When government health workers gave Thenmozhi and her friends condoms, they would blow them up and bat them back and forth in the streets of their slum in Chennai, giggling. It didn't really occur to her that she ought to use the condoms with her clients, the neighbourhood men who paid her for sex, to protect herself from AIDS.
That was a disease of other people, other countries. Thenmozhi (who like many in her community uses only one name) had many problems - a drinking, philandering husband who once set her on fire when he got angry, and feeding her children with no job and a Grade 3 education - but AIDS was not one of them.
And then five years ago, a different kind of health worker dropped by her two-room flat - a woman Thenmozhi knew, who made about $50 a month selling sex like she did. She sat on the cement floor, helped Thenmozhi pick through some rice and told her there was AIDS in India - in fact, right there in their crowded fishermen's slum in the capital of the southern Tamil Nadu province. It had recently killed a man they both knew.
She invited Thenmozhi to a community centre, where she heard informal lectures about the virus and how sex workers must band together to insist clients use protection. When she left, her handbag was full of condoms and the results of her free HIV test: negative.
Thenmozhi went on to do a most remarkable thing: "After that, I always used a condom. With every man who paid." And when she passed the age of 40 last year and transitioned into the role of madam - renting her bedroom to younger women and their clients - she handed each of them a condom, too.
Although her actions may seem logical and obvious, AIDS-education programs around the world have found that people rarely do the logical, obvious thing and use condoms once they learn about the risk of HIV. They may use condoms sometimes, in some cases. They almost never use them with the zealousness of Thenmozhi.
And yet she is no aberration. She is simply one example of the way the country has cut its rate of AIDS infection in half in the last decade, moved away from the brink of catastrophe and quietly achieved a great but unheralded public-health victory.
In southern India, HIV incidence (the rate of new infections) was 2 per cent per year in 2000; by 2007 it was just below 1 per cent. In the north, where HIV is far less prevalent, there was no large decline, but also no increase.
Only much-smaller Thailand, which implemented a mandatory-condom campaign in its sex industry in the 1990s, has ever posted similar declines.
What has happened here is starting to draw global attention. Yet so much of this story is unique to India, with its strengths (such as pro-active governments) and its weaknesses (particularly the rigid control kept over its female citizens) that it's questionable how much its example can be applied anywhere else.
When India announced in 2007 that it had 2.3 million people living with HIV, rather than the 5.7 million reported the year before, the government first attributed much of the change to better data collection. Many in the AIDS field were skeptical.
"We and all the other AIDS organizations think the number of people accessing services is increasing, so why are the infection numbers so low?" says Anjali Gopalan, head of the Delhi-based Naz Foundation and one of India's most prominent AIDS activists.
Northern states have weak mechanisms for reporting AIDS cases, while marginalized populations such as the transgender sex workers and drug users she works with are never part of door-to-door surveys, she says, so tens of thousands of cases may be missing from the official statistics. And in all regions, many people are still going without adequate treatment.
Nevertheless, more and more research points to a substantial change in sexual behaviour and with it a decline in the spread of HIV. Syphilis infections, which closely mirror HIV, have had an almost-identical drop.
"The decline is real. The numbers have plausibility and credibility," says Prabhat Jha, an epidemiologist who directs the Centre for Global Health Research in Toronto, and was an architect of India's early AIDS-control program. "There has been a profound change in behaviour among clients and sex workers that accounts for most of the drop."
While India has a significant population of intravenous drug users - as many as half of whom, in some areas, have HIV-AIDS - the bulk of the HIV transmission in this country happens through heterosexual sex. Men, often married, pay professionals for sex, get infected and then pass the virus to their wives or regular partners, who infect children at birth.
In 1997, Prof. Jha - who is Indian-born but grew up in Winnipeg - found himself in Kamithipura, the teeming red light district of Mumbai. He had just been hired by the World Bank to design a national AIDS intervention for India, and with him was another Manitoban, Frank Plummer.
Now the director of the national Centre for Infectious Disease Prevention and Control in Ottawa, Dr. Plummer was then a microbiologist renowned for his work in the early days of HIV in East Africa. Prof. Jha wanted his assessment of India's situation.
"Frank looked around, and he said, 'This looks just like Nairobi in 1984,'" Prof. Jha recalls. Those words made him shudder: He was well aware of the horrifying swath of destruction AIDS went on to cut through the slums of Kenya and so many other African nations.
But the parallel was obvious: In India in 1997, HIV infection was exploding among sex workers, drug users and truck drivers in congested urban areas. But almost no one knew about the disease and no one was doing anything to protect themselves or their other partners.
In a country of this size, with a frail public health system, it was an unfolding disaster. The newly-formed National AIDS Control Organization said the country's infection rate was the fastest-growing in the world. Prof. Jah ran mathematical models that predicted more than 20 million Indians with HIV by 2010.
