The 2005-06 National Family Health Survey (NFHS-3) - due to be out in its totality by the first quarter of 2007 - has already had a dry run. Provisional data from five states - Orissa, Maharashtra, Chhattisgarh, Gujarat and Punjab - is in, and data from seven more is due in by early 2007.

While the survey is not big enough to cull district- and village-level data, and Union Territories were not included, it did cover all 29 states and was buttressed by more than 30 organizations involved in it. The approximate sample sizes for interview at the national level are 110,000 ever-married women (i.e. women who have been married at any point in their lives), 33,000 never-married women and 78,000 men.

The complete NFHS-3 - for which fieldwork was conducted between December 2005 and August 2006 - will have extensive data on women's empowerment and the minutiae of domestic violence. For the first time since the NFHS began in 1992-93 (NFHS-1), it will also be possible to extrapolate information on violence against non-married women. Non-married women and men have been included for the first time.

NFHS surveys are conducted under the stewardship of the Ministry of Health and Family Welfare (MOHFW). The International Institute for Population Studies, which was the nodal agency for NFHS-1 and -2, is also the nodal agency for NFHS-3. USAID, the Department for International Development, the Bill and Melinda Gates Foundation, UNICEF, UNFPA and the MOHFW funded NFHS-3. ORC Macro of the US - which implements Measure DHS (demographic and health surveys), the most widely-accepted source of health and population data in the developing world - provided technical assistance to the survey.
Of as much pertinence as health information related to women is that HIV prevalence is being measured for the first time. Of course, the 112,000 tests conducted for NFHS-3 might seem a drop in the bucket (given that the government accepts that 5.2 million are living with HIV/AIDS in the country - National AIDS Control Organization (NACO), April 2006). Said Laurie Liskin, ORC Macro's Senior Advisor for Communication, "In every country we've surveyed, being surveyed doesn't translate into HIV testing."

Then, again, for NFHS-3, one in six households inhaled deep and went in for blood testing anyway. More than 200,000 anemia tests were carried out on women aged 15-49, men aged 15-54 and children under 5 years using a small, handheld, 'field-friendly' device. It was battery-operated and required no refrigeration - the one aspect that nearly proved fatal for India's pulse polio programme. It was also non-intrusive, needed only a small drop of blood for analysis and came up with results within a minute.

In all categories of gender indicators in the five states whose results are out, the statistics on women were indicative of social infirmity: Body Mass Index (BMI) was low, ranging from 41 per cent of women in Orissa and Chhattisgarh (compared to 32 per cent for men) to the relative 'normalcy' of 14 per cent of women in Punjab (compared to 12 per cent in men). The only consolation was that, relative to NFHS-2, the BMI normalized by an average of seven per cent in the said states.

Most parameters, in fact, show an improvement over the previous NFHSs, says Fred Arnold, Vice-President, ORC Macro, "Child marriage, for instance, declined drastically from NFHS-1 to NFHS-3. It's very rare, nationally." The improvement is entirely relative, though: anemia among women is still endemic, ranging from a 'low' of 38 per cent in Punjab to a 'high' of 63 per cent in Orissa; the percentage of ever-married women who have experienced spousal violence ranges from 25 per cent in Punjab to 39 per cent in Orissa. (This being the first time that spousal violence has been configured as part of the survey, there are no comparable statistics for NFHS-2.)

But unprecedented data on violence against women in the context of health might permit the reworking of policy decisions. "The survey found that infant mortality, for instance, is higher among abused women, as are other juvenile problems, and those of families and women themselves," said Sunita Kishor, senior gender specialist, ORC Macro.

Detailed data on HIV and AIDS could also determine how the NACO handles the future course of the disease in the country. NACO will use three sources of information: 1) Sentinel survey data on women; 2) NFHS-3 data on the general population; 3) Surveillance data on high-risk groups. The HIV/AIDS part of the survey - which left out women in brothels and truckers on the move on the debatable ground that the former were a minuscule proportion of the population and the latter impossible to nail down - found that more men than women had heard of AIDS (men: from 92 per cent in Punjab to 67 per cent in Chhattisgarh; women: 70 per cent in Punjab to 41 per cent in Chhattisgarh).

The survey's HIV content is expansive but not exhaustive: it includes knowledge on disease and treatment, discussions of HIV during antenatal visits, previous HIV tests, stigma and discrimination, medically safe injections, male circumcision, sexual behavior (which the surveyors concede is probably underreported), and condom use.

Samples for HIV testing were collected using the carrot approach: vouchers for free HIV testing at any local clinic were given to those who agreed to provide the samples, but there was no monitoring to check how many actually went to the clinics. No names or personal identifiers were placed on the filter paper that carried the blood sample, so no one was at risk of being identified. While this might permit a more fine-tuned macro approach to the HIV/AIDS issue, it prevents interested organizations from approximating a spot-focus that could enable them to take prophylaxis and rehab to specific locations.

For obvious reasons, ORC Macro preferred to keep out of the ongoing Indian government versus WHO controversy on HIV/AIDS statistics in India. "All that we found in the 15 countries we'd already surveyed," said Arnold, "was that HIV/AIDS figures actually fell from what had previously been conjectured."

Nonetheless, the NFHS-3's new or expanded content areas should be of interest to various organizations: NACO would be interested in family life education, as should be UNICEF in child labor. Among the other new areas are birth registration, contraceptive continuation and failure rates, prenatal mortality, male involvement in family planning and maternal health, malaria prevention, tuberculosis-related knowledge and stigma, nutritional status of man and never-married women, and Vitamin A supplementation during pregnancy.

If the NFHS-3 can provide all this, it might - despite the small survey corpus - redirect the government to a new exploration of family health, one of the plinths of a workable developmental policy. Furthermore, since all the data sets will be publicly available, they might help interested global and national organizations - government and otherwise - join in formulating a developmental strategy that doesn't have much of India's population falling through the cracks.

(Kajal Basu is a senior journalist and freelance writer based in New Delhi.)