Introduction

The Covid-19 pandemic has once again highlighted the importance of accurately counting deaths, and of classifying them by age, sex, place of residence and cause, so that health officials and decision-makers can identify health threats and high-risk populations (Sakia et al., 2023; Sankoh et al., 2020).

India, the second most populated country of the world, enacted its Registration of Birth and Death (RBD) act in 1969, mandating universal registration within 21 days of a birth or death. Today, however, things are still far from perfect. For instance, in 2020, there were around 5.6 million and 4.6 million deaths, respectively, among Indian males and females according to United Nations estimates (United Nations, 2022), but the Office of Registrar General of India (2022) “missed” some 6% of female deaths in that year.

This calls for an understanding of the determinants of death registration in India. This is the topic of our recent study (Saikia et al., 2023) based on 84,390 individuals interviewed during the National Family Health Survey, 2019-21 (NFHS-5) (IIPS & ICF, 2021). Note, however, that we relied on interviewees’ reports of family deaths and their registration in the previous year, without checking death certificates: errors of various types are therefore possible. Besides, part of the data collection took place in the post-Covid period. Due to the government compensation scheme for Covid-related deaths, higher-than-normal death registration is likely to have occurred.

Death registration level varies by place of residence, region, state and district

Between 2019 and 2021, according to our sample data, slightly less than 71% of deaths were registered, on average, with higher rates in urban than in rural areas (83% vs 66%, respectively). In the latter, most adults are employed in informal sectors such as farming, construction work and fishing, which seldom provide social security (e.g. survivor’s pension). Low registration levels may reflect a lack of motivation, few incentives and poor access to death registration services. 

Death registration coverage varies geographically, and is particularly low in the Northeast (64%), East (57%), and Central (56%) Indian regions (Figure 1). 

Death registration level vary by demographic and socioeconomic characteristics

Registration levels are higher for deaths at older ages, e.g. about 78% at 50-64 years, and 72% at 65-98 years (Figures 2 and 3). Conversely, they are very low for children: merely 35% at 0-4 years. Sex differences are large: around 74% of male deaths are registered versus 66% of female deaths. This difference is linked, among other things, to two factors: 

1) a lower proportion of women employed in the formal sector (see above) and

2) a higher proportion of accidental deaths among males. These are usually subject to police investigation, which increases the likelihood of registration (Adair et al., 2021).

Death registration was lower: 

1) among the low-educated: 63% and 71% when the household head was, respectively, illiterate or had just primary education,

2) among Muslims (65%) than Hindus (71%),

3) among scheduled castes (or STs: 67%) than other castes (77%),

4) among poorer households (52% in the first income quintile; 87% in the fifth). 

Conclusion 

Death registration levels vary considerably by sex, age, place of residence and regions in India, and are influenced by socioeconomic variables. Periodic awareness programs may therefore be needed among vulnerable subgroups and in disadvantaged districts. Providing financial assistance for funeral rites, education loans to orphans, and social security to the deceased’s family members after reporting a death to a civil authority would probably prove helpful in increasing death registration coverage.

References

Adair T, Gamage USH, Mikkelsen L, Joshi R. 2021. Are there sex differences in completeness of death registration and quality of cause of death statistics? Results from a global analysis. BMJ Glob Health. 6(10):e006660. doi:10.1136/bmjgh-2021-006660

International Institute for Population Sciences (IIPS) and ICF 2021. National Family Health Survey (NFHS-5), 2019-21. IIPS: Mumbai, India. 

Niti Aayog – SDG India Index. 2021.; 2021.

Saikia N, Kumar K, Das B. 2023. Death registration coverage 2019–2021, India. Bulletin of the World Health Organization. Feb 1; 101(2):102-10. doi: http://dx.doi.org/10.2471/BLT.22.288889 

Sankoh O, Dickson KE, Faniran S, et al. 2020. Births and deaths must be registered in Africa. Lancet Glob Health. 8(1):e33-e34. doi:10.1016/S2214-109X(19)30442-5.

Setel PW, Macfarlane SB, Szreter S, et al. 2007. A scandal of invisibility: making everyone count by counting everyone. The Lancet. 370(9598):1569-1577. doi:10.1016/S0140-6736(07)61307-5

United Nations. Department of Economic and Social Affairs, Population Division 2022. World Population Prospect. 2022.

Source fugure 1: https://figshare.com/articles/figure/Death_Registration_Level_by_Sex/21724049

The post Why and where are deaths under-reported in India? appeared first on N-IUSSP.

The literature on “missing women” in Asia tends to focus on the female deficit due to prenatal sex selection and underplays the role of postnatal sex selection. One reason is that there is a sort of fascination with the idea that modern technologies can be used to implement ancient patriarchal forms of bias against girls and women.