But it didn't happen. The aversion of this crisis has many who work in AIDS control feeling justifiably proud - theirs is a significant achievement in a field notable for its rare victories. Billions of dollars and 15 years of effort in the worst-affected nations in Africa have yielded declines in HIV infection of at best 5 or 6 per cent (except possibly in Uganda, which saw a very sharp drop in the late 1980s - the reasons are hotly debated today).
Ask those involved, such as Prof. Jha - who had a $200-million budget from the World Bank to try to stop infections - what worked, and they list mass distribution of free condoms; the use of "peer educators" such as the sex worker who visited Thenmozhi; and a media campaign full of frank messages.
All of which have been tried, and tried, and tried in Africa, with only meagre impact. So why did they work in India?
Windows of opportunity
The first difference is timing. HIV was spreading quickly in Africa by the mid-1960s, yet it was 20 years before anyone tested an African for it.
By the time real efforts to stop the virus got under way, hundreds of thousands of people had already died and in many countries 10 to 15 per cent of the general population was infected.
The first HIV cases in India, on the other hand, were not found until 1986, here in Chennai, when two men tested positive. For the next few years, there were only a handful of cases.
By the time some far-sighted bureaucrats got on the job in the mid-1990s, HIV was just at the 1-per-cent rate which epidemiologists consider its viral tipping point - and, most critically, not yet in the general population but still largely confined to high-risk groups.
"We still had a window of opportunity," recalls S. Ramasundaram.
Today he heads the department of public works in Tamil Nadu's state government, but back then, he was a director in the health department, and was handed the new AIDS file.
He had a background in demography - and looking at Prof. Jha's models of projected spread of HIV, he had a similar full-body shudder of horror.
"I argued with the government. Very rarely in development do you get a crystal ball. I said, 'If you don't do something, it's going to be doomsday.'"
Mr. Ramasundaram was talking - urgently and often - to state leaders, and that is the next key difference from Africa. Tamil Nadu is the state that has posted the greatest gains on HIV and its government has been active and engaged since the day those first cases were reported - in sharp contrast to the denial with which many African leaders first confronted HIV.
(The huge sizes of Indian states - Tamil Nadu's population is 64 million, larger than three-quarters of African countries - and the differences between them make comparing Indian states and African nations more accurate than using India as a whole.)
Mr. Ramasundaram launched an AIDS-control organization at arm's length from the government, to minimize bureaucratic slowdowns and corruption.
He brought in the best private advertising agencies in the state, gave them a budget bigger than Coke or Pepsi (then the biggest spenders) and had them vet all their ads with people living with HIV.
Then he blanketed the state: billboards, cricket stadiums, movies and newspapers in every language.
"There were so many messages on HIV then - the Chief Minister called me and said, 'You're frightening people.' I said, 'Sir, that's the point.' He said, 'It looks like we have a big epidemic!' I said, 'No, we want to prevent a big epidemic.'"
Next he turned to community groups that were already working in slums and with sex workers, and tasked them with going door-to-door to spread the word.
The leader was Lakshmi Bai, a fast-talking, whirling-sari-clad social scientist with years of experience with sex workers, including not only women but gay and transgender men.
She eschewed the idea of a straight-up AIDS program; instead, she involved the sex workers in projects to build their self-esteem, organized them into collectives with food and clothing banks, and pushed them to confront government with their needs.
"You don't think only about just one disease. You can't talk just about AIDS," explains Ms. Bai, who now runs the non-governmental Tamil Nadu AIDS Initiative. "But when you are doing all these things together, they are going to listen to what you are saying about HIV."
"My body is a temple and I have to take care of it - the director has told us we are precious people and God's spirit lives in us," says Thenmozhi, speaking with the kind of reverence that many sex workers seem to have for Ms. Bai. So, she added, condoms only make sense.
The Tamil Nadu program benefited from bitter experience in Africa, where the thinking had been that for each dollar of AIDS prevention funding, 20 cents should be spent on the high-risk groups and the rest on the general population.
By the late 1990s, Mr. Ramasundaram explained, research had made clear that there would be far more impact from spending the whole dollar on those high-risk groups - and keeping the virus out of the general population. Condom use by sex workers rose from 40 per cent to 90 per cent in three years.
Meanwhile, the billboards and movie ads were addressing a particular population: "Don't treat Tamil Nadu as an African country … the literacy level is higher, the adaptation to change is faster and the technology absorption is higher," Dr. S. Vijayakumar, now head of the state AIDS agency, says (with a certain smugness that often characterizes the reflections of those in the field here).
However, in terms of one key bit of technology, there was indeed a crucial difference in India: Condoms had been actively promoted here since Indira Gandhi's population-control policies of the 1970s.
Also, there was little of the cultural distaste and discomfort that has greeted condom campaigns in Africa - and no conservative Christian church to lead a public outcry about abstinence.
There were, however, plenty of trained lab technicians and statisticians and the sort of qualified staff an AIDS program needs, the human resources that are so often lacking in Africa.
And there was cash. When Mr. Ramasundaram set up his new state AIDS organization, the World Bank offered up millions of dollars, which meant he didn't have to compete for scarce state-health resources.
