Abstract: India’s endeavour to achieve the Sustainable Development Goals (SDGs) in a well-defined and time-bound manner is critical for national and global development. This paper examines India’s data availability to assess the SDGs related to health and nutrition. While India is still facing severe challenges of generating disaggregated information on mortality and cause-specific deaths, the desired data on nutrition and healthcare utilisation are largely available, enabling the computation of SDG indicators at the state or district level. However, indicators for socioeconomic groups cannot generally be obtained at levels of disaggregation below the state. There is scope to raise the preciseness of morbidity-related indicators in the current data systems. To address health and nutrition data issues, India should place emphasis on completing the Civil Registration System, modifying existing surveys in light of SDG indicators, strengthening the HMIS and Surveillance systems, and exploring application of indirect estimation techniques.
Introduction
By setting the eight Millennium Development Goals (MDGs) in the early 2000s, the United Nations galvanised unprecedented efforts to eradicate extreme poverty and hunger, achieve universal primary education, promote gender equality and empower women, reduce child mortality, improve maternal health, combat HIV/AIDS, malaria and other diseases, ensure environmental sustainability, and promote global partnership for development. Overall, substantial progress was made in reaching these goals within the predefined timeline, particularly by the developing countries. MDGs were also successful in promoting global awareness, political accountability, improved metrics, social feedback, and public pressures to achieve the objectives.[1] Yet, for all the remarkable success, inequalities between and within countries have persisted, calling for the continuation of efforts beyond 2015. In September 2015 the international community, led by the UN, declared a set of 17 Sustainable Development Goals (SDGs) related to economic development, environmental sustainability, and social inclusion, and to be achieved by 2030.
India’s endeavour to achieve the SDGs in a well-defined and time-bound manner is critical not only for the nation’s own development, but also that of the world—due to India’s mammoth population size and the extreme inequalities in demographic and socio-economic indicators across the country. Despite being one of the largest emerging economies in recent decades, India’s human development index is lower than that of other emerging countries such as China, Brazil, and Russia. India has been the single largest contributor to global under-five deaths since the 1970s.[2] The annual number of under-five deaths in India was as high as 1.2 million in 2015 and is more than six times higher than the number for another giant, China. Similarly, India is one of the top five countries where 60 percent of the world’s one billion extreme poor people live.[3] The poverty head count ratio at the national level was as high as 21.9 percent in 2011-2012.[4] India continues to have widespread hunger, ranking a lowly 97 among 118 developing countries in the Global Hunger Index.[5]
What of India’s health-related MDGs—were they achieved? It is found that India successfully achieved the MDG 6 of halting and reversing the HIV epidemic.[6] However, the country failed in the MDGs related to child and maternal health. First, though India is close to attaining the goal set for the under-five mortality rate, it has missed the targets for infant mortality (39 per 1000 births in 2014 vs. targeted 27 for 2015) and maternal mortality (167 for 2011-13 vs. target of 109 in 2015). India is also reported as moderately off-track for the reversal of the incidence of malaria and other major diseases. Secondly, the national achievement in many of the indicators masks the poor performance of many populous districts from north-central and eastern India which have been lagging behind. Finally, although the level of mortality has reduced substantially in the MDG period, the absolute numbers of maternal and child deaths are massive.
Therefore, it is important for India to gear up for SDGs with well-thought out and systematic efforts especially at the local level. It is noteworthy that the SDGs are not only target-oriented but also inclusive, unlike the MDGs which emphasised on the overall attainment of goals. The battlecry for the SDGs—“No one is to be left behind”—means that for India, the overarching goal is to reduce inequalities across gender, region, class, and caste. An immediate imperative is an assessment of data availability for the precise estimation of health and nutrition metrics, at small geographical areas and for any disadvantaged socio-economic groups.
This paper examines India’s readiness to assess its progress towards the SDGs related to nutrition and health (including reproductive, maternal, newborn and child health). The following section discusses relevant SDG indicators and sources of data in India, to be followed by an examination of appropriate strategies to track SDGs. The study concludes with specific recommendations for improving data systems in India for the short and long term.