Another reason is more technical. It is easy to evidence skewed sex ratios at birth with basic birth data, whereas signs of postnatal discrimination are far more blurred, at least ever since deliberate selective infanticide has disappeared from China and India. This is primarily due to the absence in many countries of reliable civil registration data that would otherwise allow for easy computation of mortality rates by age and sex. In addition, the very concept of “excess female mortality” and the way to calculate it are still debated, since existing series demonstrate that, in “normal conditions” (i.e., without discrimination), the expected male-female mortality ratio varies greatly by age and overall mortality level.

As a result, estimates usually require the use of heavy statistical machinery and lead to a single estimate of excess female deaths for an entire country (Alkema et al. 2014; Bongaarts and Guilmoto 2015; Costa et al. 2017). However, a single number cannot adequately reflect the situation of a country as diverse as India, where history and geography have shaped very different regional gender systems.

Estimating excess under-five female mortality

This consideration led us to focus on the existing demographic datasets describing India’s 640-odd districts. Among these, an often neglected source is the decadal census, and in particular the detailed, district-level fertility tables of its 2011 edition, which detail the number of newborns by sex and mother’s age (Fertility Series data F1, F5 and F9; Guilmoto et at 2018). We decided to test the strength of the classical Brass method, in which the proportion of surviving children by mother’s age can be used for estimating past infant and child mortality, taking into account age at childbearing, which varies by district. Under-five mortality is an ideal measure, since earlier studies have shown that most of the excess female mortality is concentrated during the first years of life, a period during which girls have a distinct biological mortality advantage over boys.

The procedures involved quite a few stages:¹ here, let us focus on the results. We found that the average level of excess mortality in girls aged 0–4 in 2000–2005 was no less than 18.5 per 1000 live births, translating into an average of 239,000 excess deaths of young girls per year. Around 22% of the overall female under-5 mortality is therefore due to gender bias. Incidentally, India narrowly missed Millennium Development Goal 4 on child mortality (about 2 units per 1000 live births). This implies that without excess female mortality, India’s Millennium Development Goal of reducing under five mortality (both sexes combined) from 125 per 1000 live births in 1990 to 42 by 2015 could have easily been achieved.

A map of gender discrimination

The situation at district level is depicted in Figure 1. What emerges, in particular, is the strong spatial patterning of excess female mortality on the Ganga River Basin in North central India. There is a large cluster of almost 60 adjacent districts in Madhya Pradesh, Uttar Pradesh, and Rajasthan where excess female mortality exceeds 30 per 1000 live births. In contrast, gender discrimination is much lower, even if not totally absent, in many states in the South and East of the country.

Beyond their unique geography around the Hindi belt, the districts where postnatal discrimination towards girls takes the heaviest toll in India tend to be rural, agricultural districts, characterized by high population density, high fertility, low social development and high son preference. On the contrary, excess female mortality weakens with increasing female education and the substantial presence of Muslim and tribal populations.

To understand the dynamics of gender bias, it is worth underlining that this map does not coincide at all with the more prosperous districts to the West of the country, from Punjab to Maharashtra, where prenatal sex selection is rife. Son preference and large families go hand in hand with excess female mortality. In contrast, the rapid fertility decline in Western India where the preference for male offspring remains pervasive has been accompanied by a skewed sex ratio at birth visible since the 1990s.

Conclusions

Excess female mortality resulting from gender discrimination in the postnatal period is a rarely studied phenomenon, and this, in our opinion, confers a special interest to our study and to Figure 1. It also reveals the distinct contours of the old discriminatory gender regime, characterized by high fertility and strong discrimination towards young girls. As the regional estimates of excess deaths of girls demonstrate, any intervention to reduce the discrimination against girls in food and health care allocation should therefore target in priority the regions of Bihar and Uttar Pradesh where poverty, low social development, and patriarchal institutions persist and investment in girls is limited.

The transition from postnatal to prenatal sex selection across districts is likely to extend soon to the large states of North India featuring on our map, where fertility is currently declining rapidly and economic progress is reaching even remote rural districts. This does not bode well for the future of the sex ratio at birth in India, and reinforces the need to address directly the issue of gender discrimination in addition to encouraging social and economic development that benefits Indian women.

References

Alkema L, Chao F, You D, Pedersen J, Sawyer CC. (2014) National, regional, and global sex ratios of infant, child, and under-5 mortality and identification of countries with outlying ratios: a systematic assessment. The Lancet Global Health. September 30;2(9):e521-30.

Bongaarts J, Guilmoto CZ. (2015) How many more missing women? Excess female mortality and prenatal sex selection, 1970–2050. Population and Development Review. June 1;41(2):241-69.

Costa JC, da Silva IC, Victora CG. (2017) Gender bias in under-five mortality in low/middle-income countries. BMJ Global Health. July 1;2(2):e000350.

Guilmoto CZ, Saikia N, Tamrakar V, Bora J (2018). Excess under-5 female mortality across India: a spatial analysis using 2011 census data. Lancet Global Health.