"It was crucial that we had that source of funding," he says.
International agencies rushed to support India's AIDS response in its infancy; the Bill and Melinda Gates Foundation alone pledged $342-million (U.S.) over 10 years from 2004, its largest program anywhere.
But there are also less-pleasant truths about India's victory over HIV. Beyond literacy, condoms, blunt ads and brilliant bureaucrats, one thing more than any other has checked the spread of the virus here: the oppression of Indian women.
The extreme control exerted over women's personal lives - first by their parents, then by their husbands and in-laws - means that very few ever have the opportunity to have a sexual partner other than their husbands.
Where 25 per cent of men report more than one sexual partner, less than 2 per cent of women do. Married women get infected by their husbands, and sometimes pass HIV to their children, but the virus stops there: They do not have other partners to pass HIV on to.
This is a marked contrast to Africa, where it is now clear that the "concurrent sexual network" - the tendency for both men and women to have overlapping partners rather than serial ones - has been the key driver of the epidemic.
(Meanwhile, discrimination has played a sharply different role in the spread of HIV among men who have sex with men - it has extremely limited AIDS organizations' ability effectively to provide these men condoms and information. As a consequence, they have HIV infection rates 10 times those of the general population.)
There is, in fact, a broader issue of culture at play in India's AIDS success story, the sort of squishy subject that makes AIDS researchers flinch because it lies so far outside tidy quantifiable data.
But many in the field agree that Indian society remains rigidly hierarchical, still infused with the powerful role of the caste system, and people are accustomed to the strong role of government in their lives.
That's a contrast to many African countries with weaker states and more egalitarian societies. And it meant that when the Indian government sternly told people to use condoms and cut back on partners, they listened.
Mr. Vijayakumar believes that in the next couple of years, his AIDS control agency can drive new infections down to zero. It's a breathtakingly ambitious goal - it has never been done anywhere else - but he brandishes an impressive array of maps, charts and software programs to demonstrate just how he is going to do it.
His office collects data from every possible source - from blood banks to maternity hospitals to neighbourhood clinics for sex workers - and can pinpoint where each new infection comes from.
He has a three-pronged strategy based on continued prevention messages, better reach of the interventions that prevent parents from infecting children, and continued work with the high-risk groups.
"We should be able to do this," he says, working long past dark in an office where a steady flow of assistants ebbed in and out bearing yet more charts and data sets.
"I have a plan in place - my problem is my high-risk groups. If I can bring them into the health fold we'll certainly be able to do it."
But Mr. Vijayakumar is watching his budget shrink, and government, donor and public attention shift away from HIV, as success itself eases the sense of panic.
Many say the shift in government funds is justified, given how few people HIV kills in comparison with basic public-health problems such as water-borne diarrhea, child malnutrition, smoking or road accidents. The Gates Foundation is redirecting its funding to issues such as maternal and newborn care.
Yet HIV remains of critical concern here: With 2.3-million infected people, this country has the third-largest burden of HIV-AIDS in the world, and has succeeded in getting treatment to fewer than half of the people who need it.
The successes achieved have been mostly in the richer south of the country. The outstanding question is the north, with much weaker governments and health systems that have yet to embark on serious AIDS-control programs.
Other factors make the north vulnerable too. "There are large numbers of migrant workers from Uttar Pradesh and Bihar, and they are a huge worry," says Mr. Ramasundaram. A major factor in the African pandemic has been workers spending months or years away from their wives and paying for sex in their host cities.
Overall, the precise situation with HIV in the north is unknown - mother-child transmission could be exploding.
India's AIDS interventions have been relatively cheap - a tenth the cost of Thailand's sex-worker and condom intervention - but they are not free. "The era of 'Big HIV' in India may be over, and we know that once condoms become common in commercial sex, they stay common," says Prof. Jha. "But there is a huge 'if' - the Indian government has to continue to pay for the cheap and effective ways to curb HIV for at least the next decade, especially in North India."
In Chennai, Lakshmi Bai, who has lived through a rare, radical shift in sexual behaviour, is as often gloomy as she is encouraged. "So much life has been lost," she says - a fact often glossed over by the bureaucrats and researchers excited about the falling infection rates. While many HIV-related deaths are not reported as AIDS, at minimum several hundred thousand Indians have already died of the disease, Prof. Jha estimates.
"Everything is not rosy here," says Ms. Bai. "There is so much to do! I'm quite worried about sustainability, with the funders leaving. Even now, with all this 'empowerment' blah-blah, sexual decisions are taken by men."
There are 50,000 Tamil Nadu sex workers on her books today, but new young girls - and young men - show up all the time. "If you're not going to continuously address, what will happen? If these things are not done - disaster only."
But when Thenmozhi sits in a circle with the women at the drop-in centre, when they joke and gently mock their clients, the married men "who say they can't eat the same food every day," disaster seems far away. The women bemoan their troubles - shirking husbands and mounting bills. But AIDS is not one of them.
Stephanie Nolen is The Globe and Mail's correspondent in South Asia.