In a country perennially thirsty for reliable health data, the National Family Health Survey (NFHS) is like an oasis. It has a large volume of data that is openly accessible. The report of the fifth round of the NFHS was recently released (covering phase 2 States where data collection was delayed due to the novel coronavirus pandemic). Since then, we have had a spate of articles by journalists and scientists covering different aspects (malnutrition, fertility, domestic violence to name a few). It is the go-to source for many researchers and policy makers and is frequently used for various rankings by NITI Aayog.
Range and scope
For the uninitiated, the NFHS is a large survey conducted in a representative sample of households throughout India which started in 1992-93 and is repeated at an interval of about four to five years. It is the Indian version of the Demographic and Health Surveys (DHS), as it is known in other countries. Currently, the survey provides district-level information on fertility, child mortality, contraceptive practices, reproductive and child health (RCH), nutrition, and utilisation and quality of selected health services. The respondents are largely women in the reproductive age group (15-49 years) with husbands included. The fifth round covered 6,36,699 households, 7,24,115 women, and 1,01,839 men across the country. Each survey costs upwards of ₹250 crore and the funding for different rounds of NFHS has been provided by the United States Agency for International Development (USAID), the Department for International Development (DFID), the Bill and Melinda Gates Foundation (BMGF), UNICEF, the United Nations Population Fund (UNFPA), and the Ministry of Health and Family Welfare, Government of India.
Over the years its scope has been expanded to include HIV, non-communicable diseases, or NCDs (tobacco and alcohol use, hypertension, blood sugar, etc.), Vitamin D3. It has now become an omnibus train where anyone and everyone is free to climb into for a ride. It offers something for everyone. While there is a level of efficiency in adding some questions to an existing survey, this has been lost a long time ago in the NFHS. In NFHS-4, the household questionnaire had 74 questions, the women’s questionnaire was 93 pages long with 1,139 questions and the men’s questionnaire was 38 pages long with 843 questions. The NFHS-5 questionnaire was even longer. The size of the survey has obvious implications for data quality.
Other surveys and goals
The NFHS is coordinated by the International Institute for Population Sciences (IIPS Mumbai) and the actual survey is outsourced. There is an entrenched set of agencies which survive on this survey. Issues have been raised on the quality of these agencies and their workers. The NFHS is not the only survey that the Health Ministry conducts. In the last five years, it has conducted the National NCD Monitoring Survey (NNMS), the National Mental Health Survey (NMHS), the Global Adult Tobacco Survey (GATS), the alcohol survey, the Comprehensive National. Nutrition Survey (CNNS) and many others. Many of these have been implemented by premier academic institutions at costs below ₹25 crore, though none of these generated district-level estimates.
Some of these surveys are done to meet the global commitments on targets (NCDs, tobacco, etc.). However, the requirements for the monitoring of NCD targets are not met by the NFHS, as it covers an age group different than that needed for the global set of indicators. Yet, efforts to get the NNMS sanctioned met with stiff resistance as decision-makers felt that the NFHS was enough to answer those questions. As already said, for tobacco we have another vertical survey. Then why do we have questions on these in the NFHS? It is because we are confusing research with programme monitoring and surveillance needs. Questions on domestic violence and blood collection for vitamin D3 levels are good examples of this lopsided thinking.
Alignment is difficult
There have been previous attempts to align these surveys but they have failed as different advocates have different “demands” and push for inclusion of their set of questions. While the Department of Planning, Statistics and Programme Monitoring is supposed to take a final call, it lacks the technical capacity and the heft to do so and ends up using a “please-all” approach of accepting all requests with some effort at alignment. Everyone is happy, except perhaps the stakeholder with no power of negotiation — the household which is selected for the survey.
Another reason why these questions are not dropped altogether is that the NFHS is the only major survey that India has a record of doing regularly. One does not know if and when the other surveys will be repeated. For example, we do not have any surety that the second round of the NNMS will be conducted, though it is due. So, the general thinking is that “do whatever is possible, as something is better than nothing”. Multiple surveys also raise the problem of differing estimates, as is likely, due to sampling differences in the surveys. We noted this for example in tobacco, where differences in tobacco use estimates of the Global Adult Tobacco Survey (GATS) and the NNMS needed a lot of effort at reconciliation and explanation. Another example is the issue of wide divergence in sex ratio at birth reported by the NFHS and the Sample Registration System (SRS). The SRS is a better system for it as it continuously enumerates the population unlike the NFHS which is a cross-sectional survey well known for recall biases.
There must be purpose
It is time we questioned this rationale and end the over-dependence on one omnibus survey to provide all public health data for India. The experience of the NFHS and other surveys has conclusively demonstrated our capacity to conduct large-scale surveys with computer-assisted interviews and reasonable quick turnaround and cost. Can we now show that we have the capacity to plan the public health data needs for the country and ensure that these data are collected in an orderly and regular manner with appropriate budgetary allocation? This requires clarity of purpose and a hard-nosed approach to the issue. Some tough calls will have to be taken including questioning the need for vertical surveys, irrespective of national or international funding.
We have to identify a set of national-level indicators and surveys that will be done using national government funds at regular intervals. I propose just three national surveys — an abridged NFHS focusing on Reproductive and Child Health (RCH) issues, a Behavioral Surveillance Survey (focusing on HIV, NCD, water sanitation and hygiene (WASH)-related and other behaviours) and one nutrition-biological survey (entails collection of data on blood pressure, anthropometry, blood sugar, serology, etc.) done every three to five years in a staggered manner. We need to look at alternate models and choose what suits us best. This does not include data sources on mortality and the health system.
A road map
I also propose, as was done for the NNMS, that we take a national-level sample for such surveys and ask States to invest in conducting focused State-level surveys. States have to become active partners including providing financial contributions to these surveys. For a detailed understanding on some issues, each round of survey can focus on a specific area of interest. Other important public health questions can be answered by specific studies (which may or may not need a national-level study), conducted by academic institutions on a research mode based on availability of funding. It is also very important to ensure that the data arising from these surveys are in the public domain. This enables different analyses and viewpoints to be presented on the same set of data enriching the discussion and unlocks the full potential of the survey.
Are we ready to establish a public health data architecture that a country of our complexity needs? We have the technical capacity to do so. All it requires now is the political will.
Dr. Anand Krishnan is Professor at the Centre for Community Medicine, All India Institute of Medical Sciences, New Delhi. He has been involved in many of these surveys in an advisory capacity. The views expressed are personal
