Wearing sandals and draped in a dark-blue sari, Aparajita Chakraborty glides into the cluster of hilltop homes with the self-assurance of someone who has long been making house calls.
She has. For more than 30 years, Chakraborty has been visiting this extended family, doing checkups and dispensing advice. But she is no doctor, she's a community health worker who has been dispatched by the local hospital. Yet she has won the trust and gratitude of the surrounding villages by saving lives – mainly from cholera and other deadly diarrhoeal diseases.
With all the men away, either working in the rice fields or having migrated to the city, Chakraborty quickly gets down to business in the family compound of half a dozen homes. She and a colleague conduct a group interview, asking four women personal questions such as: when did they last menstruate? Are they taking the pill, or using another method of family planning?
One woman explains that she stopped taking the pill when her husband began working in Chittagong, a day's journey away. She resumed her use of contraceptives immediately after his surprise visit. By then it was too late, and she is now expecting their third child.
Another woman says she is not using any form of contraception. The woman's husband, it transpires, had a vasectomy after their fourth child. But he doesn't want his brothers to know for fear they will think him impotent. So it's a secret, albeit one that has been documented by hospital staff, along with every birth, death, marriage, divorce and other vital statistic of 225,000 people in the region.
Chakraborty knows more intimate details about the community than they know about one another. But discretion is paramount, she says: "I keep what I hear to myself. I feel like I'm part of the family."
She is part of an all-female cadre of community health workers who span this portion of Bangladesh's low-lying delta, carefully maintaining one of the longest-running and most detailed health and population data sets in the developing world.
The Matlab hospital that dispatched Chakraborty has grown extensively since 1963, when it was launched as a cholera research station atop a barge near Dhaka. The institution was set up by the International Centre for Diarrhoeal Disease Research, Bangladesh.
Half a century later, this hub for child and maternal health is widely credited for demonstrating how poor Muslim women with little or no formal education can plan their families. The approach has spread throughout this densely populated, poverty-stricken country, curbing its rapid population growth.
In 2000, the UN projected that Bangladesh's 160 million-strong population would soar to 265 million by 2050. The latest projections show the numbers are likely to climb to slightly more than 200 million by mid-century before stabilising soon after.
"Matlab showed us the way," says Ubaidur Rob, the non-profit Population Council's Bangladesh director. "Women were employed as field workers in the 1970s, when fertility was very high and female employment was virtually zero. This is where change began."
Its well-chronicled successes have made Matlab something of a mecca for public health researchers. Initially, many were drawn to the rural outpost because it suffered regular cholera epidemics. Now, the lure is the detailed population and health database that can reflect the success or failure of a drug trial or health intervention.
In the mid-70s, family planning advocates decided this was an ideal place to test whether poor, under-educated women in a religiously conservative area would adopt the use of contraceptives. To set up the experiment, researchers divided 149 villages into two groups. One half participated in the Matlab centre's maternal and child healthcare initiatives, including home delivery of modern contraceptives, while the other had access only to government services.
The communities were otherwise identical: poor fishing and farming families living in bamboo houses, in villages with little or no electricity, running water or sanitation. Most had dirt floors and cooked on fires of wood, rice chaff and cow dung.
The area was 88% Muslim and contraception was denounced by Islamic clerics. Most households practised purdah – allowing women out of the house only if "properly" covered and escorted by a male relative.
Researchers quickly learned it was not enough just to make contraceptives available, says Dr Mohammad Yunus, who ran the Matlab centre for nearly 40 years. What worked, he says, was a comprehensive doorstep service with trained female health workers making regular follow-up visits to help mothers pick a method of contraception that was best for them, treat side-effects and provide basic maternal and child healthcare.
Differences emerged immediately. Married women were more likely to use contraceptives and, over time, had an average of 1.5 fewer children than their counterparts in the comparison area. Their children were healthier, fewer women died of pregnancy-related causes, and child mortality fell.
These families grew wealthier, too. With fewer children to support, parents accumulated more farmland, built more valuable homes and gained access to running water. Their children stayed in school longer, and women enjoyed higher incomes.
The results suggested family planning was a cost-effective way to improve public health and help lift people out of poverty, health experts said. And it showed that communities do not have to become wealthier or better educated before birthrates can fall – if contraception is made available in an appropriate way.
The Matlab centre's programme drew the attention of government officials, who decided to roll it out in two areas. Matlab workers trained government employees in the door-to-door approach. By the early 80s, those areas had experienced a similar increase in contraceptive use, and the government set about training tens of thousands of female health workers using the Matlab model.
"Over the next five years, it was phased in across the whole country," Yunus says. "Bangladesh became a success story for family planning and reducing infant mortality."
Since then, average birthrates have tumbled from six children a woman to slightly more than two, and Bangladesh has become one of the first impoverished countries to meet the UN millennium development goal of reducing child mortality by two-thirds.