Sustainable Development Goals, Targets and Indicators in Health and Nutrition
The targets and indicators linked to health and nutrition are covered in SDG 2 and 3[7] (see appendix A). In all, there are 13 targets and 30 indicators: mortality (9), morbidity (5), nutrition (4), healthcare service utilisation (5), and the rest are in reproductive health (adolescent birth rate and met need for contraception), lifestyle factors (tobacco use), and substance abuse. Two other indicators (3.b.2 and 3.d.1) which refer to development assistance and international regulations are not addressed in this paper.
In addition to the health and nutrition targets related to SDG 2 and 3, there are other indicators (often labelled as WASH) under SDG 6 (management of sanitation and water) which directly influence health; these will be discussed in this paper. Some indicators from SDG 5 (Gender equality), SDG 10 (reduce inequality), and SDG 16 (peaceful and inclusive societies) also have an influence on health, but these are beyond the scope of this study.
The UN has decreed that SDG indicators should be disaggregated, wherever relevant, by income, sex, age, race, ethnicity, migratory status, disability status, and geographic location, or other characteristics (General Assembly resolution 68/261). Thus India needs to monitor all of these indicators by gender, geographical regions (districts or below districts, say, parliamentary constituencies), income groups, religions, castes, and disability condition. Moreover, these are to be obtained at a series of time points or for periods between 2015 and 2030, to facilitate continuous monitoring of the country’s progress.
Major sources of data on health and nutrition in India
India’s legacy of collecting data on population and health goes back to the colonial period when the first modern Census was organised, culminating in 1872. This was followed soon by the passing of the Births, Deaths and Marriages Registration Act, 1886. Eventually, Independent India will have a long list of sources of data of various kinds, such as registration, large-scale sample surveys, and surveillance systems. Yet, due to lack of integration of these data systems, they have proved inadequate to precisely answer even some of the most basic questions in health and mortality.[8]
A Civil Registration System (CRS) is ideally suited to gather mortality information by age, sex, place, and basic socio-economic characteristics such as marital status, religion, occupation, cause of death, medical certification and type of medical attention received. It is imperative that such a system be adopted universally.[9] CRS has been fully functioning in developed countries for a long time, and the emerging countries such as Brazil, China, South Africa too have reached complete coverage in recent decades.[10] India, meanwhile, has long had a CRS overseen by the Office of Registrar General, offering vital statistics. However, despite the provision of legal action in case of failure to register births and deaths, coverage of death registration is poor and varies widely by age, sex, state, and over time. Therefore, a Sample Registration System (SRS) was introduced in 1970 which follows a dual recording scheme in a sample of villages and urban blocks. Though meant as an interim measure until the coverage of the CRS reaches a satisfactory level, the SRS has continued for long and estimates derived from it have been, by default, widely accepted as reliable. The SRS, based on a population of about 7.5 million population, publishes a yearly bulletin, yearly report, and abridged life tables at regular intervals. For large states, the SRS gives estimates of age-specific death rates by sex and place of residence (rural/urban); for smaller states and union territories, only the crude death rate and the infant mortality rates are available. For large states, some key indicators are also given for natural divisions (which are groups of districts within states).
The major limitations of SRS include lack of district-level indicators, inadequate information for smaller states, absence of information on basic socio-economic characteristics, and non-availability of individual-level data in the public domain. With the introduction of MDGs, district-level demographic indicators became essential especially in the most populous and backward states of central, northern, and eastern India, popularly known as EAG states (Bihar, Chhattisgarh, Jharkhand, Madhya Pradesh, Odisha, Rajasthan, Uttar Pradesh, and Uttarakhand), as well as Assam. Therefore, instead of emphasising on CRS or expanding the SRS to district-level disaggregation, the Office of Registrar General of India (ORGI) introduced the Annual Health Survey (AHS) in the EAG states and Assam in 2011. Unlike CRS or SRS, AHS provides data on other health indicators such as morbidity prevalence and healthcare service utilisation, especially for maternal and child health.
Having reliable statistics on the causes of death in India is still a faraway dream. The two major systems of data on causes of death–i.e., Survey of Causes of Death (SCD-Rural) and Medical Certification on Causes of Death (MCCD)—are unable to provide cause-specific death rates due to their unscientific sampling design, poor coverage, non-availability of the size of exposed population, and irregularities in publishing reports. The most recent ‘Special Survey of Deaths’, for example, conducted under SRS by the Office of Registrar General India, appears to be an improvement over the earlier data on causes of death; however, individual-level data is still not available, making it impossible for researchers to evaluate the nuances of cause-specific death rates across the country.
