold subnationally. Specifically, Maharashtra and Gujarat were ranked just below Punjab and Haryana in income in the early 1990s, but post 1991, their income accelerated at the fastest pace; [22] both states demonstrated consistent declines in mean child mortality, but inequality increased in Maharashtra in these rates during 1998 to 2005. The premise that economic growth increases inequality of health outcomes subnationally is reinforced with data from Tamil Nadu; the declines in child mortality in this state were accompanied by an increase in inequalities. Tamil Nadu had one of the highest income growth rates in the 1990s.

In light of our findings, the government launch of the National Rural Health Mission (NRHM) and the National Urban Health Mission (NUHM) in 2005 is particularly timely. The NRHM was launched [22] to improve rural health by targeting phased increases in funding for the health sector of 2% or 3% of the GDP. The immediate expected outcome is reduced child mortality. India also developed NUHM to provide accessible and affordable health care to nearly 220 million of India’s urban poor in 429 cities. Our findings suggest that the NRHM and NUHM could identify the districts with the highest number of child deaths in the high-mortality states and target them with low-cost interventions like vaccinations, oral rehydration salts, and vitamin A supplementation. /p>

DDespite reductions in national child mortality in India in 1992 to 2005, significant rural-urban, interstate, and intrastate disparities remain. In addition, although we found no conclusive evidence of a tradeoff between "goodness" and "fairness," the results indicate that, on average, declines in child mortality rates do not automatically reduce wealth-related inequalities in these rates.

Funding: None; Competing interests: None stated.

2. Wagstaff A, Bustreo F, Bryce J, Claeson M. Child health: reaching the poor. Am J Public Health. 2004;94:726-36.

3. Victora CG, Wagstaff A, Schellenberg JA, Gwatkin D, Claeson M, Habicht JP. Applying an equity lens to child health and mortality: more of the same is not enough. Lancet. 2003;362:233-41.

4. Marmot M. Social determinants of health inequalities. Lancet. 2005;365:1099-104./p>

55. O’Donnell O, Doorslaer Ev, Wagstaff A, Lindelow M. Analyzing Health Equity Using Household Survey Data: A Guide To Techniques And Their Implementation. Washington, D.C.: World Bank; 2008.

6. Kakwani N, Wagstaff A, van Doorslaer E. Socioeconomic inequalities in health: Measurement, computation, and statistical inference. Journal of Econometrics. [doi: 10.1016/S0304-4076(96)01807-6]. 1997;77:87-103.

7. Wagstaff A, Paci P, van Doorslaer E. On the measurement of inequalities in health. Social Science & Medicine. [doi: DOI: 10.1016/0277-9536(91)90212-U]. 1991;33:545-57.

8. International Institute for Population Sciences (IIPS), Macro International. National Family Health Survey (NHFS-1), 1992-93: India. 1994; Mumbai, India: IIPS.

9. International Institute for Population Sciences (IIPS), Macro International. National Family Health Survey (NHFS-2), 1998-99: India. 2000; Mumbai, India: IIPS.

10. International Institute for Population Sciences (IIPS), Macro International. (2007). National Family Health Survey (NHFS-3), 2005-06: India. 2007; Mumbai, India: IIPS.

11. Rutstein S, Rojas G. Online Guide to DHS Statistics. 2006; Available from: http://www.measuredhs.com/help/Datasets/whnjs.htm.

12. Kakwani N. Measurement of tax progressivity: An international comparison. Economic Journal1977;87:71-80./p>

13. Wagstaff A. Inequality aversion, health inequalities and health achievement. J Health Econ2002 Jul;21(4):627-41.

14. Dholakia R, Kumar A, Datta S. Millennium Development Goals needs assessment at state level in India: A study of Madhya Pradesh, Rajasthan and Uttar Pradesh. Background paper for the UN Millenium Project Task force on Project & Economic Development2004.

15. Bhat R, Jain N. Analysis of public expenditure on health using state level data. Ahmedabad: Indian Institute of Management2004 Contract No.: WP2004-06-08.

16. National Rural Health Mission. Government of India; [cited 2009 January 2]; Available from: http://mohfw.nic.in/NRHM.htm.

17. Ahluwalia M. Economic reforms in India since 1991: Has gradualism worked? Journal of Economic Perspectives2002;16(2):67-88.

18. Parker B, Kozel B, Kukerja M. In search of a chance: Urban opportunities, poverty and vulnerability in northern India. Urban Research symposium: Urban Development for Economic Growth and Poverty Reduction; December 15-17, 2003; Washington, D.C.: The World Bank; 2003.

19. Alderman H, Appleton S, Haddad L, Song L, Yohannes Y. Reducing child malnutrition: how far does income growth take us? Washington DC: World Bank2000.

20. Preston SH. The changing relation between mortality and level of economic development. Popul Stud (Camb)1975 Jul;29:231-48.

21. Pritchett L, Summers L. Wealthier is healthier. Journal of Human Resources1996;31:841-68.

22. Ahluwalia MS. State-level performance under economic reforms in India. Conference on Indian Economic Prospects: Advancing Policy Reform; Stanford University, Palo Alto, CA: Center for research on Economic Development and Policy Reform; 2000.