Foot notes

¹ We first obtained a series of under-five mortality rates by sex and district. We then used a set of existing life-tables drawn from countries with no gender bias to model the relationship between the sex ratio of under-five mortality and overall child mortality. Finally, we applied these standardized mortality sex ratios to our mortality estimates to derive excess under-5 female mortality for each district of India, and then converted this information into absolute numbers of excess female deaths. For more details, see Guilmoto et al. (2018).

The post Counting excess under-5 female mortality in Indian districts appeared first on N-IUSSP.

A large part of under-five mortality is due to infant mortality, which is also declining but remains high: the most reliable data of the Sample Registration System (SRS) show that India’s infant mortality rate (IMR) decreased from 129 to 40 (per thousand live births) between 1971 and 2013.

Finally, as a large, and recently increasing, share of infant and under-five mortality is due to neonatal mortality, concern about this very early mortality has intensified in India. Previous studies focused mainly on individual-level determinants and the medical causes of neonatal deaths. But it is also worth considering the macro dimension of the problem, and to examine the trend, tempo, and quantum of changes in the neonatal mortality rate (NMR) at state level, assessing the amount of inequality across all states and between their urban and rural subpopulations (Saikia et al. 2016).

Regional patterns and the unequal speed of reduction

India’s regional pattern of neonatal mortality has remained almost unaltered in the past 30 years, with the centre-east lagging behind the south (Figure 1).

While the NMR generally declined over the period, a huge gap remains between the vanguard state of Kerala and laggard states such as Uttar Pradesh, Madhya Pradesh and Odisha (Figure 2). Besides, the NMR decline slowed down between 1993 and 2000 and again between 2003 and 2007, but not everywhere and, in particular, not in the medium-income group of states.

Note also that there is no simple correspondence between the socioeconomic gradient and neonatal mortality: for instance, the NMR has traditionally been very similar in medium- and high-income states, and since the early 2000s, it has been even lower in the group of medium-income states.

Persistent rural-urban divide

Although the absolute magnitude of the rural excess in NMRs has diminished greatly, the rural NMR remains twice as high as in urban areas (Figure 3). Throughout the period under consideration, the rural NMR was higher than the urban NMR in each state, with the sole exception of Kerala in the 1980s (Figure 3). The variation in NMRs across states was higher for the rural population.

The NMR of the high-income states has been higher than that of middle-income states since the 2000s. This is interesting, since the NMR of high-income states has not yet reached the minimum level of NMR experienced in some other Indian states. The highest absolute decrease in rural NMR occurred in middle-income state Tamil Nadu, followed by the low-income state Uttar Pradesh. Tamil Nadu also experienced the highest relative decrease in rural NMR, followed by the state of Kerala. Literatures suggest that differential levels of social-political development (including women’s autonomy and rights, literacy, etc.) and differential accessibility and functioning of the health care system widen the NMR differences between states.

Did geographical disparities in NMR decrease over time?

Table 1 presents two types of measures of absolute and relative disparity with respect to NMR across 30 sub-populations (urban and rural populations in 15 states). These are simple range measures (Max-Min), the Lorenz curve-based dispersion measure of mortality (DMM) and the Gini coefficient (G). While the range measures depend on the two extreme NMR values, the DMM and G measures express the absolute and relative inequality across the entire distribution. DMM has decreased over time, but it stabilized between 2001 and 2011; G decreased from 1991 to 2001, but then increased again in 2011 to a level higher than in 1981. In short, absolute geographical disparity in NMRs has shrunk over time, but relative disparity has instead increased in the last decade.

What do we learn from this study?

Despite falling by more than half over the past decades, the national NMR in India in 2011 was still ten times higher than the corresponding figure in urban Kerala or in the majority of developed countries. Temporal NMR decreases have been uneven across sub-populations: both high- and low-income groups of states were unable to maintain the pace of NMR reduction of the 1980s: they experienced episodes of stagnation or even marginal increase, which may reflect internal (i.e., between individuals, within states) economic inequality, which is on the rise. Also worrying is the persistence of the NMR disparity across the regional and urban-rural divide.

In short: in terms of NMR, as well as in a number of other dimensions in India, progress is under way, but much remains to be done.

References

Saikia, N., Shkolnikov, V. M., & Jasilionis, D. (2016) Trends and Sub-National Disparities in Neonatal Mortality in India from 1981 to 2011. Asian Population Studies, 12(1), 88-107.

UNICEF (2016) The state of the World’s children 2016. Accessed on 14^th July 2016.

¹ Under-five (or child) mortality refers to deaths in the first five years of life; infant mortality to deaths in the first year of life; neonatal mortality to deaths in the first 28 days of life.

The post Neonatal mortality trend in Indian states, 1981-2011 appeared first on N-IUSSP.