Much of this data vacuum in health and mortality was filled recently by several multi-purpose, large-scale surveys such as the National Family Health Survey (NFHS), District Level Household Surveys (DLHS), National Sample Survey (NSS), India Human Development Survey (IHDS), Rapid Survey on Children (RSOC), and WHO-SAGE. Many of these data sets allowed estimating the most important mortality indicators (infant mortality rate, child mortality rate, and in some cases, adult mortality rate) by demographic and socio-economic characteristics, at least at the state level. Some of these surveys also provide reported morbidity prevalence in India across population subgroups, though only a few diseases are covered and these have varied across survey rounds. While the NSS provided data on nutritional intake, surveys like NFHS, DLHS, IHDS, and RSOC did an impressive job by offering anthropometric measures of mothers and children in India. Finally, these data served as a rich source for information on healthcare service utilisation, particularly by women of reproductive age and children under five.
The biggest advantage of these surveys over SRS is that individual-level data are shared with researchers and policymakers, giving them more scope to analyse the dynamics of mortality, morbidity, disability and healthcare service utilisation across the breadth of the Indian population. However, because of the limited sample size and uncertain periodicity of these surveys, they fail to meet the requirement of evaluation of target-oriented policies at the district or below-district level. Although the overall sample size in the surveys is huge, district sample size is not conducive for assessing mortality differentials due to large sampling errors.
Among all health management data tools, the most important one may yet be the Health Management Information System (HMIS), launched in 2008 by the Ministry of Health and Family Welfare. HMIS was introduced to meet the demand for micro-level data on population and health at the level of the health facility. It has been providing maternal and child health indicators at small geographical/administrative/health facility level. The HMIS seeks to capture all vital events and MCH indicators and transmit the information electronically to ensure quick collation and tabulation. It is supposed to be the most powerful source in tracking health programme’s performance since data are available at health facility level and on a monthly or weekly basis.
However, the problems with HMIS begins with its structure: it does not provide any information on exposed population, and as a result, the calculation of mortality rate is not feasible. Overall, HMIS data are found to be incomplete and of poor quality; there are irregularities in report generation, data duplication and data inconsistencies at various levels of healthcare facilities.[11] There is also over-reporting of the positive indicators (for instance, maternal and child healthcare service utilisation), and under-reporting of the negative indicators (such as under-five deaths). A most recent evaluation shows that the quality of records in HMIS is sub-optimal and there is over-reporting in some MCH indicators.[12] HMIS data on mortality are grossly under-reported due to the following reasons: 1) under-reporting of deaths by government health facilities; 2) absence of reporting by the private health facilities; and 3) variation of coverage among states/districts.[13] These inadequacies in HMIS at the grassroots level may be attributed to a host of factors, including shortfalls in human resources, lack of proper training, and deficiency in infrastructure (such as internet connectivity and power supply).
For morbidity indicators, another major source of information comprises the recorded incidence of specific diseases obtained from surveillance or notification. India has launched programmes to control/eradicate a number of diseases and as part of such programmes, the incidences (cases) of these are recorded. This has been done for malaria, tuberculosis, leprosy, and HIV/AIDS. Many diseases are covered by the National Vector Borne Disease Control Programs (now called Expanded Vector Borne Disease Control Programme- EVBCDP). An integrated system of recording incidence, the Integrated Disease Surveillance Programme (IDSP), has recently been established under the Ministry of Health and Family Welfare. The mission of this system is to detect epidemic- prone diseases early and take timely and effective public health action.[14] IDSP publishes weekly reports on outbreaks (malaria, viral hepatitis, food poisoning, and other neglected tropical diseases like dengue, chikungunya, and leprosy). There are, however, limitations in disease surveillance systems in India such as under-reporting of deaths and inadequate laboratory diagnosis.
Though notification of many diseases is mandatory, one is not sure how many patients who seek care in private healthcare institutions get covered. As a large proportion of the population obtain care from the private sector, completeness of the data based on reporting-recording is questionable. It is pertinent to note that according to the 71st round of the NSS, private doctors are the source of treatment for about half of all spells of illnesses treated.[15] For HIV/AIDS, Sentinel Surveillance has been in place for long and model-based estimation has been adopted.
Measurement readiness for Health and Nutrition goals in current system of Data
As noted in appendix A, a set of health and nutrition metrics has been finalised to track regularly the progress in the achievement of the SDGs. This section discusses the availability of each of those indicators at disaggregated level. For this purpose, the indicators have been grouped according to: i) mortality; ii) morbidity; iii) health services utilisation, reproductive health, and substance use; iv) nutrition; and v) water supply and sanitation. (See Tables 1 to 5).
In general, most of the mortality indicators are available in India at the state level (See Table 1). However, except AHS (in Empowered Action Groups of states and Assam), none of the data sources provide district-level indicators of mortality. At the same time, mortality indicators are not available regularly by socio-economic characteristics at the district level or small state level. While such estimates can be obtained from unit-level data (the NFHS and DLHS data are in the public domain), relative sampling errors are quite large for small socioeconomic groups, preventing assessments of changes over short time intervals.
Table 1: Availability of SDG mortality indicators at disaggregated level, India
| Indicators | Sources of Data | Periodicity | Reference period | Lowest Geographical disaggregation | Availability of indicator for socioeconomic groups | Availability of Individual data | Comment |
| 3.1.1 Maternal Mortality Ratio | NFHS , 1, 2, 4 | Not regular | 2 years before survey year | States | No | Yes | Relative Sampling error is large |
| SRS | Every three year | 3 years | Large States and groups of small states | No | No | Relative Sampling error is large | |
| AHS* | Annual | 3 years | Only EAG states and Assam ‘s districts | No | Yes | Only nine states covered | |
3.2.1 Under-five mortality rate
and
3.2.2 Neonatal Mortality Rate
| SRS | Annual | Yearly | States | Yes, at state level | No | Problematic data in small states |
| NFHS 1, 2, 3, 4 | Not regular | 5 years prior to surveys | States; rural/urban | Yes, at state level | Yes | Relative Sampling error is large below state level | |
| DLHS* 1, 2, 3, 4 | Not regular |
3 years
prior to surveys
| District | Yes, at state level | Yes | Relative Sampling error is large below state level | |
| AHS * | Yearly | 3 years prior to surveys | Only EAG states and Assam ‘s districts | No | Yes | Only nine states covered | |
| 3.4.1 Mortality rate attributed to cardiovascular disease, cancer, diabetes or chronic respiratory disease; Cause Specific Death Rate (CSDR) | SRS | Not regular | 2-3 years | India | No | No | No published report on CSDR; small sample for states |
| 3.4.2 Suicide Mortality Rate (SMR) | SRS | Not regular | 2-3 years | India | No | No | No published report on SMR; small sample for states |
| 3.6.1 Death rate due to road traffic Injuries (TIDR) | SRS | Not regular | 2-3 years | India | No | No | No published report on TIDR; small sample for states |
| 3.9.1 Mortality rate attributed to household and ambient air pollution | No | NA | NA | NA | NA | NA | NA |
| 3.9.2 Mortality rate attributed to unsafe water, unsafe sanitation and lack of hygiene | No | NA | NA | NA | NA | NA | NA |
| 3.9.3 Mortality rate attributed to unintentional poisoning | No | NA | NA | NA | NA | NA | NA |
*: The DLHS and AHS have now been discontinued.
Data availability is much worse for cause-specific death rates. There is no published report by ORGI giving cause-specific death rates in India. The published reports on causes on deaths (2001-2003; 2004-2006 and 2010-2013) provide only the distribution of deaths by causes. Since these reports do not provide detailed information on the number of persons exposed and events by age, sex and causes of death, it is not possible for researchers to calculate SDG indicators related to cause-specific deaths. Besides, the age groups for which distributions are published are broad and for a large proportion of deaths, the cause of death is not known. Further, there is hardly any available data on mortality attributable to pollution and poor sanitation and hygiene.
Table 2. Availability of SDG morbidity indicators at disaggregated level, India
| Indicators | Sources of Data | Periodicity | Reference period | Lowest Geographical disaggregation | Availability of indicator for socioeconomic groups | Availability of Individual data | Comment |
| 3.3.1 Number of new HIV infections per 1,000 uninfected population, by sex, age and key populations | Integrated Biological and Behavioral Surveillance (IBBS), India/ HIV Sentinel Surveillance | Yearly | 1 Year | National | No | No | Model based estimation |
| NFHS 3 | Not regular | 1 year | Some States/group of states | Yes | Yes | Reliable for the areas covered | |
| 3.3.2 Tuberculosis incidence per 1,000 population | NFHS 3, 4, | Not regular | 1 year | State | Yes | Yes | Lay reporting |
| IHDS 1& 2 | Not regular | Prevalence at the time of survey | State | Yes | Yes | Lay reporting | |
| NSSO 71stround | Not regular | 15 days prior to survey | NSSO region | Yes | Yes | Lay reporting | |
| DLHS 4 * | Not regular | 1 year | Districts of Non-EAG states and Assam | Yes | Yes | Lay reporting, Not nationally representative | |
Revised National
Tuberculosis Control Programme (RNTCP) and NIKSHAY
| Regular | Yearly | State | No | No | Only reported cases get covered | |
| 3.3.3 Malaria incidence per 1,000 population | NFHS 2 | Not regular | 3 months prior to survey | State | Yes | Yes | Lay reporting, seasonal variations in Malaria |
| DLHS 4 * | Not regular | 15 days prior to survey | State and district excluding EAG states and Assam | Yes | Yes | Lay reorting,Nationally not representative | |
| National Malaria Eradication Programme (NMEP); now Expanded Vector Borne Disease Control Programme (EVBDCP) | Regular | Monthly | District | No | No | Only reported cases get covered | |
| 3.3.4 Hepatitis B incidence per 100,000 population | No data on population based information | NA | NA | NA | NA | NA | NA |
| 3.3.5 Number of people requiring interventions against neglected tropical diseases | Integrated Disease Surveillance Programme (IDSP); Expanded Vector Borne Disease Control Programme (EVBDCP) | Regular | Weekly | District | No | Yes | Coverage likely to be poor |
*: The DLHS has now been discontinued.
Table 2 presents the availability of SDG related morbidity indicators at disaggregated level. Majority of the morbidity indicators are available at national and state level through large-scale surveys. NFHS is an effective source for some selected morbidities such as HIV/AIDS and TB since questions are designed to estimate the prevalence of these diseases. DLHS, IHDS and NSSO also collected information on diagnosed morbidity. But surveys rely on reports by survey respondents (lay reporting) rather than by professionals and thus morbidity incidence/prevalence based on these surveys may not be correctly estimated. These indicators are also not available at district or below-district level. Further, as there is seasonality in incidence of some diseases, prevalence on the survey date or in a reference period (usually two weeks) prior to the survey may not give an accurate representation of the average prevalence level. Surveys with several rounds spread over the year can address seasonality and this was done in some rounds of the NSS but not in other surveys. As noted earlier, estimates based on reported incidence from the system suffer from incompleteness of coverage. Unless notification becomes universal, these are bound to underestimate the true incidence. Compliance of the private health sector to the requirement of notification is essential but difficult to achieve.
Most of the healthcare service utilisation indicators are available at the district level and by socio-economic characteristics within states from various surveys (See Table 3) and can be obtained within districts for large groups from the unit level data. In particular, most surveys capture information on maternal and child healthcare quite comprehensively. Unmet need for family planning is also estimated in the NFHS, DLHS, and AHS, though the criteria for ascertaining met needs have varied. Though estimates of adolescent birth rates are also available, the relative sampling errors can be large for this, given the small number of women in this age group. On the number of workers in the health sector, data are available from the Census, but as the Census is decennial, recent changes cannot be assessed. Government reports give the number of health workers in the public sector but it is difficult to get information on the number in the private sector which provides a major portion of curative services. The professional registries could serve the purpose provided these are updated regularly. Moreover, there is no information from nationally representative surveys/Census on proportion of the population with access to affordable medicines and vaccines on a sustainable basis and on substance abuse.
No comments:
Post a Comment